Marketing Metrics

MARKETING METRICS
SECOND EDITION

This page intentionally left blank

MARKETING
METRICS
SECOND EDITION

THE DEFINITIVE GUIDE TO
MEASURING MARKETING
PERFORMANCE

Paul W. Farris
Neil T. Bendle
Phillip E. Pfeifer
David J. Reibstein

Vice President, Publisher: Tim Moore
Associate Publisher and Director of Marketing: Amy Neidlinger
Executive Editor: Jeanne Glasser
Editorial Assistant: Myesha Graham
Operations Manager: Gina Kanouse
Senior Marketing Manager: Julie Phifer
Publicity Manager: Laura Czaja
Assistant Marketing Manager: Megan Colvin
Cover Designer: Chuti Prasertsith
Managing Editor: Kristy Hart
Senior Project Editor: Lori Lyons
Copy Editor: Geneil Breeze
Proofreader: Debbie Williams
Senior Indexer: Cheryl Lenser
Compositor: Nonie Ratcliff
Manufacturing Buyer: Dan Uhrig
© 2010 by Pearson Education, Inc.
Publishing as FT Press
Upper Saddle River, New Jersey 07458
FT Press offers excellent discounts on this book when ordered in quantity for bulk purchases or special
sales. For more information, please contact U.S. Corporate and Government Sales, 1-800-382-3419,
corpsales@pearsontechgroup.com. For sales outside the U.S., please contact International Sales at
international@pearson.com.
Company and product names mentioned herein are the trademarks or registered trademarks of their
respective owners.
All rights reserved. No part of this book may be reproduced, in any form or by any means, without
permission in writing from the publisher.
Printed in the United States of America
First Printing February 2010
ISBN-10: 0-13-705829-2
ISBN-13: 978-0-13-705829-7
Pearson Education LTD.
Pearson Education Australia PTY, Limited.
Pearson Education Singapore, Pte. Ltd.
Pearson Education North Asia, Ltd.
Pearson Education Canada, Ltd.
Pearson Educación de Mexico, S.A. de C.V.
Pearson Education—Japan
Pearson Education Malaysia, Pte. Ltd.
Library of Congress Cataloging-in-Publication Data
Marketing metrics : The Definitive Guide to Measuring Marketing Performance/ Paul W. Farris ... [et al.].
p. cm.
Rev. ed. of: Marketing metrics : 50+ metrics every executive should master. 2006.
Includes bibliographical references and index.
ISBN 978-0-13-705829-7 (hbk. : alk. paper) 1. Marketing research. 2. Marketing—Mathematical
models. I. Farris, Paul.
HF5415.2.M35543 2010
658.8’3—dc22
2009040210

We dedicate this book to our students, colleagues,
and consulting clients who convinced us that
a book like this would fill a real need.

CONTENTS

Acknowledgments ix
About the Authors xi
Foreword xiii
Foreword to Second Edition xv
1
INTRODUCTION 1

2
SHARE OF HEARTS, MINDS, AND MARKETS 27

3
MARGINS AND PROFITS 65

4
PRODUCT AND PORTFOLIO MANAGEMENT 109

5
CUSTOMER PROFITABILITY 153

6
SALES FORCE AND CHANNEL MANAGEMENT 181

vii

7
PRICING STRATEGY 219

8
PROMOTION 263

9
ADVERTISING MEDIA AND WEB METRICS 287

10
MARKETING AND FINANCE 337

11
THE MARKETING METRICS X-RAY 357

12
SYSTEM OF METRICS 369

Appendix A
SURVEY OF MANAGERS’ USE OF METRICS 385

Bibliography 393
Endnotes 397
Index 405

viii MARKETING METRICS

ACKNOWLEDGMENTS
We hope this book will be a step, however modest, toward clarifying the language,
construction, and meaning of many of our important marketing metrics. If we have
succeeded in making such a step, we owe thanks to a number of people.
Jerry Wind reviewed our initial concept and encouraged us to set our sights higher.
Rob Northrop, Simon Bendle, and Vince Choe read early drafts and gave valuable
feedback on the most important chapters. Eric Larson, Jordan Mitchell, Tom Disantis,
and Francisco Simon helped develop material for important sections and provided their
research skills. Gerry Allan and Alan Rimm-Kauffman allowed us to cite liberally from
their materials on customers and Internet marketing. We thank Valerie Redd and Kelly
Brandon for their help in designing, testing, and administering the survey of the metrics
that senior marketing managers use to monitor and manage their businesses.
Marc Goldstein combined business savvy with deft editing touches that improved the
readability of almost every chapter. Paula Sinnott, Tim Moore, Kayla Dugger, and their
colleagues also made significant improvements in moving from a raw manuscript to the
book in your hands.
Erv Shames, Erjen van Nierop, Peter Hedlund, Fred Telegdy, Judy Jordan, Lee Pielemier,
and Richard Johnson have collaborated on our “Allocator” management simulation and
“Management by the Numbers” online tutorials. That work helped us set the stage for
this volume. Finally, we thank Kate, Emily, Donna, and Karen, who graciously tolerated
the time sacrificed from home and social lives for the writing of this book.

ix

ABOUT THE AUTHORS
Paul W. Farris is Landmark Communications Professor and Professor of Marketing at
The Darden Graduate Business School, University of Virginia, where he has taught since
1980. Professor Farris’s research has produced award-winning articles on retail power
and the measurement of advertising effects. He has published more than 50 articles in
journals such as the Harvard Business Review, Journal of Marketing, Journal of Adver-
tising Research, and Marketing Science. He is currently developing improved techniques
for integrating marketing and financial metrics and is coauthor of several books, includ-
ing The Profit Impact of Marketing Strategy Project: Retrospect and Prospects. Farris’s con-
sulting clients have ranged from Apple and IBM to Procter & Gamble and Unilever. He
has served on boards of manufacturers, retailers, and e-Business companies. Currently,
he is a director of GSI Group, Sto Corp., and The Ohio Art Company.
Neil T. Bendle is a Ph.D. candidate in marketing at the Carlson School of Management,
University of Minnesota. While studying for his Ph.D. he has won awards for his teach-
ing, and his thesis has focused on managers’ difficulties in understanding consumer
tastes. He holds an MBA from Darden and has nearly a decade’s experience in market-
ing management, consulting, business systems improvement, and financial manage-
ment. He was responsible for measuring the success of marketing campaigns for the
British Labour Party.
Phillip E. Pfeifer, Richard S. Reynolds Professor of Business Administration at The
Darden Graduate Business School, currently specializes in direct/interactive marketing.
He has published a popular MBA textbook and more than 35 refereed articles in jour-
nals such as the Journal of Interactive Marketing, Journal of Database Marketing, Decision
Sciences, and the Journal of Forecasting. In addition to academic articles and a textbook,
Mr. Pfeifer is a prolific case writer, having been recognized in 2004 as the Darden
School’s faculty leader in terms of external case sales, and in 2008 with a Wachovia
Award for Distinguished Case writer. His teaching has won student awards and has been
recognized in Business Week’s Guide to the Best Business Schools. Recent consulting
clients include Circuit City, Procter & Gamble, and CarMax.
David J. Reibstein is Managing Director of CMO Partners and William Stewart
Woodside Professor of Marketing at the Wharton School. Regarded as one of the world’s
leading authorities on marketing, he served as Executive Director of the Marketing
Sciences Institute, and co-founded Wharton’s CMO Summit, which brings together
leading CMOs to address their most pressing challenges. Reibstein architected and

xi

teaches the Wharton Executive Education course on marketing metrics. He has an
extensive track record consulting with leading businesses, including GE, AT&T Wireless,
Shell Oil, HP, Novartis, Johnson & Johnson, Merck, and Major League Baseball. He has
served as Vice Dean and Director of Wharton’s Graduate Division, as visiting professor
at Stanford and INSEAD, and as faculty member at Harvard. He serves on the Board of
Directors of Shopzilla, And1, and several other organizations.

xii MARKETING METRICS

FOREWORD
Despite its importance, marketing is one of the least understood, least measurable func-
tions at many companies. With sales force costs, it accounts for 10 percent or more of
operating budgets at a wide range of public firms. Its effectiveness is fundamental to
stock market valuations, which often rest upon aggressive assumptions for customer
acquisition and organic growth. Nevertheless, many corporate boards lack the under-
standing to evaluate marketing strategies and expenditures. Most directors—and a ris-
ing percentage of Fortune 500 CEOs—lack deep experience in this field.
Marketing executives, for their part, often fail to develop the quantitative, analytical
skills needed to manage productivity. Right-brain thinkers may devise creative cam-
paigns to drive sales but show little interest in the wider financial impact of their work.
Frequently, they resist being held accountable even for top-line performance, asserting
that factors beyond their control—including competition—make it difficult to monitor
the results of their programs.
In this context, marketing decisions are often made without the information, expertise,
and measurable feedback needed. As Procter & Gamble’s Chief Marketing Officer has
said, “Marketing is a $450 billion industry, and we are making decisions with less data
and discipline than we apply to $100,000 decisions in other aspects of our business.”
This is a troubling state of affairs. But it can change.
In a recent article in The Wall Street Journal, I called on marketing managers to take con-
crete steps to correct it. I urged them to gather and analyze basic market data, measure
the core factors that drive their business models, analyze the profitability of individual
customer accounts, and optimize resource allocation among increasingly fragmented
media. These are analytical, data-intensive, left-brain practices. Going forward, I believe
they’ll be crucial to the success of marketing executives and their employers. As I con-
cluded in the Journal:

“Today’s boards want chief marketing officers who can speak the language of pro-
ductivity and return on investment and are willing to be held accountable. In
recent years, manufacturing, procurement and logistics have all tightened their
belts in the cause of improved productivity. As a result, marketing expenditures
account for a larger percentage of many corporate cost structures than ever before.
Today’s boards don’t need chief marketing officers who have creative flair but no
financial discipline. They need ambidextrous marketers who offer both.”

xiii

In Marketing Metrics, Farris, Bendle, Pfeifer, and Reibstein have given us a valuable
means toward this end. In a single volume, and with impressive clarity, they have
outlined the sources, strengths, and weaknesses of a broad array of marketing metrics.
They have explained how to harness those data for insight. Most importantly, they have
explained how to act on this insight—how to apply it not only in planning campaigns,
but also in measuring their impact, correcting their courses, and optimizing their
results. In essence, Marketing Metrics is a key reference for managers who aim to become
skilled in both right- and left-brain marketing. I highly recommend it for all ambidex-
trous marketers.

John A. Quelch, Lincoln Filene Professor of Business Administration and Senior
Associate Dean for International Development, Harvard Business School

xiv MARKETING METRICS

FOREWORD TO THE
SECOND EDITION
At Google, we have a saying we use quite frequently: “Data beats opinion.” In practice,
this means that for any endeavor, we first determine our key success metrics and then
measure how we are doing against them on a regular basis. This allows us to optimize
and expand those programs that are working, while sunsetting those that are not.
In today’s hyper-competitive business landscape, most marketers are compelled to take
this approach versus relying on conventional wisdom, rules of thumb, or intuition that
may have been sufficient in the past.
The challenge, of course, is knowing what to measure and exactly how to measure it.
That’s where Marketing Metrics comes in. It is the most comprehensive and authorita-
tive guide to defining, constructing, and using the metrics every marketer needs today.
This second edition adds advice on how to measure emerging topics such as social mar-
keting and brand equity, in addition to explaining indispensable marketing metrics
ranging from Return on Sales to Cannibalization Rate.
Perhaps the most pressing question in marketing today is not simply how to measure
any single outcome, but understanding how all the various metrics interconnect—and
the resulting financial consequences of your marketing decisions. Marketing Metrics
moves this discussion a major step forward by reviewing alternative integrated market-
ing measurement systems and how companies are assembling such systems for better
diagnostics and more transparent marketing models. I predict that those enterprises
who develop a deep understanding of this marketing interconnectivity will gain a sig-
nificant competitive advantage over time.
What does your boss or client think about all this? Marketing Metrics surveyed senior
marketing managers on the metrics they use to monitor and manage their business. The
results tellingly reveal that your boss and client think you should already know what to
measure and how to measure it, so there’s a sense of urgency for all of us to become
masters of marketing metrics.
In our experience at Google, marketers who move with speed, center their messages
around relevance, and use data (it beats opinion!) are best-positioned for success with
today’s buyers and modern media vehicles. I therefore heartily recommend Marketing
Metrics as the foundation of the data portion of this three-pronged marketing strategy!
Jim Lecinski
Managing Director, U.S. Sales & Service, Google

xv

1
INTRODUCTION

In recent years, data-based marketing has swept through the business world. In its wake,
measurable performance and accountability have become the keys to marketing success.
However, few managers appreciate the range of metrics by which they can evaluate
marketing strategies and dynamics. Fewer still understand the pros, cons, and nuances
of each.
In this environment, we have come to recognize that marketers, general managers, and
business students need a comprehensive, practical reference on the metrics used to
judge marketing programs and quantify their results. In this book, we seek to provide
that reference. We wish our readers great success with it.

1.1 What Is a Metric?
A metric is a measuring system that quantifies a trend, dynamic, or characteristic.1
In virtually all disciplines, practitioners use metrics to explain phenomena, diagnose
causes, share findings, and project the results of future events. Throughout the worlds of
science, business, and government, metrics encourage rigor and objectivity. They make
it possible to compare observations across regions and time periods. They facilitate
understanding and collaboration.

1.2 Why Do You Need Metrics?
“When you can measure what you are speaking about, and express it in numbers, you
know something about it; but when you cannot measure it, when you cannot express
it in numbers, your knowledge is of a meager and unsatisfactory kind: it may be
the beginning of knowledge, but you have scarcely, in your thoughts, advanced to the
stage of science.”––William Thomson, Lord Kelvin, Popular Lectures and Addresses
(1891–94)2

1

Lord Kelvin, a British physicist and the manager of the laying of the first successful
transatlantic cable, was one of history’s great advocates for quantitative investigation. In
his day, however, mathematical rigor had not yet spread widely beyond the worlds of
science, engineering, and finance. Much has changed since then.
Today, numerical fluency is a crucial skill for every business leader. Managers must
quantify market opportunities and competitive threats. They must justify the financial
risks and benefits of their decisions. They must evaluate plans, explain variances, judge
performance, and identify leverage points for improvement––all in numeric terms.
These responsibilities require a strong command of measurements and of the systems
and formulas that generate them. In short, they require metrics.
Managers must select, calculate, and explain key business metrics. They must under-
stand how each is constructed and how to use it in decision-making. Witness the fol-
lowing, more recent quotes from management experts:

“. . . every metric, whether it is used explicitly to influence behavior, to evaluate future
strategies, or simply to take stock, will affect actions and decisions.” 3
“If you can’t measure it, you can’t manage it.”4

1.3 Marketing Metrics: Opportunities, Performance,
and Accountability
Marketers are by no means immune to the drive toward quantitative planning and eval-
uation. Marketing may once have been regarded as more an art than a science. Executives
may once have cheerfully admitted that they knew they wasted half the money they spent
on advertising, but they didn’t know which half. Those days, however, are gone.
Today, marketers must understand their addressable markets quantitatively. They must
measure new opportunities and the investment needed to realize them. Marketers
must quantify the value of products, customers, and distribution channels––all under
various pricing and promotional scenarios. Increasingly, marketers are held accountable
for the financial ramifications of their decisions. Observers have noted this trend in
graphic terms:

“For years, corporate marketers have walked into budget meetings like neighborhood
junkies. They couldn’t always justify how well they spent past handouts or what
difference it all made. They just wanted more money––for flashy TV ads, for big-ticket
events, for, you know, getting out the message and building up the brand. But those
heady days of blind budget increases are fast being replaced with a new mantra:
measurement and accountability.”5

2 MARKETING METRICS

1.4 Choosing the Right Numbers
The numeric imperative represents a challenge, however. In business and economics,
many metrics are complex and difficult to master. Some are highly specialized and
best suited to specific analyses. Many require data that may be approximate, incomplete,
or unavailable.
Under these circumstances, no single metric is likely to be perfect. For this reason, we
recommend that marketers use a portfolio or “dashboard” of metrics. By doing so, they
can view market dynamics from various perspectives and arrive at “triangulated” strate-
gies and solutions. Additionally, with multiple metrics, marketers can use each as a
check on the others. In this way, they can maximize the accuracy of their knowledge.6
They can also estimate or project one data point on the basis of others. Of course, to use
multiple metrics effectively, marketers must appreciate the relations between them and
the limitations inherent in each.
When this understanding is achieved, however, metrics can help a firm maintain a
productive focus on customers and markets. They can help managers identify the
strengths and weaknesses in both strategies and execution. Mathematically defined
and widely disseminated, metrics can become part of a precise, operational language
within a firm.

Data Availability and Globalization of Metrics
A further challenge in metrics stems from wide variations in the availability of
data between industries and geographies. Recognizing these variations, we have tried to
suggest alternative sources and procedures for estimating some of the metrics in
this book.
Fortunately, although both the range and type of marketing metrics may vary between
countries,7 these differences are shrinking rapidly. Ambler,8 for example, reports that
performance metrics have become a common language among marketers, and that they
are now used to rally teams and benchmark efforts internationally.

1.5 Mastering Metrics
Being able to “crunch the numbers” is vital to success in marketing. Knowing which
numbers to crunch, however, is a skill that develops over time. Toward that end, man-
agers must practice the use of metrics and learn from their mistakes. By working
through the examples in this book, we hope our readers will gain both confidence and
a firm understanding of the fundamentals of data-based marketing. With time and

Chapter 1 Introduction 3

experience, we trust that you will also develop an intuition about metrics, and learn to
dig deeper when calculations appear suspect or puzzling.
Ultimately, with regard to metrics, we believe many of our readers will require not
only familiarity but also fluency. That is, managers should be able to perform relevant
calculations on the fly––under pressure, in board meetings, and during strategic
deliberations and negotiations. Although not all readers will require that level of
fluency, we believe it will be increasingly expected of candidates for senior manage-
ment positions, especially those with significant financial responsibility. We
anticipate that a mastery of data-based marketing will become a means for many of our
readers to differentiate and position themselves for career advancement in an ever more
challenging environment.

Organization of the Text
This book is organized into chapters that correspond to the various roles played by mar-
keting metrics in enterprise management. Individual chapters are dedicated to metrics
used in promotional strategy, advertising, and distribution, for example. Each chapter is
composed of sections devoted to specific concepts and calculations.
We must present these metrics in a sequence that will appear somewhat arbitrary. In
organizing this text, we have sought to strike a balance between two goals: (1) to estab-
lish core concepts first and build gradually toward increasing sophistication, and (2) to
group related metrics in clusters, helping our readers recognize patterns of mutual rein-
forcement and interdependence. In Figure 1.1, we offer a graphical presentation of this
structure, demonstrating the interlocking nature of all marketing metrics––indeed of all
marketing programs––as well as the central role of the customer.
The central issues addressed by the metrics in this book are as follows:
■ Chapter 2––Share of Hearts, Minds, and Markets: Customer perceptions, market
share, and competitive analysis.
■ Chapter 3––Margins and Profits: Revenues, cost structures, and profitability.
■ Chapter 4––Product and Portfolio Management: The metrics behind product
strategy, including measures of trial, growth, cannibalization, and brand equity.
■ Chapter 5––Customer Profitability: The value of individual customers and
relationships.
■ Chapter 6––Sales Force and Channel Management: Sales force organization,
performance, and compensation. Distribution coverage and logistics.
■ Chapter 7––Pricing Strategy: Price sensitivity and optimization, with an eye
toward setting prices to maximize profits.

4 MARKETING METRICS

Customers and Market Research

Logistics
Operations
Product and
Customer
Portfolio
Profitability
Management Sales Force

Sales Force
Margins and and
Profits Channel
Management
Share of Hearts, Minds,
and Markets

Marketing
Pricing
and
Strategy
Finance
Finance

Advertising
Media and Web Promotions
Metrics

The Trade

Advertising
Agency

Figure 1.1 Marketing Metrics: Marketing at the Core of the Organization

■ Chapter 8––Promotion: Temporary price promotions, coupons, rebates, and
trade allowances.
■ Chapter 9––Advertising Media and Web Metrics: The central measures of adver-
tising coverage and effectiveness, including reach, frequency, rating points, and
impressions. Models for consumer response to advertising. Specialized metrics
for Web-based campaigns.
■ Chapter 10––Marketing and Finance: Financial evaluation of marketing programs.
■ Chapter 11––The Marketing Metrics X-Ray: The use of metrics as leading indi-
cators of opportunities, challenges, and financial performance.
■ Chapter 12—System of Metrics: Decomposing marketing metrics into compo-
nent parts can improve measurement accuracy, add managerial insight into
problems, and assist marketing model building.


Components of Each Chapter
As shown in Table 1.1, the chapters are composed of multiple sections, each dedicated to
specific marketing concepts or metrics. Within each section, we open with definitions,
formulas, and a brief description of the metrics covered. Next, in a passage titled
Construction, we explore the issues surrounding these metrics, including their formu-
lation, application, interpretation, and strategic ramifications. We provide examples to
illustrate calculations, reinforce concepts, and help readers verify their understanding of
key formulas. That done, in a passage titled Data Sources, Complications, and Cautions,
we probe the limitations of the metrics under consideration and potential pitfalls in
their use. Toward that end, we also examine the assumptions underlying these metrics.
Finally, we close each section with a brief survey of Related Metrics and Concepts.
In organizing the text in this way, our goal is straightforward: Most of the metrics in this
book have broad implications and multiple layers of interpretation. Doctoral theses
could be devoted to many of them, and have been written about some. In this book,
however, we want to offer an accessible, practical reference. If the devil is in the details,
we want to identify, locate, and warn readers against him, but not to elaborate his entire
demonology. Consequently, we discuss each metric in stages, working progressively
toward increasing levels of sophistication. We invite our readers to sample this informa-
tion as they see fit, exploring each metric to the depth that they find most useful
and rewarding.
With an eye toward accessibility, we have also avoided advanced mathematical notation.
Most of the calculations in this book can be performed by hand, on the back of the
proverbial envelope. More complex or intensive computations may require a spread-
sheet. Nothing further should be needed.

Reference Materials
Throughout this text, we have highlighted formulas and definitions for easy reference.
We have also included outlines of key terms at the beginning of each chapter and
section. Within each formula, we have followed this notation to define all inputs
and outputs.

$—(Dollar Terms): A monetary value. We have used the dollar sign and “dollar
terms” for brevity, but any other currency, including the euro, yen, dinar, or yuan,
would be equally appropriate.
%—(Percentage): Used as the equivalent of fractions or decimals. For readability,
we have intentionally omitted the step of multiplying decimals by 100 to obtain
percentages.

6 MARKETING METRICS

#––(Count): Used for such measures as unit sales or number of competitors.
R––(Rating): Expressed on a scale that translates qualitative judgments or prefer-
ences into numeric ratings. Example: A survey in which customers are asked to
assign a rating of “1” to items that they find least satisfactory and “5” to those that
are most satisfactory. Ratings have no intrinsic meaning without reference to their
scale and context.
I––(Index): A comparative figure, often linked to or expressive of a market
average. Example: the consumer price index. Indexes are often interpreted as
a percentage.

$––Dollar. %––Percentage. #––Count. R––Rating. I––Index.

References and Suggested Further Reading
Abela, Andrew, Bruce H. Clark, and Tim Ambler. “Marketing Performance Measurement,
Performance, and Learning,” working paper, September 1, 2004.
Ambler, Tim, and Chris Styles. (1995). “Brand Equity: Toward Measures That Matter,” working
paper No. 95-902, London Business School, Centre for Marketing.
Barwise, Patrick, and John U. Farley. (2003). “Which Marketing Metrics Are Used and Where?”
Marketing Science Institute, (03-111), working paper, Series issues two 03-002.
Clark, Bruce H., Andrew V. Abela, and Tim Ambler. “Return on Measurement: Relating
Marketing Metrics Practices to Strategic Performance,” working paper, January 12, 2004.
Hauser, John, and Gerald Katz. (1998). “Metrics: You Are What You Measure,” European
Management Journal, Vo. 16, No. 5, pp. 517–528.
Kaplan, R. S., and D. P. Norton. (1996). The Balanced Scorecard: Translating Strategy into Action,
Boston, MA: Harvard Business School Press.


Table 1.1 Major Metrics List

Section Metric Section Metric

Share of Hearts, Minds, and Markets 3.2 Channel Margins
2.1 Revenue Market Share 3.3 Average Price per Unit
2.1 Unit Market Share 3.3 Price Per Statistical Unit
2.2 Relative Market Share 3.4 Variable and Fixed Costs
2.3 Brand Development Index 3.5 Marketing Spending
2.3 Category Development 3.6 Contribution per Unit
Index 3.6 Contribution Margin (%)
2.4–2.6 Decomposition of Market 3.6 Break-Even Sales
Share
3.7 Target Volume
2.4 Market Penetration
3.7 Target Revenues
2.4 Brand Penetration
2.4 Penetration Share Product and Portfolio Management
2.5 Share of Requirements 4.1 Trial
2.6 Heavy Usage Index 4.1 Repeat Volume
2.7 Hierarchy of Effects 4.1 Penetration
2.7 Awareness 4.1 Volume Projections
2.7 Top of Mind 4.2 Year-on-Year Growth
2.7 Ad Awareness 4.2 Compound Annual Growth
Rate (CAGR)
2.7 Knowledge
4.3 Cannibalization Rate
2.7 Consumer Beliefs
4.3 Fair Share Draw Rate
2.7 Purchase Intentions
4.4 Brand Equity Metrics
2.7 Purchase Habits
4.5 Conjoint Utilities
2.7 Loyalty
4.6 Segment Utilities
2.7 Likeability
4.7 Conjoint Utilities and
2.8 Willingness to Recommend
Volume Projections
2.8 Customer Satisfaction
2.9 Net Promoter Customer Profitability

2.10 Willingness to Search 5.1 Customers
5.1 Recency
Margins and Profits
5.1 Retention Rate
3.1 Unit Margin
5.2 Customer Profit
3.1 Margin (%)
5.3 Customer Lifetime Value

8 MARKETING METRICS

Table 1.1 Continued


5.4 Prospect Lifetime Value 7.3 Price Elasticity of Demand
5.5 Average Acquisition Cost 7.4 Optimal Price
5.5 Average Retention Cost 7.5 Residual Elasticity

Sales Force and Channel Management Promotion
6.1 Workload 8.1 Baseline Sales
6.1 Sales Potential Forecast 8.1 Incremental
6.2 Sales Goal Sales/Promotion Lift

6.3 Sales Force Effectiveness 8.2 Redemption Rates

6.4 Compensation 8.2 Costs for Coupons and
Rebates
6.4 Break-Even Number of
Employees 8.2 Percentage Sales with
Coupon
6.5 Sales Funnel, Sales Pipeline
8.3 Percent Sales on Deal
6.6 Numeric Distribution
8.3 Pass-Through
6.6 All Commodity Volume
(ACV) 8.4 Price Waterfall

6.6 Product Category Volume Advertising Media and Web Metrics
(PCV)
9.1 Impressions
6.6 Total Distribution
9.1 Gross Rating Points (GRPs)
6.6 Category Performance Ratio
9.2 Cost per Thousand
6.7 Out of Stock Impressions (CPM)
6.7 Inventories 9.3 Net Reach
6.8 Markdowns 9.3 Average Frequency
6.8 Direct Product Profitability 9.4 Frequency Response
(DPP) Functions
6.8 Gross Margin Return on 9.5 Effective Reach
Inventory Investment
9.5 Effective Frequency
(GMROII)
9.6 Share of Voice
Pricing Strategy 9.7 Pageviews
7.1 Price Premium 9.8 Rich Media Display Time
7.2 Reservation Price
7.2 Percent Good Value

Continues


Table 1.1 Continued


9.9 Rich Media Interaction Rate Marketing and Finance
9.10 Clickthrough Rate 10.1 Net Profit
9.11 Cost per Click 10.1 Return on Sales (ROS)
9.11 Cost per Order 10.1 Earnings Before Interest,
9.11 Cost per Customer Acquired Taxes, Depreciation, and
Amortization (EBITDA)
9.12 Visits
10.2 Return on Investment (ROI)
9.12 Visitors
10.3 Economic Profit (aka EVA®)
9.12 Abandonment Rate
10.4 Payback
9.13 Bounce Rate
10.4 Net Present Value (NPV)
9.14 Friends/Followers/Supporters
10.4 Internal Rate of Return (IRR)
9.15 Downloads
10.5 Return on Marketing
Investment (ROMI); Revenue

1.6 Marketing Metrics Survey
Why Do a Survey of Which Metrics Are Most Useful?
From the beginning of our work on this book, we have fielded requests from colleagues,
editors, and others to provide a short list of the “key” or “top ten” marketing metrics.
The intuition behind this request is that readers (managers and students) ought to be
able to focus their attention on the “most important” metrics. Until now we have resis-
ted that request.
Our reasons for not providing the smaller, more concentrated list of “really important”
metrics are as follows. First, we believe that any ranking of marketing metrics from most
to least useful will depend on the type of business under consideration. For example,
marketers of business-to-business products and services that go to market through
a direct sales force don’t need metrics that measure retail availability or dealer pro-
ductivity.
The second reason we believe that different businesses will have different rankings is
that metrics tend to come in matched sets. For example, if customer lifetime value is
important to your business (let’s say, financial services), then you are likely to value

10 MARKETING METRICS

measures of retention and acquisition costs as well. The same notion applies to retail,
media, sales force, and Web traffic metrics. If some of these are important to you, oth-
ers in the same general categories are likely to be rated as useful, too.
Third, businesses don’t always have access (at a reasonable cost) to the metrics they
would like to have. Inevitably, some of the rankings presented will reflect the cost of
obtaining the data that underlie the particular metrics.
Fourth, some metrics might be ranked lower, but ultimately prove to be useful, after
managers fully understand the pros and cons of a particular metric. For example, many
believe that Economic Value Added (EVA) is the “gold standard” of profitability metrics,
but it ranks far below other financial performance measures such as ROI. We believe
one reason for the low ranking of EVA is that this metric is less applicable at the “oper-
ating level” than for overall corporate performance. The other reason is that the meas-
ure is relatively new, and many managers don’t understand it as well. Customer Lifetime
Value is another metric that is gaining acceptance, but is still unfamiliar to many man-
agers. If all these metrics were well understood, there would be no need for a book of
this type.
In summary, while we believe the rankings resulting from our survey can be useful, we
ask readers to keep the above points in mind. We report in Tables 1.2 (page 13) and 1.3
(page 21) the overall ranking of the usefulness of various metrics as well as the different
rankings for different types of businesses and different categories of metrics. Although
no business is likely to be exactly like yours, we thought readers might find it useful to
see what other marketers thought which metrics were most useful in monitoring and
managing their businesses. For a look at the complete survey, see Appendix A.

Survey Sample
Our survey was completed by 194 senior marketing managers and executives. More than
100 held the title of Vice President/Director/Manager or “Head” of Marketing, some
with global responsibility. Most held titles such as VP of Marketing, Marketing Director,
and Director Sales and Marketing. There were 10 presidents and C-level managers with
heavy marketing responsibilities, and the remaining respondents included product/
project/category managers, trade marketing managers, pricing managers, key account
managers, development managers, and assistant/associate vice presidents.
Industries represented in our survey are too diverse to easily summarize. No more than
10 responses from a single industry were recorded, and the respondents listed their mar-
kets as aerospace, automobiles, banking, chemicals, consumer goods, construction,
computers, consulting, education, industrial distribution, investments, government,
health care, housing, insurance, information technology, manufacturing, materials,
medical devices, paints, pharmaceuticals, retailing, software, telecommunications, and
transportation. Roughly 20% of respondents did not provide a specific industry.


Survey questions asked respondents to rate the usefulness of particular metrics in mon-
itoring and managing their businesses. Note that this survey asks managers to give rat-
ings with respect to how these metrics are actually used but does not inquire about the
reason. Nor did the survey offer guidance concerning the meaning of “useful”—that was
left as a matter of interpretation for survey participants.
Financial metrics are generally rated very high in usefulness compared to any true mar-
keting metrics. This is not surprising given that financial metrics are common to almost
every business.


Table 1.2 Survey of Senior Marketing Managers on the Perceived
Usefulness of Various Marketing Metrics (n = 194)

All Who What Does Who Are
Responded to Customer Your Business Your
Question Relationship Sell? Customers?

Consumers
Infrequent
Purchase
Frequent

Purchase
Contract

Products

Business
Services

Mixed

Mixed
End
Group

# of People in Group 194 65 69 41 105 36 31 44 85 48

Chapter % Saying
Question in Very
Metric Number Book Useful Rank Rank Rank Rank Rank Rank Rank Rank Rank Rank
Net Profit Q8.10#1 10 91% 1 1 1 1 1 1 1 1 1 1
Margin % Q8.3#2 3 78% 2 10 2 3 2 6 2 2 3 6
Return on Investment Q8.10#3 10 77% 3 4 5 2 3 5 3 3 2 8
Customer Satisfaction Q8.2#12 2 71% 4 2 17 11 13 3 5 19 6 4
Chapter 1 Introduction

Target Revenues Q8.4#2 3 71% 5 8 12 5 12 8 3 13 7 6
Sales Total Q8.6#3 6 70% 6 7 10 8 10 8 8 16 3 12
Target Volumes Q8.4#1 3 70% 7 5 6 11 8 13 10 8 7 10
Return on Sales Q8.10#2 10 69% 8 12 12 3 9 17 8 4 17 2
Loyalty Q8.2#8 2 69% 9 70 71 98 4 11 17 13 5 16
Annual Growth % Q8.4#7 4 69% 10 13 3 11 7 11 15 8 10 10
Dollar Market Share Q8.1#1 2 67% 11 13 7 7 5 13 21 8 11 13
13

Continues

From the Library of Ross Hagglun

14

Table 1.2 Continued

MARKETING METRICS


Consumers
Infrequent
Purchase
Frequent

Purchase
Contract

Products

Business
Services

Mixed

Mixed
End
Group


Chapter % Saying
Question in Very
Customers Q8.5#1 5 67% 12 5 16 11 19 4 5 26 13 3
Unit Margin Q8.3#1 3 65% 13 17 9 5 11 21 10 13 12 13
Retention Rate Q8.5#3 5 63% 14 3 26 26 28 2 5 76 9 5
Sales Potential Forecast Q8.6#2 6 62% 15 11 18 11 17 18 10 23 14 18
Unit Market Share Q8.1#2 2 61% 16 23 4 16 5 54 30 8 18 17
Brand Awareness Q8.2#1 2 61% 17 23 7 16 14 33 10 4 25 9
Variable and Fixed Costs Q8.3#6 3 60% 18 15 11 32 15 8 30 19 21 13
Willingness to
Recommend Q8.2#10 2 57% 19 9 32 26 30 6 19 36 16 29
Volume Projections Q8.4#6 4 56% 20 23 14 21 16 31 24 45 15 27
Sales Force Effective Q8.6#4 6 54% 21 21 22 21 25 31 15 42 23 18
Price Premium Q8.8#1 7 54% 22 28 27 8 23 33 17 56 19 25


Marketing Spending Q8.3#7 3 52% 23 51 15 16 18 67 21 6 46 21
Average Price per Unit Q8.3#4 3 51% 24 23 23 32 21 33 38 27 26 25
Penetration Q8.4#5 4 50% 25 39 19 21 22 54 24 39 24 32
Top of Mind Q8.2#2 2 50% 26 33 25 26 30 33 30 39 27 21
Compensation Q8.6#5 6 49% 27 17 30 52 32 18 46 42 20 58
Return on Marketing
Investment (ROMI) Q8.10#8 10 49% 27 47 32 8 26 45 24 19 39 24
Consumer Beliefs Q8.2#5 2 48% 29 33 35 21 47 21 10 30 29 36
Contribution Margin % Q8.3#9 3 47% 30 56 21 21 29 46 24 45 32 21
Net Present Value Q8.10#6 10 46% 31 31 37 26 39 27 20 39 41 20
Market Penetration Q8.1#6 2 45% 32 17 41 58 38 41 38 45 35 33
Sales Funnel, Sales Q8.6#7 6 44% 33 17 60 32 54 21 21 74 21 58
Pipeline
Relative Market Share Q8.1#3 2 44% 34 36 38 40 32 33 65 58 41 27
Purchase Habits Q8.2#7 2 43% 35 39 35 43 27 41 80 30 29 69
Inventories Q8.7#7 6 43% 36 62 20 48 20 109 59 24 45 46

Likeability Q8.2#9 2 43% 37 28 54 38 47 21 46 45 37 39
Effective Reach Q8.9#6 9 42% 38 48 40 32 37 46 44 7 61 46
Economic Profit (EVA) Q8.10#4 10 41% 39 31 63 26 50 27 30 71 36 38
Impressions Q8.9#1 9 41% 40 36 61 26 50 41 24 19 64 29
Customer Profit Q8.5#4 5 41% 41 16 69 52 59 18 54 73 28 46
Optimal Price Q8.8#5 7 41% 42 39 47 36 36 46 46 45 49 36

Continues
15


16

Table 1.2 Continued

MARKETING METRICS


Consumers
Infrequent
Purchase
Frequent

Purchase
Contract

Products

Business
Services

Mixed

Mixed
End
Group


Chapter % Saying
Question in Very
Payback Q8.10#5 10 41% 42 51 51 20 54 27 43 67 34 44
Incremental Sales or Q8.8#8 8 41% 44 66 24 52 24 96 65 24 50 51
Promotional Lift
Consumer Knowledge Q8.2#4 2 40% 45 36 57 43 64 21 30 58 37 51
Contribution per Unit Q8.3#8 3 40% 46 71 29 48 39 62 46 63 54 29
Break-Even Sales Q8.3#10 3 40% 46 51 39 43 43 40 59 58 41 46
Customer Lifetime Value Q8.5#5 5 39% 48 23 77 40 69 21 30 76 46 33
Price Elasticity Q8.8#4 7 39% 48 71 31 38 35 72 54 34 56 39
Purchase Intentions Q8.2#6 2 39% 50 54 67 19 62 41 30 45 32 79
Growth CAGR Q8.4#8 4 38% 51 45 32 74 41 54 72 83 31 45
Internal Rate of Return Q8.10#7 10 38% 52 44 63 36 66 27 29 71 53 35
Effective Frequency Q8.9#7 9 37% 53 56 52 43 45 67 44 12 74 46


Visitors Q8.9#15 9 37% 54 39 58 58 60 46 38 53 51 62
Average Acquisition Cost Q8.5#7 5 36% 55 21 95 43 77 13 38 83 41 43
Share of Voice Q8.9#8 9 36% 55 66 43 52 45 62 64 33 72 39
Visits Q8.9#14 9 36% 57 39 58 66 61 46 38 53 55 51
Workload Q8.6#1 6 36% 58 50 48 66 53 54 59 79 40 58
Repeat Volume Q8.4#4 4 36% 59 56 46 58 50 54 65 64 52 58
Clickthrough Rate Q8.9#10 9 35% 60 33 61 77 63 33 54 29 67 51
Baseline Sales Q8.8#7 8 34% 61 71 42 56 42 72 80 45 56 69
Total Distribution Q8.7#4 6 34% 62 84 43 48 44 96 59 28 66 69
Net Reach Q8.9#4 9 34% 62 62 48 66 58 72 51 37 62 62
Brand Penetration Q8.1#7 2 34% 64 62 54 62 47 62 75 30 69 62
Out of Stock % Q8.7#6 6 33% 65 86 27 88 34 109 86 18 64 85
Average Retention Cost Q8.5#8 5 33% 66 30 98 40 82 13 51 91 48 51
Product Category Volume Q8.7#3 6 33% 67 84 45 57 57 92 58 62 62 51
Cost per Customer Q8.9#13 9 32% 68 48 72 66 70 54 51 74 60 51
Acquired

Average Frequency Q8.9#5 9 31% 69 76 48 71 54 83 75 16 77 86
Channel Margin Q8.3#3 3 30% 70 66 80 48 70 83 37 67 82 39
Direct Product Q8.7#9 6 30% 71 76 56 62 67 72 54 66 69 62
Profitability
Recency Q8.5#2 5 29% 72 56 74 71 75 33 80 94 59 62
Cost per Thousand Q8.9#3 9 28% 73 62 81 62 70 62 75 38 83 75
Impression
17

Continues


18

Table 1.2 Continued

MARKETING METRICS


Consumers
Infrequent
Purchase
Frequent

Purchase
Contract

Products

Business
Services

Mixed

Mixed
End
Group


Chapter % Saying
Question in Very
Pageview Q8.9#9 9 28% 74 45 84 88 87 54 46 56 83 69
Cost per Click Q8.9#11 9 27% 75 56 86 77 79 46 65 53 88 75
Brand Equity Metrics Q8.4#10 4 26% 76 76 76 77 68 72 89 58 90 74
Markdowns Q8.7#8 6 26% 77 96 52 84 65 106 80 34 90 86
Cannibalization Rate Q8.4#9 4 24% 78 88 65 95 74 83 97 78 76 91
Abandonment Rate Q8.9#16 9 24% 79 56 90 95 90 62 71 81 87 68
Ad Awareness Q8.2#3 2 23% 80 76 88 77 78 72 80 64 104 75
Cost per Order Q8.9#12 9 23% 81 71 91 74 90 67 65 95 73 75
Gross Rating Points Q8.9#2 9 23% 82 88 91 58 84 67 80 42 99 92
Break-Even Number Q8.6#6 6 23% 83 66 96 71 100 46 59 85 69 96
of Employees
Hierarchy of Effects Q8.1#11 2 23% 84 81 83 84 80 72 86 92 83 69


Numeric Distribution % Q8.7#1 6 22% 85 108 75 62 73 106 103 69 89 97
All Commodity Volume Q8.7#2 6 22% 85 96 67 93 75 83 89 69 78 104
Penetration Share Q8.1#8 2 22% 87 76 93 74 84 72 75 95 75 79
Brand Development Q8.1#4 2 21% 88 91 79 94 89 83 75 80 94 79
Index
Prospect Lifetime Value Q8.5#6 5 21% 89 81 106 66 95 46 104 98 67 97
Percentage Sales on Deal Q8.8#12 8 21% 89 91 82 87 92 83 72 87 79 92
Willingness to Search Q8.2#13 2 20% 91 71 102 77 86 72 107 85 79 100
Trial Volume Q8.4#3 4 19% 92 90 72 108 82 96 97 90 79 103
Net Promoter Score Q8.2#11 2 19% 93 55 101 103 94 61 107 106 58 109
Facings Q8.7#5 6 19% 94 99 66 105 81 72 107 45 99 110
Redemption Rates Q8.8#9 8 19% 95 102 69 100 92 96 104 82 94 92
Cost of Coupons/ Rebates Q8.8#10 8 19% 95 102 77 90 87 96 97 87 102 79
Category Development Q8.1#5 2 18% 97 95 87 103 97 83 86 99 92 79
Index
Reservation Price Q8.8#2 7 17% 98 99 93 84 96 72 89 100 86 99

GMROII Q8.7#10 6 16% 99 102 84 99 98 96 89 87 94 100
Percent Good Value Q8.8#3 7 16% 99 91 108 77 107 67 72 100 109 62
Percentage Sales with Q8.8#11 8 16% 99 109 88 90 98 96 89 93 105 86
Coupon
Price per Statistical Unit Q8.3#5 3 16% 102 91 102 90 104 83 65 104 94 79
Conjoint Utilities Q8.4#11 4 14% 103 81 99 108 101 92 89 107 94 89
Residual Elasticity Q8.8#6 7 14% 104 98 109 77 102 92 97 109 92 92
19

Continues


20
MARKETING METRICS

Table 1.2 Continued


Consumers
Infrequent
Purchase
Frequent

Purchase
Contract

Products

Business
Services

Mixed

Mixed
End
Group


Chapter % Saying
Question in Very
Percent Time on Deal Q8.8#13 8 14% 105 102 96 95 105 96 89 97 102 104
Conjoint Utilities &
Volume Projection Q8.4#12 4 13% 106 87 99 108 103 92 89 103 105 89

Pass-Through Q8.8#15 8 11% 107 102 107 100 108 83 97 102 108 100
Share of Requirements Q8.1#9 2 10% 108 102 102 105 106 106 106 108 99 108
Average Deal Depth Q8.8#14 8 10% 109 110 105 100 109 96 97 105 107 104
Heavy Usage Index Q8.1#10 2 6% 110 101 110 107 110 96 110 110 110 104


Table 1.3 Ranking of Metrics by Category/Chapter (See Appendix A for complete survey)

% Saying Ranking in
Section in Question Chapter in Very Survey
Metric Survey Number Book Useful Section
Dollar Market Share 1 Q8.1#1 2 67% 1
Unit Market Share 1 Q8.1#2 2 61% 2
Market Penetration 1 Q8.1#6 2 45% 3
Relative Market Share 1 Q8.1#3 2 44% 4
Brand Penetration 1 Q8.1#7 2 34% 5
Hierarchy of Effects 1 Q8.1#11 2 23% 6
Penetration Share 1 Q8.1#8 2 22% 7
Brand Development
Index 1 Q8.1#4 2 21% 8
Category Development
Index 1 Q8.1#5 2 18% 9
Share of Requirements 1 Q8.1#9 2 10% 10
Heavy Usage Index 1 Q8.1#10 2 6% 11
Customer Satisfaction 2 Q8.2#12 2 71% 1
Loyalty 2 Q8.2#8 2 69% 2
Brand Awareness 2 Q8.2#1 2 61% 3
Willingness to
Recommend 2 Q8.2#10 2 57% 4
Top of Mind 2 Q8.2#2 2 50% 5
Consumer Beliefs 2 Q8.2#5 2 48% 6
Purchase Habits 2 Q8.2#7 2 43% 7
Likeability 2 Q8.2#9 2 43% 8
Consumer Knowledge 2 Q8.2#4 2 40% 9
Purchase Intentions 2 Q8.2#6 2 39% 10
Ad Awareness 2 Q8.2#3 2 23% 11
Willingness to Search 2 Q8.2#13 2 20% 12
Net Promoter Score 2 Q8.2#11 2 19% 13
Margin % 3 Q8.3#2 3 78% 1
Unit Margin 3 Q8.3#1 3 65% 2
Variable and Fixed Costs 3 Q8.3#6 3 60% 3

Continues


Table 1.3 Continued

% Saying Ranking in
Marketing Spending 3 Q8.3#7 3 52% 4
Average Price per Unit 3 Q8.3#4 3 51% 5
Contribution Margin % 3 Q8.3#9 3 47% 6
Contribution per Unit 3 Q8.3#8 3 40% 7
Break-Even Sales 3 Q8.3#10 3 40% 8
Channel Margin 3 Q8.3#3 3 30% 9
Price per Statistical Unit 3 Q8.3#5 3 16% 10
Target Revenues 4 Q8.4#2 3 71% 1
Target Volumes 4 Q8.4#1 3 70% 2
Annual Growth % 4 Q8.4#7 4 69% 3
Volume Projections 4 Q8.4#6 4 56% 4
Penetration 4 Q8.4#5 4 50% 5
Growth CAGR 4 Q8.4#8 4 38% 6
Repeat Volume 4 Q8.4#4 4 36% 7
Brand Equity Metrics 4 Q8.4#10 4 26% 8
Cannibalization Rate 4 Q8.4#9 4 24% 9
Trial Volume 4 Q8.4#3 4 19% 10
Conjoint Utilities 4 Q8.4#11 4 14% 11
Conjoint Utilities & 4 Q8.4#12 4 13% 12
Volume Projection
Customers 5 Q8.5#1 5 67% 1
Retention Rate 5 Q8.5#3 5 63% 2
Customer Profit 5 Q8.5#4 5 41% 3
Customer Lifetime 5 Q8.5#5 5 39% 4
Value
Average Acquisition 5 Q8.5#7 5 36% 5
Cost
Average Retention Cost 5 Q8.5#8 5 33% 6
Recency 5 Q8.5#2 5 29% 7
Prospect Lifetime Value 5 Q8.5#6 5 21% 8
Sales Total 6 Q8.6#3 6 70% 1


% Saying Ranking in
Sales Potential Forecast 6 Q8.6#2 6 62% 2
Sales Force Effective 6 Q8.6#4 6 54% 3
Compensation 6 Q8.6#5 6 49% 4
Sales Funnel, Sales 6 Q8.6#7 6 44% 5
Pipeline
Workload 6 Q8.6#1 6 36% 6
Break-Even Number 6 Q8.6#6 6 23% 7
of Employees
Inventories 7 Q8.7#7 6 43% 1
Total Distribution 7 Q8.7#4 6 34% 2
Out of Stock % (OOS) 7 Q8.7#6 6 33% 3
Product Category 7 Q8.7#3 6 33% 4
Volume (PCV)
Direct Product 7 Q8.7#9 6 30% 5
Profitability (DPP)
Markdowns 7 Q8.7#8 6 26% 6
Numeric Distribution % 7 Q8.7#1 6 22% 7
All Commodity 7 Q8.7#2 6 22% 8
Volume (ACV)
Facings 7 Q8.7#5 6 19% 9
Gross Margin Return 7 Q8.7#10 6 16% 10
on Inventory
Investment
(GMROII)
Price Premium 8 Q8.8#1 7 54% 1
Optimal Price 8 Q8.8#5 7 41% 2
Incremental Sales 8 Q8.8#8 8 41% 3
or Promotional Lift
Price Elasticity 8 Q8.8#4 7 39% 4
Baseline Sales 8 Q8.8#7 8 34% 5
Percentage Sales 8 Q8.8#12 8 21% 6
on Deal
Redemption Rates 8 Q8.8#9 8 19% 7

Continues


Table 1.3 Continued

% Saying Ranking in
Cost of Coupons/ 8 Q8.8#10 8 19% 8
Rebates
Reservation Price 8 Q8.8#2 7 17% 9
Percent Good Value 8 Q8.8#3 7 16% 10
Percentage Sales with 8 Q8.8#11 8 16% 11
Coupon
Residual Elasticity 8 Q8.8#6 7 14% 12
Percent Time on Deal 8 Q8.8#13 8 14% 13
Pass-Through 8 Q8.8#15 8 11% 14
Average Deal Depth 8 Q8.8#14 8 10% 15
Effective Reach 9 Q8.9#6 9 42% 1
Impressions 9 Q8.9#1 9 41% 2
Effective Frequency 9 Q8.9#7 9 37% 3
Visitors 9 Q8.9#15 9 37% 4
Share of Voice 9 Q8.9#8 9 36% 5
Visits 9 Q8.9#14 9 36% 6
Clickthrough Rate 9 Q8.9#10 9 35% 7
Net Reach 9 Q8.9#4 9 34% 8
Cost per Customer 9 Q8.9#13 9 32% 9
Acquired
Average Frequency 9 Q8.9#5 9 31% 10
Cost per Thousand
Impression (CPM) 9 Q8.9#3 9 28% 11
Pageview 9 Q8.9#9 9 28% 12
Cost per Click (CPC) 9 Q8.9#11 9 27% 13
Abandonment Rate 9 Q8.9#16 9 24% 14
Cost per Order 9 Q8.9#12 9 23% 15
Gross Rating Points 9 Q8.9#2 9 23% 16
Net Profit 10 Q8.10#1 10 91% 1
Return on Investment 10 Q8.10#3 10 77% 2
(ROI)


% Saying Ranking in
Return on Sales (ROS) 10 Q8.10#2 10 69% 3
Return on Marketing 10 Q8.10#8 10 49% 4
Investment (ROMI)
Net Present Value (NPV) 10 Q8.10#6 10 46% 5
Economic Profit (EVA) 10 Q8.10#4 10 41% 6
Payback 10 Q8.10#5 10 41% 7
Internal Rate of Return 10 Q8.10#7 10 38% 8
(IRR)


2
SHARE OF HEARTS, MINDS,
AND MARKETS

Introduction

Key concepts covered in this chapter:
Market Share Heavy Usage Index
Relative Market Share Awareness, Attitudes, and Usage (AAU)
Market Concentration Customer Satisfaction
Brand Development Index (BDI) Willingness to Recommend
Category Development Index (CDI) Net Promoter
Penetration Willingness to Search
Share of Requirements

“As Wal-Mart aggressively rolls out more stores, it continues to capture an increasing
share of wallet. Three out of five consumers shopped for gifts at Wal-Mart this past
holiday season. U.S. households now buy, on average, 22% of their groceries at
Wal-Mart. A quarter of all shoppers indicate that they are spending more of their
clothing budget at Wal-Mart now compared with a year ago. These ShopperScape
findings lend credence to Retail Forward’s premise that Wal-Mart will continue to
push the boundaries of what consumers will allow it to be.”1

27

At first glance, market share appears to involve a relatively simple calculation: “us/
(us them).” But this raises a host of questions. Who, for example, are “they?” That is,
how broadly do we define our competitive universe? Which units are used? Where in the
value chain do we capture our information? What time frame will maximize our signal-
to-noise ratio? In a metric as important as market share, and in one as closely moni-
tored for changes and trends, the answers to such questions are crucial. In this chapter,
we will address them and also introduce key components of market share, including
penetration share, heavy usage index, and share of requirements.
Probing the dynamics behind market share, we’ll explore measures of awareness, atti-
tude, and usage––major factors in the decision-making process by which customers
select one brand over another. We’ll discuss customer satisfaction with products and
dealers, the quantification of which is growing in importance among marketing profes-
sionals. Finally, we’ll consider metrics measuring the depth of consumer preference and
satisfaction, including customers’ willingness to search if a brand is unavailable and
their disposition to recommend that brand to others. Increasingly, marketers rely on
these as leading indicators of future changes in share.

Metric Construction Considerations Purpose
2.1 Revenue Market Sales revenue as a Scope of market Measure of
Share percentage of definition. competitiveness.
market sales Channel level
revenue. analyzed. Before/
after discounts.
Time period
covered.
2.1 Unit Market Share Unit sales as a Scope of market Measure of
percentage of definition. competitiveness.
market unit sales. Channel level
analyzed. Time
period covered.
2.2 Relative Market Brand market Can use either Assesses
Share share divided by unit or revenue comparative
largest competi- shares. market strength.
tor’s market
share.



2.3 Brand Brand sales in a Can use either Regional or
Development specified segment, unit or revenue segment differ-
Index compared with sales. ences in brand
sales of that purchases and
brand in the consumption.
market as a
whole.
2.3 Category Category sales in Can use either Regional or
Development a specified seg- unit or revenue segment differ-
Index ment, compared sales. ences in category
with sales of that purchases and
category in the consumption.
market as a
whole.
2.4 Decomposition of Penetration Share Can be based on Calculation of
2.5 Market Share * Share of unit or revenue market share.
Requirements * shares. Time Competitive
2.6 Heavy Usage period covered. analysis.
Index. Historical
trends analysis.
Formulation
of marketing
objectives.
2.4 Market Purchasers of a Based on popula- Measures cate-
Penetration product category tion. Therefore, gory acceptance
as a percentage of unit/revenue by a defined pop-
total population. consideration not ulation. Useful in
relevant. tracking accept-
ance of new prod-
uct categories.
2.4 Brand Purchasers of a Based on popula- Measures brand
Penetration brand as a per- tion. Therefore, acceptance by
centage of total unit/revenue a defined
population. consideration not population.
relevant.

Continues

Chapter 2 Share of Hearts, Minds, and Markets 29

2.4 Penetration Share Brand penetration A component of Comparative
as a percentage the market share acceptance of
of market formula. brand within
penetration. category.
2.5 Share of Brand purchases Can use either Level of commit-
Requirements as a percentage of unit or revenue ment to a brand
total category shares. May rise by its existing
purchases by buy- even as sales customers.
ers of that brand. decline, leaving
only most loyal
customers.
2.6 Heavy Usage Category pur- Can use either Measures relative
Index chases by cus- unit or revenue usage of a
tomers of a sales. category by
brand, compared customers for a
with purchases specific brand.
in that category
by average cus-
tomers in the
category.

2.7 Hierarchy of Awareness; Strict sequence is Set marketing
Effects attitudes, beliefs; often violated and and advertising
importance; can be reversed. objectives.
intentions to try; Understand
buy; trial, repeat. progress in stages
of customer
decision process.
2.7 Awareness Percentage of Is this prompted Consideration of
total population or unprompted who has heard of
that is aware of a awareness? the brand.
brand.
2.7 Top of Mind First brand to May be subject Saliency of brand.
consider. to most recent
advertising or
experience.

2.7 Ad Awareness Percentage of May vary by One measure
total population schedule, reach, of advertising
that is aware and frequency of effects. May
of a brand’s advertising. indicate “stopping
advertising. power” of ads.


2.7 Knowledge Percentage of Not a formal Extent of
population with metric. Is this familiarity with
knowledge of prompted or product beyond
product, recollec- unprompted name recognition.
tion of its adver- knowledge?
tising.
2.7 Consumer Customers/ Customers/ Perception of
Beliefs consumers view consumers may brand by
of product, gener- hold beliefs with attribute.
ally captured via varying degrees of
survey responses, conviction.
often through
ratings on a
scale.
2.7 Purchase Probability of To estimate prob- Measures pre-
Intentions intention to ability of pur- shopping disposi-
purchase. chase, aggregate tion to purchase.
and analyze rat-
ings of stated
intentions (for
example, top two
boxes).
2.7 Purchase Habits Frequency of pur- May vary widely Helps identify
chase. Quantity among shopping heavy users.
typically pur- trips.
chased.
2.7 Loyalty Measures include “Loyalty” itself is Indication of base
share of require- not a formal met- future revenue
ments, willingness ric, but specific stream.
to pay premium, metrics measure
willingness to aspects of this
search. dynamic. New
product entries
may alter loyalty
levels.
2.7 Likeability Generally meas- Often believed to Shows overall
ured via ratings correlate with preference prior
across a number persuasion. to shopping.
of scales.

Continues


2.8 Willingness to Generally meas- Nonlinear in Shows strength of
Recommend ured via ratings impact. loyalty, potential
across a 1–5 scale. impact on others.

2.8 Customer Generally meas- Subject to Indicates
Satisfaction ured on a 1–5 response bias. likelihood of
scale, in which Captures views repurchase.
customers declare of current Reports of
their satisfaction customers, not dissatisfaction
with brand in lost customers. show aspects that
general or specific Satisfaction is a require improve-
attributes. function of ment to enhance
expectations. loyalty.
2.9 Net Promoter Percentage of cus- Requires a survey Some claim it to
tomers willing to of intentions. be the single
recommend to best metric for
others less the marketers.
percentage
unwilling to
recommend
the product or
service.
2.10 Willingness to Percentage of cus- Hard to capture. Indicates
Search tomers willing to importance of
delay purchases, distribution
change stores, or coverage.
reduce quantities
to avoid switching
brands.

2.1 Market Share
Market share is the percentage of a market (defined in terms of either units or rev-
enue) accounted for by a specific entity.
Unit Sales (#)
Unit Market Share (%)
Total Market Unit Sales (#)

Sales Revenue ($)
Revenue Market Share (%)
Total Market Revenue ($)


Marketers need to be able to translate sales targets into market share because this will
demonstrate whether forecasts are to be attained by growing with the market or by
capturing share from competitors. The latter will almost always be more difficult to
achieve. Market share is closely monitored for signs of change in the competitive
landscape, and it frequently drives strategic or tactical action.

Purpose: Key indicator of market competitiveness.
Market share is an indicator of how well a firm is doing against its competitors. This
metric, supplemented by changes in sales revenue, helps managers evaluate both pri-
mary and selective demand in their market. That is, it enables them to judge not only
total market growth or decline but also trends in customers’ selections among competi-
tors. Generally, sales growth resulting from primary demand (total market growth) is
less costly and more profitable than that achieved by capturing share from competitors.
Conversely, losses in market share can signal serious long-term problems that require
strategic adjustments. Firms with market shares below a certain level may not be viable.
Similarly, within a firm’s product line, market share trends for individual products are
considered early indicators of future opportunities or problems.

Construction
Market Share: The percentage of a market accounted for by a specific entity.
Unit Market Share: The units sold by a particular company as a percentage of
total market sales, measured in the same units.
Unit Sales (#)

This formula, of course, can be rearranged to derive either unit sales or total market unit
sales from the other two variables, as illustrated in the following:
Unit Sales (#) Unit Market Share (%) * Total Market Unit Sales (#)
Unit Sales (#)

Revenue Market Share: Revenue market share differs from unit market share in
that it reflects the prices at which goods are sold. In fact, a relatively simple way to
calculate relative price is to divide revenue market share by unit market share (see
Section 7.1).
Sales Revenue ($)
Total Market Sales Revenue ($)


As with the unit market share, this equation for revenue market share can be rearranged
to calculate either sales revenue or total market sales revenue from the other two variables.

Data Sources, Complications, and Cautions
Market definition is never a trivial exercise: If a firm defines its market too broadly, it
may dilute its focus. If it does so too narrowly, it will miss opportunities and allow
threats to emerge unseen. To avoid these pitfalls, as a first step in calculating market
share, managers are advised to define the served market in terms of unit sales or
revenues for a specific list of competitors, products, sales channels, geographic areas,
customers, and time periods. They might posit, for example, that “Among grocery
stores, we are the revenue market share leader in sales of frozen Italian food entrées in
the Northeastern U.S.”
Data parameters must be carefully defined: Although market share is likely the single
most important marketing metric, there is no generally acknowledged best method for
calculating it. This is unfortunate, as different methods may yield not only different
computations of market share at a given moment, but also widely divergent trends over
time. The reasons for these disparities include variations in the lenses through which
share is viewed (units versus dollars), where in the channel the measurements are taken
(shipments from manufacturers versus consumer purchases), market definition (scope
of the competitive universe), and measurement error. In the situation analysis that
underlies strategic decisions, managers must be able to understand and explain these
variations.
Competitive dynamics in the automobile industry, and at General Motors in particular,
illustrate the complexities involved in quantifying market share:

“With market share sliding in the first two months of the year, from 27.2% to
24.9%––the lowest level since a two-month strike shut the company down in 1998––GM
as a whole expects a net loss of $846 million the first quarter.”2
Reviewing this statement, drawn from Business Week in 2005, a marketing manager
might immediately pose a number of questions:
■ Do these figures represent unit (auto) or revenue (dollar) market shares?
■ Does this trend hold for both unit and revenue market shares at GM?
■ Was revenue market share calculated before or after rebates and discounts?
■ Do the underlying sales data reflect factory shipments, which relate directly to
the manufacturer’s current income statement, or sales to consumers, which are
buffered by dealer inventories?


■ Does the decline in market share translate to an equivalent percentage decrease
in sales, or has the total market size changed?
Managers must determine whether a stated market share is based on shipment data,
channel shipments, retail sales, customer surveys, or some other source. On occasion,
share figures may represent combinations of data (a firm’s actual shipments, for exam-
ple, set against survey estimates of competitors’ sales). If necessary, managers must also
adjust for differences in channels.
The time period measured will affect the signal-to-noise ratio: In analyzing short-
term market dynamics, such as the effects of a promotion or a recent price change, man-
agers may find it useful to measure market share over a brief period of time. Short-term
data, however, generally carry a low signal-to-noise ratio. By contrast, data covering a
longer time span will be more stable but may obscure important, recent changes in the
market. Applied more broadly, this principle also holds in aggregating geographic areas,
channel types, or customers. When choosing markets and time periods for analysis,
managers must optimize for the type of signal that is most important.
Potential bias in reported shares: One way to find data for market sizing is through
surveys of customer usage (see Section 2.7). In interpreting these data, however, man-
agers must bear in mind that shares based on reported (versus recorded) sales tend to be
biased toward well-known brands.

Related Metrics and Concepts
Served Market: That portion of the total market for which the firm competes. This
may exclude geographic regions or product types. In the airline industry, for exam-
ple, as of mid 2009, Ryan Air did not fly to the United States. Consequently, the
U.S. would not be considered part of its served market.

2.2 Relative Market Share and Market Concentration
Relative market share indexes a firm’s or a brand’s market share against that of its
leading competitor.
Brand’s Market Share ($,#)
Relative Market Share (I) (%)
Largest Competitor’s Market Share ($,#)
Market concentration, a related metric, measures the degree to which a comparatively
small number of firms accounts for a large proportion of the market.
These metrics are useful in comparing a firm’s or a brand’s relative position across
different markets and in evaluating the type and degree of competition in those
markets.


Purpose: To assess a firm’s or a brand’s success and its position
in the market.
A firm with a market share of 25% would be a powerful leader in many markets but a
distant “number two” in others. Relative market share offers a way to benchmark a firm’s
or a brand’s share against that of its largest competitor, enabling managers to compare
relative market positions across different product markets. Relative market share gains
some of its significance from studies––albeit controversial ones––suggesting that major
players in a market tend to be more profitable than their competitors. This metric was
further popularized by the Boston Consulting Group in its famous matrix of relative
share and market growth (see Figure 2.1).

High
Question Mark or
Star
Problem Child

Market
Growth

Cash Cow Dog
Low

High Relative Market Share Low

Figure 2.1 The BCG Matrix

In the BCG matrix, one axis represents relative market share––a surrogate for compet-
itive strength. The other represents market growth––a surrogate for potential. Along
each dimension, products are classified as high or low, placing them in one of four
quadrants. In the traditional interpretation of this matrix, products with high relative
market shares in growing markets are deemed stars, suggesting that they should be
supported with vigorous investment. The cash for that investment may be generated
by cash cows, products with high relative shares in low-growth markets. Problem child
products may have potential for future growth but hold weak competitive positions.
Finally, dogs have neither strong competitive position nor growth potential.


Construction
Brand’s Market Share ($,#)
Relative Market Share (I)
Largest Competitor’s Market Share ($,#)

Relative market share can also be calculated by dividing brand sales (#,$) by largest
competitor’s sales (#,$) because the common factor of total market sales (or revenue)
cancels out.

EXAMPLE: The market for small urban cars consists of five players (see Table 2.1).

Table 2.1 Market for Small Urban Cars

Units Sold (Thousands) Revenue (Thousands)
Zipper 25 €375,000
Twister 10.0 €200,000
A-One 7.5 €187,500
Bowlz 5 €125,000
Chien 2.5 €50,000
Market Total 50.0 €937,500

In the market for small urban cars, managers at A-One want to know their firm’s market
share relative to its largest competitor. They can calculate this on the basis of revenues or
unit sales.
In unit terms, A-One sells 7,500 cars per year. Zipper, the market leader, sells 25,000.
A-One’s relative market share in unit terms is thus 7,500/25,000 or 0.30. We arrive at the
same number if we first calculate A-One’s share (7,500/50,000 = .15) and Zipper’s share
(25,000/50,000 .50) and then divide A-One’s share by Zipper’s share (.15/.50 = .30).
In revenue terms, A-One generates €187.5 million in car sales each year. Zipper, the mar-
ket leader, generates €375 million. A-One’s relative market share in revenue terms is thus
€187.5m/€375m, or 0.5. Due to its comparatively high average price per car, A-One’s rel-
ative market share is greater in revenue than in unit terms.


Market Concentration: The degree to which a relatively small number of firms
accounts for a large proportion of the market. This is also known as the concentra-
tion ratio. It is usually calculated for the largest three or four firms in a market.3
Three (Four) Firm Concentration Ratio: The total (sum) of the market shares
held by the leading three (four) competitors in a market.

EXAMPLE: In the small urban car market, the three firm concentration ratio is
comprised of the market shares of the top three competitors—Zipper, Twister, and
A-One (see Table 2.2).

Table 2.2 Market Share––Small Urban Cars

Units Sold Revenue
(Thousands) Unit Share (Thousands) Revenue Share
Zipper 25.0 50% €375,000 40.0%
Twister 10.0 20% €200,000 21.3%
A-One 7.5 15% €187,500 20.0%
Bowlz 5.0 10% €125,000 13.3%
Chien 2.5 5% €50,000 5.3%
Market Total 50.0 100% €937,500 100%

In unit terms, the three firm concentration ratio is 50% 20% 15% 85%.
In revenue terms, it is 40% 21.3% 20% 81.3%.

Herfindahl Index: A market concentration metric derived by adding the squares
of the individual market shares of all the players in a market. As a sum of squares,
this index tends to rise in markets dominated by large players.

EXAMPLE: The Herfindahl Index dramatically highlights market concentration in
the small urban car market (see Table 2.3).


Table 2.3 Calculation of the Herfindahl Index for Small Urban Cars

Units Sold Herfindahl Revenue Revenue Herfindahl
(Thousands) Unit Share Index (Thousands) Share Index
Zipper 25.0 50% 0.25 €375,000 40% 0.16
Twister 10.0 20% 0.04 €200,000 21% 0.0455
A-One 7.5 15% 0.0225 €187,500 20% 0.04
Bowlz 5.0 10% 0.01 €125,000 13% 0.0178
Chien 2.5 5% 0.0025 €50,000 5% 0.0028
Market Total 50.0 100% 0.325 €937,500 100% 0.2661

On a unit basis, the Herfindahl Index is equal to the square of the unit market share of
Zipper (50% ^ 2 = 0.25), plus that of Twister (20% ^ 2 = 0.04), plus those of A-One,
Bowlz, and Chien = 0.325.
On a revenue basis, the Herfindahl Index comprises the square of the revenue market
share of Zipper (40% ^ 2 0.16), plus those of all its competitors 0.2661.
As demonstrated by the Herfindahl Index, the market for small urban cars is slightly
more concentrated in unit terms than in revenue terms. The reason for this is straightfor-
ward: Higher-priced cars in this market sell fewer units.
Note: For a given number of competitors, the Herfindahl Index would be lowest if shares
were equally distributed. In a five-firm industry, for example, equally distributed shares
would yield a Herfindahl Index of 5 * (20% ^ 2) 0.2.

As ever, appropriate market definition and the use of comparable figures are vital pre-
requisites to developing meaningful results.

Market Share Rank: The ordinal position of a brand in its market, when competi-
tors are arranged by size, with 1 being the largest.
Share of Category: This metric is derived in the same manner as market share,
but is used to denote a share of market within a certain retailer or class of retailers
(for example, mass merchandisers).


2.3 Brand Development Index and Category
Development Index

The brand development index (BDI) quantifies how well a brand is performing
within a specific group of customers, compared with its average performance among
all consumers.
[Brand Sales to Group (#)/Households (#)
in the Group]
Brand Development Index (I)
[Total Brand Sales (#)/Total Household (#)]

The category development index (CDI) measures the sales performance of a category
of goods or services within a specific group, compared with its average performance
among all consumers.
[Category Sales to Group (#)/Households in Group (#)]
Category Development Index (I)
[Total Category Sales (#)/Total Household (#)]

The brand and category development indexes are useful for understanding specific
customer segments relative to the market as a whole. Although defined here with
respect to households, these indexes could also be calculated for customers, accounts,
businesses, or other entities.

Purpose: To understand the relative performance of a brand
or category within specified customer groups.
The brand and category development indexes help identify strong and weak segments
(usually, demographic or geographic) for particular brands or categories of goods and
services. For example, by monitoring the CDI (category development index), marketers
might determine that Midwesterners buy twice as many country-western music CDs
per capita as Americans in general, while consumers living on the East Coast buy less
than the national average. This would be useful information for targeting the launch
campaign for a new country-western performer. Conversely, if managers found that a
particular product had a low brand development index in a segment that carried a high
CDI for its category, they might ask why that brand suffered relatively poor perform-
ance in such a promising segment.


Construction
Brand Development Index—BDI (I): An index of how well a brand performs
within a given market group, relative to its performance in the market as
a whole.
[Brand Sales to Group (#)/Households in Group (#)]
Brand Development Index––BDI (I)
[Total Brand Sales (#)/Total Household (#)]

The BDI (brand development index) is a measure of brand sales per person or per
household within a specified demographic group or geography, compared with its
average sales per person or household in the market as a whole. To illustrate its use:
One might hypothesize that sales per capita of Ben & Jerry’s brand ice cream would be
greater in the brand’s home state, Vermont, than in the rest of the country. By
calculating Ben & Jerry’s BDI for Vermont, marketers could test this hypothesis
quantitatively.

EXAMPLE: Oaties is a minor brand of breakfast cereal. Among households without
children, its sales run one packet per week per 100 households. In the general population,
Oaties’ sales run one packet per week per 80 households. This translates to 1/100 of a
packet per household in the childless segment, versus 1/80 of a packet in the general
populace.
(Brand Sales/Household)
BDI =
(Total Brand Sales/Household)

1/100
= = 0.8
1/80

Oaties performs slightly less well in the childless segment than in the market as a whole.

Category Development Index—CDI: An index of how well a category performs
within a given market segment, relative to its performance in the market as a whole.

[Category Sales to Group (#)/Households in Group (#)]
[Total Category Sales (#)/Total Household (#)]

Similar in concept to the BDI, the category development index demonstrates where a
category shows strength or weakness relative to its overall performance. By way of
example, Boston enjoys high per-capita consumption of ice cream. Bavaria and Ireland
both show higher per-capita consumption of beer than Iran.


Data Sources and Complications
In calculating BDI or CDI, a precise definition of the segment under study is vital.
Segments are often bounded geographically, but they can be defined in any way for
which data can be obtained.

The term category development index has also been applied to retail organizations. In
this application, it measures the extent to which a retailer emphasizes one category
versus others.

Retailer’s Share of Category Sales (%)
Retailer’s Total Share of Market (%)

This use of the term is similar to the category performance ratio (see Section 6.6).

2.4 Penetration
Penetration is a measure of brand or category popularity. It is defined as the number
of people who buy a specific brand or a category of goods at least once in a given
period, divided by the size of the relevant market population.

Customers Who Have Purchased a Product in the Category (#)
Market Penetration (%)
Total Population (#)

Customers Who Have Purchased the Brand (#)
Brand Penetration (%)

Penetration Share (%)


Often, managers must decide whether to seek sales growth by acquiring existing cate-
gory users from their competitors or by expanding the total population of category
users, attracting new customers to the market. Penetration metrics help indicate
which of these strategies would be most appropriate and help managers to monitor
their success. These equations might also be calculated for usage instead of purchase.


Construction
Penetration: The proportion of people in the target who bought (at least once in
the period) a specific brand or a category of goods.


Two key measures of a product’s “popularity” are penetration rate and penetration
share. The penetration rate (also called penetration, brand penetration, or market pen-
etration as appropriate), is the percentage of the relevant population that has purchased
a given brand or category at least once in the time period under study.

EXAMPLE: Over a period of a month, in a market of 10,000 households, 500 house-
holds purchased Big Bomb brand flea foggers.
Big Bomb Customers
Brand Penetration, Big Bomb =
Total Population

500
= 5%
10,000

A brand’s penetration share, in contrast to penetration rate, is determined by compar-
ing that brand’s customer population to the number of customers for its category in the
relevant market as a whole. Here again, to be considered a customer, one must have
purchased the brand or category at least once during the period.

EXAMPLE: Returning to the flea fogger market, during the month in which 500
households purchased Big Bomb, 2,000 households bought at least one product of any
brand in this category. This enables us to calculate Big Bomb’s penetration share.


Big Bomb Customers
Penetration Share, Big Bomb =
Category Customers

500
= 25%
20,000

DECOMPOSING MARKET SHARE
Relationship of Penetration Share to Market Share: Market share can be calcu-
lated as the product of three components: penetration share, share of requirements,
and heavy usage index.
Market Share (%) Penetration Share (%) * Share of Requirements (%)
* Heavy Usage Index (I)
Share of Requirements: The percentage of customers’ needs in a category that are
served by a given brand or product (see Section 2.5).
Heavy Usage Index: A measure of how heavily the people who use a specific prod-
uct use the entire category of such products (see Section 2.6).
In light of these relationships, managers can use this decomposition of market share to
reveal penetration share, given the other inputs.
Market Share (%)
[Heavy Usage Index (I) * Share of Requirements (%)]

EXAMPLE: Eat Wheats brand cereal has a market share in Urbanopolis of 6%. The
heavy usage index for Eat Wheats cereal is 0.75 in Urbanopolis. Its share of requirements
is 40%. From these data, we can calculate the penetration share for Eat Wheats brand
cereal in Urbanopolis:
Market Share
Penetration Share =
(Heavy Usage Index * Share of Requirements)

6% 6%
= = 20%
(0.75 * 40%) .30


The time period over which a firm measures penetration can have a significant impact
on the penetration rate. For example, even among the most popular detergent brands,
many are not purchased weekly. As the time period used to define penetration becomes
shorter, managers can expect penetration rates to decline. By contrast, penetration share
may be less subject to this dynamic because it represents a comparison between brands,
among which the effects of shorter periods may fall approximately evenly.

RELATED METRICS AND CONCEPTS
Total Number of Active Customers: The customers (accounts) who purchased at
least once in a given time period. When assessed at a brand level, this is equivalent
to brand penetration. This term is often used in shorthand form––total number of
customers––though this would not be appropriate when a distinction must be made
for ex-customers. This is discussed in more detail in Section 5.1 (customers of a spec-
ified recency).
Accepters: Customers who are disposed to accept a given product and its benefits:
the opposite of rejecters.
Ever-tried: The percentage of a population that has tried a given brand at any
time. (See Section 4.1 for more on trial.)

2.5 Share of Requirements
Share of requirements, also known as share of wallet, is calculated solely among buy-
ers of a specific brand. Within this group, it represents the percentage of purchases
within the relevant category, accounted for by the brand in question.
Brand Purchases (#)
Unit Share of Requirements (%)
Total Category Purchases by Brand Buyers (#)

Brand Purchases ($)
Revenue Share of Requirements (%)
Total Category Purchases by Brand Buyers ($)

Many marketers view share of requirements as a key measure of loyalty. This metric
can guide a firm’s decisions on whether to allocate resources toward efforts to expand
a category, to take customers from competitors, or to increase share of requirements
among its established customers. Share of requirements is, in essence, the market
share for a brand within a market narrowly defined as the people who have already
purchased that brand.


Purpose: To understand the source of market share in terms of breadth
and depth of consumer franchise, as well as the extent of relative category
usage (heavy users/larger customers versus light users/smaller customers).

Construction
Share of Requirements: A given brand’s share of purchases in its category, meas-
ured solely among customers who have already purchased that brand. Also known
as share of wallet.
When calculating share of requirements, marketers may consider either dollars or
units. They must ensure, however, that their heavy usage index is consistent with
this choice.

Brand Purchases (#)
Unit Share of Requirements (%)
Total Category Purchases by Brand Buyers (#)

Brand Purchases ($)
Revenue Share of Requirements (%)
Total Category Purchases by Brand Buyers ($)

The best way to think about share of requirements is as the average market share
enjoyed by a product among the customers who buy it.

EXAMPLE: In a given month, the unit purchases of AloeHa brand sunscreen ran
1,000,000 bottles. Among the households that bought AloeHa, total purchases of sun-
screen came to 2,000,000 bottles.
AloeHa Purchases
Share of Requirements =
Category Purchases by AloeHa Customers

1,000,000
= 50%
2,000,000

Share of requirements is also useful in analyzing overall market share. As previously
noted, it is part of an important formulation of market share.

Market Share Penetration Share * Share of Requirements * Heavy Usage Index

Share of requirements can thus be calculated indirectly by decomposing market share.


Market Share (%)
Share of Requirements (%)
[Penetration Share (%) * Heavy Usage Index (I)]

EXAMPLE: Eat Wheats brand cereal has a market share in Urbanopolis of 8%. The
heavy usage index for Eat Wheats in Urbanopolis is 1. The brand’s penetration share in
Urbanopolis is 20%. On this basis, we can calculate Eat Wheats’ share of requirements in
Urbanopolis:
Market Share
(Heavy Usage Index * Penetration Share)

8% 8%
= = 40%
(1 * 20%) 20%

Note that in this example, market share and heavy usage index must both be defined in
the same terms (units or revenue). Depending on the definition of these two metrics, the
calculated share of requirements will be either unit share of requirements (%) or revenue
share of requirements (%).

Double Jeopardy: Some marketers strive for a “niche” positioning that yields high mar-
ket share through a combination of low penetration and high share of requirements.
That is, they seek relatively few customers but very loyal ones. Before embarking on this
strategy, however, a phenomenon known as “double jeopardy” should be considered.
Generally, the evidence suggests that it’s difficult to achieve a high share of requirements
without also attaining a high penetration share. One reason is that products with high
market share generally have high availability, whereas those with low market share may
not. Therefore, it can be difficult for customers to maintain loyalty to brands with low
market share.

Sole Usage: The fraction of a brand’s customers who use only the brand in question.
Sole Usage Percentage: The proportion of a brand’s customers who use only that
brand’s products and do not buy from competitors. Sole users may be die-hard,
loyal customers. Alternatively, they may not have access to other options, perhaps
because they live in remote areas. Where sole use is 100%, the share of wallet
is 100%.


Customers Who Buy Only the Brand in Question (#)
Sole Usage (%)
Total Brand Customers (#)

Number of Brands Purchased: During a given period, some customers may buy only
a single brand within a category, whereas others buy two or more. In evaluating loyalty
to a given brand, marketers can consider the average number of brands purchased
by consumers of that brand versus the average number purchased by all customers in
that category.

EXAMPLE: Among 10 customers for cat food, 7 bought the Arda brand, 5 bought
Bella, and 3 bought Constanza. Thus, the 10 customers made a total of 15 brand pur-
chases (7 5 3), yielding an average of 1.5 brands per customer.
Seeking to evaluate customer loyalty, a Bella brand manager notes that of his firm’s five
customers, 3 bought only Bella, whereas two bought both Arda and Bella. None of
Bella’s customers bought Constanza. Thus, the five Bella customers made seven brand
purchases (1 1 1 2 2), yielding an average of 1.4 (that is, 7/5) brands per Bella
customer. Compared to the average category purchaser, who buys 1.5 brands, Bella
buyers are slightly more loyal.

Repeat Rate: The percentage of brand customers in a given period who are also
brand customers in the subsequent period.
Repurchase Rate: The percentage of customers for a brand who repurchase that
brand on their next purchase occasion.
Confusion abounds in this area. In these definitions, we have tried to distinguish a met-
ric based on calendar time (repeat rate) from one based on “customer time” (repurchase
rate). In Chapter 5, “Customer Profitability,” we will describe a related metric, retention,
which is used in contractual situations in which the first non-renewal (non-purchase)
signals the end of a customer relationship. Although we suggest that the term retention
be applied only in contractual situations, you will often see repeat rates and repurchase
rates referred to as “retention rates.” Due to a lack consensus on the use of these terms,
marketers are advised not to rely on the names of these metrics as perfect indicators of
how they are calculated.
The importance of repeat rate depends on the time period covered. Looking at one
week’s worth of purchases is unlikely to be very illuminating. In a given category, most
consumers only buy one brand in a week. By contrast, over a period of years, consumers
may buy several brands that they do not prefer, on occasions when they can’t find the
brand to which they seek to be loyal. Consequently, the right period to consider depends


on the product under study and the frequency with which it is bought. Marketers are
advised to take care to choose a meaningful period.

2.6 Heavy Usage Index
The heavy usage index is a measure of the relative intensity of consumption. It
indicates how heavily the customers for a given brand use the product category to
which that brand belongs, compared with the average customer for that category.

Average Total Purchases in Category by Brand Customers (#,$)
Heavy Usage Index (I)
Average Total Purchases in Category by All Customers for
That Category (#,$)

or
Market Share (%)
[Penetration Share (%) * Share of Requirements (%)]

The heavy usage index, also called the weight index, yields insight into the source of
volume and the nature of a brand’s customer base.

Purpose: To define and measure whether a firm’s consumers
are “heavy users.”
The heavy usage index answers the question, “How heavily do our customers use the
category of our product?” When a brand’s heavy usage index is greater than 1.0, this sig-
nifies that its customers use the category to which it belongs more heavily than the aver-
age customer for that category.

Construction
Heavy Usage Index: The ratio that compares the average consumption of products
in a category by customers of a given brand with the average consumption of prod-
ucts in that category by all customers for the category.
The heavy usage index can be calculated on the basis of unit or dollar inputs. For a given
brand, if the heavy usage index is greater than 1.0, that brand’s customers consume an
above-average quantity or value of products in the category.
Average Total Purchases in Category by Brand Customers (#,$)
Average Total Purchases in Category by All Customers
for That Category (#,$)


EXAMPLE: Over a period of one year, the average shampoo purchases by households
using Shower Fun brand shampoo totaled six 15-oz bottles. During the same period,
average shampoo consumption by households using any brand of shampoo was four 15-
oz bottles.
The heavy usage index for households buying Shower Fun is therefore 6/4, or 1.5.
Customers of Shower Fun brand shampoo are disproportionately heavy users. They buy
50% more shampoo than the average shampoo consumer. Of course, because Shower
Fun buyers are part of the overall market average, when compared with non-users of
Shower Fun, their relative usage is even higher.

As previously noted, market share can be calculated as the product of three compo-
nents: penetration share, share of requirements, and heavy usage index (see Section 2.4).
Consequently, we can calculate a brand’s heavy usage index if we know its market share,
penetration share, and share of requirements, as follows:
Market Share (%)
[Penetration Share (%) * Share of Requirements (%)]

This equation works for market shares defined in either unit or dollar terms. As noted
earlier, the heavy usage index can measure either unit or dollar usage. Comparing a
brand’s unit heavy usage index to its dollar heavy usage index, marketers can determine
whether category purchases by that brand’s customers run above or below the average
category price.

The heavy usage index does not indicate how heavily customers use a specific brand,
only how heavily they use the category. A brand can have a high heavy usage index, for
example, meaning that its customers are heavy category users, even if those customers
use the brand in question to meet only a small share of their needs.

See also the discussion of brand development index (BDI) and category development
index (CDI) in Section 2.3.


2.7 Awareness, Attitudes, and Usage (AAU): Metrics
of the Hierarchy of Effects
Studies of awareness, attitudes, and usage (AAU) enable marketers to quantify levels
and trends in customer knowledge, perceptions, beliefs, intentions, and behaviors. In
some companies, the results of these studies are called “tracking” data because they
are used to track long-term changes in customer awareness, attitudes, and behaviors.
AAU studies are most useful when their results are set against a clear comparator.
This benchmark may comprise the data from prior periods, different markets,
or competitors.

Purpose: To track trends in customer attitudes and behaviors.
Awareness, attitudes, and usage (AAU) metrics relate closely to what has been called the
Hierarchy of Effects, an assumption that customers progress through sequential stages
from lack of awareness, through initial purchase of a product, to brand loyalty (see
Figure 2.2). AAU metrics are generally designed to track these stages of knowledge,
beliefs, and behaviors. AAU studies also may track “who” uses a brand or product––in
which customers are defined by category usage (heavy/light), geography, demographics,
psychographics, media usage, and whether they purchase other products.

Awareness Customers must first become aware of a product, then . . .

Attitudes They develop attitudes and beliefs about that product, and finally . . .

Usage Customers purchase and experience the product.

Figure 2.2 Awareness, Attitudes, and Usage: Hierarchy of Effects

Information about attitudes and beliefs offers insight into the question of why specific
users do, or do not, favor certain brands. Typically, marketers conduct surveys of large
samples of households or business customers to gather these data.


Construction
Awareness, attitudes, and usage studies feature a range of questions that aim to shed
light on customers’ relationships with a product or brand (see Table 2.4). For example,
who are the acceptors and rejecters of the product? How do customers respond to a
replay of advertising content?

Table 2.4 Awareness, Attitudes, and Usage: Typical Questions

Type Measures Typical Questions

Awareness Awareness and Knowledge Have you heard of Brand X?
What brand comes to
mind when you think
“luxury car?”
Attitudes Beliefs and Intentions Is Brand X for me?
On a scale of 1 to 5, is Brand
X for young people?
What are the strengths and
weaknesses of each brand?
Usage Purchase Habits and Loyalty Did you use Brand X this
week?
What brand did you
last buy?

Marketers use answers to these questions to construct a number of metrics. Among
these, certain “summary metrics” are considered important indicators of performance.
In many studies, for example, customers’ “willingness to recommend” and “intention to
purchase” a brand are assigned high priority. Underlying these data, various diagnostic
metrics help marketers understand why consumers may be willing––or unwilling––to
recommend or purchase that brand. Consumers may not have been aware of the brand,
for example. Alternatively, they may have been aware of it but did not subscribe to one
of its key benefit claims.

AWARENESS AND KNOWLEDGE
Marketers evaluate various levels of awareness, depending on whether the consumer
in a given study is prompted by a product’s category, brand, advertising, or usage
situation.
Awareness: The percentage of potential customers or consumers who recognize––or
name––a given brand. Marketers may research brand recognition on an “aided” or


“prompted” level, posing such questions as, “Have you heard of Mercedes?”
Alternatively, they may measure “unaided” or “unprompted” awareness, posing such
questions as, “Which makes of automobiles come to mind?”
Top of Mind: The first brand that comes to mind when a customer is asked an
unprompted question about a category. The percentage of customers for whom a
given brand is top of mind can be measured.
Ad Awareness: The percentage of target consumers or accounts who demo-
nstrate awareness (aided or unaided) of a brand’s advertising. This metric can be
campaign- or media-specific, or it can cover all advertising.
Brand/Product Knowledge: The percentage of surveyed customers who demon-
strate specific knowledge or beliefs about a brand or product.

ATTITUDES
Measures of attitude concern consumer response to a brand or product. Attitude is a
combination of what consumers believe and how strongly they feel about it. Although a
detailed exploration of attitudinal research is beyond the scope of this book, the follow-
ing summarizes certain key metrics in this field.
Attitudes/Liking/Image: A rating assigned by consumers––often on a scale of 1–5 or
1–7––when survey respondents are asked their level of agreement with such proposi-
tions as, “This is a brand for people like me,” or “This is a brand for young people.” A
metric based on such survey data can also be called relevance to customer.
Perceived Value for Money: A rating assigned by consumers––often on a scale of
1–5 or 1–7––when survey respondents are asked their level of agreement with such
propositions as, “This brand usually represents a good value for the money.”
Perceived Quality/Esteem: A consumer rating––often on a scale of 1–5 or 1–7––of
a given brand’s product when compared with others in its category or market.
Relative Perceived Quality: A consumer rating (often from 1–5 or 1–7) of brand
product compared to others in the category/market.
Intentions: A measure of customers’ stated willingness to behave in a certain way.
Information on this subject is gathered through such survey questions as, “Would
you be willing to switch brands if your favorite was not available?”
Purchase Intentions: A specific measure or rating of consumers’ stated purchase
intentions. Information on this subject is gathered through survey respondents’
reactions to such propositions as, “It is very likely that I will purchase this
product.”


USAGE
Measures of usage concern such market dynamics as purchase frequency and units
per purchase. They highlight not only what was purchased, but also when and where
it was purchased. In studying usage, marketers also seek to determine how many
people have tried a brand. Of those, they further seek to determine how many
have “rejected” the brand, and how many have “adopted” it into their regular portfo-
lio of brands.
Usage: A measure of customers’ self-reported behavior.
In measuring usage, marketers pose such questions as the following: What brand of
toothpaste did you last purchase? How many times in the past year have you purchased
toothpaste? How many tubes of toothpaste do you currently have in your home? Do you
have any Crest toothpaste in your home at the current time?
In the aggregate, AAU metrics concern a vast range of information that can be tailored
to specific companies and markets. They provide managers with insight into customers’
overall relationships with a given brand or product.

Sources of AAU data include
■ Warranty cards and registrations, often using prizes and random drawings to
encourage participation.
■ Regularly administered surveys, conducted by organizations that interview
consumers via telephone, mail, Web, or other technologies, such as hand-held
scanners.
Even with the best methodologies, however, variations observed in tracking data from
one period to the next are not always reliable. Managers must rely on their experience to
distinguish seasonality effects and “noise” (random movement) from “signal” (actual
trends and patterns). Certain techniques in data collection and review can also help
managers make this distinction.
1. Adjust for periodic changes in how questions are framed or administered.
Surveys can be conducted via mail or telephone, for example, among paid or
unpaid respondents. Different data-gathering techniques may require adjust-
ment in the norms used to evaluate a “good” or “bad” response. If sudden
changes appear in the data from one period to the next, marketers are advised
to determine whether methodological shifts might play a role in this result.
2. Try to separate customer from non-customer responses; they may be very dif-
ferent. Causal links among awareness, attitudes, and usage are rarely clear-cut.


Though the hierarchy of effects is often viewed as a one-way street, on which
awareness leads to attitudes, which in turn determine usage, the true causal flow
might also be reversed. When people own a brand, for example, they may be
predisposed to like it.
3. Triangulate customer survey data with sales revenue, shipments, or other data
related to business performance. Consumer attitudes, distributor and retail
sales, and company shipments may move in different directions. Analyzing
these patterns can be a challenge but can reveal much about category dynamics.
For example, toy shipments to retailers often occur well in advance of the
advertising that drives consumer awareness and purchase intentions. These, in
turn, must be established before retail sales. Adding further complexity, in the
toy industry, the purchaser of a product might not be its ultimate consumer. In
evaluating AAU data, marketers must understand not only the drivers of
demand but also the logistics of purchase.
4. Separate leading from lagging indicators whenever possible. In the auto indus-
try, for example, individuals who have just purchased a new car show a height-
ened sensitivity to advertisements for its make and model. Conventional
wisdom suggests that they’re looking for confirmation that they made a good
choice in a risky decision. By helping consumers justify their purchase at this
time, auto manufacturers can strengthen long-term satisfaction and willingness
to recommend.

Likeability: Because AAU considerations are so important to marketers, and
because there is no single “right” way to approach them, specialized and
proprietary systems have been developed. Of these, one of the best known is the Q
scores rating of “likeability.” A Q Score is derived from a general survey of selected
households, in which a large panel of consumers share their feelings about brands,
celebrities, and television shows.4
Q Scores rely upon responses reported by consumers. Consequently, although the
system used is sophisticated, it is dependent on consumers understanding and being
willing to reveal their preferences.
Segmentation by Geography, or Geo-clustering: Marketers can achieve insight
into consumer attitudes by separating their data into smaller, more homogeneous
groups of customers. One well-known example of this is Prizm. Prizm assigns U.S.
households to clusters based on ZIP Code,5 with the goal of creating small groups of
similar households. The typical characteristics of each Prizm cluster are known,
and these are used to assign a name to each group. “Golden Ponds” consumers, for


example, comprise elderly singles and couples leading modest lifestyles in small
towns. Rather than monitoring AAU statistics for the population as a whole, firms
often find it useful to track these data by cluster.

2.8 Customer Satisfaction and Willingness
to Recommend
Customer satisfaction is generally based on survey data and expressed as a rating. For
example, see Figure 2.3.

Very Somewhat Neither Satisfied Somewhat Very
Dissatisfied Dissatisfied nor Dissatisfied Satisfied Satisfied

1 2 3 4 5

Figure 2.3 Ratings

Within organizations, customer satisfaction ratings can have powerful effects.
They focus employees on the importance of fulfilling customers’ expectations.
Furthermore, when these ratings dip, they warn of problems that can affect sales and
profitability.
A second important metric related to satisfaction is willingness to recommend. When
a customer is satisfied with a product, he or she might recommend it to friends, rela-
tives, and colleagues. This can be a powerful marketing advantage.

Purpose: Customer satisfaction provides a leading indicator of consumer
purchase intentions and loyalty.
Customer satisfaction data are among the most frequently collected indicators of mar-
ket perceptions. Their principal use is twofold.

1. Within organizations, the collection, analysis, and dissemination of these data
send a message about the importance of tending to customers and ensuring
that they have a positive experience with the company’s goods and services.
2. Although sales or market share can indicate how well a firm is performing
currently, satisfaction is perhaps the best indicator of how likely it is that the
firm’s customers will make further purchases in the future. Much research has
focused on the relationship between customer satisfaction and retention.


Studies indicate that the ramifications of satisfaction are most strongly realized
at the extremes. On the scale in Figure 2.3, individuals who rate their satisfac-
tion level as “5” are likely to become return customers and might even evangel-
ize for the firm. Individuals who rate their satisfaction level as “1,” by contrast,
are unlikely to return. Further, they can hurt the firm by making negative
comments about it to prospective customers. Willingness to recommend is a
key metric relating to customer satisfaction.

Construction
Customer Satisfaction: The number of customers, or percentage of total customers,
whose reported experience with a firm, its products, or its services (ratings) exceeds
specified satisfaction goals.
Willingness to Recommend: The percentage of surveyed customers who indicate
that they would recommend a brand to friends.
These metrics quantify an important dynamic. When a brand has loyal customers, it
gains positive word-of-mouth marketing, which is both free and highly effective.
Customer satisfaction is measured at the individual level, but it is almost always
reported at an aggregate level. It can be, and often is, measured along various dimen-
sions. A hotel, for example, might ask customers to rate their experience with its front
desk and check-in service, with the room, with the amenities in the room, with the
restaurants, and so on. Additionally, in a holistic sense, the hotel might ask about over-
all satisfaction “with your stay.”
Customer satisfaction is generally measured on a five-point scale (see Figure 2.4).


1 2 3 4 5

Figure 2.4 A Typical Five-Point Scale

Satisfaction levels are usually reported as either “top box” or, more likely, “top
two boxes.” Marketers convert these expressions into single numbers that show the
percentage of respondents who checked either a “4” or a “5.” (This term is the same as
that commonly used in projections of trial volumes; see Section 4.1.)


EXAMPLE: The general manager of a hotel in Quebec institutes a new system of cus-
tomer satisfaction monitoring (see Figure 2.5). She leaves satisfaction surveys at check-
out. As an incentive to respond, all respondents are entered into a drawing for a pair of
free airline tickets.


Score 1 2 3 4 5
Responses 3 7 40 100 50
(200 useable)
% 2% 4% 20% 50% 25%

Figure 2.5 Hotel Customer Survey Response

The manager collects 220 responses, of which 20 are unclear or otherwise unusable.
Among the remaining 200, 3 people rate their overall experience at the hotel as very
unsatisfactory, 7 deem it somewhat unsatisfactory, and 40 respond that they are neither
satisfied nor dissatisfied. Of the remainder, 50 customers say they are very satisfied, while
the rest are somewhat satisfied.
The top box, comprising customers who rate their experience a “5,” includes 50 people
or, as a percentage, 50/200 25%. The top two boxes comprise customers who are
“somewhat” or “very” satisfied, rating their experience a “4” or “5.” In this example, the
“somewhat satisfied” population must be calculated as the total usable response pool, less
customers accounted for elsewhere, that is, 200 3 7 40 50 = 100. The sum of
the top two boxes is thus 50 100 150 customers, or 75% of the total.

Customer satisfaction data can also be collected on a 10-point scale. Regardless of the
scale used, the objective is to measure customers’ perceived satisfaction with their expe-
rience of a firm’s offerings. Marketers then aggregate these data into a percentage of top-
box responses.
In researching satisfaction, firms generally ask customers whether their product or
service has met or exceeded expectations. Thus, expectations are a key factor behind
satisfaction. When customers have high expectations and the reality falls short, they
will be disappointed and will likely rate their experience as less than satisfying. For this
reason, a luxury resort, for example, might receive a lower satisfaction rating than a
budget motel––even though its facilities and service would be deemed superior in
“absolute” terms.


Surveys constitute the most frequently used means of collecting satisfaction data. As a
result, a key risk of distortion in measures of satisfaction can be summarized in a single
question: Who responds to surveys?
“Response bias” is endemic in satisfaction data. Disappointed or angry customers often
welcome a means to vent their opinions. Contented customers often do not.
Consequently, although many customers might be happy with a product and feel no
need to complete a survey, the few who had a bad experience might be disproportion-
ately represented among respondents. Most hotels, for example, place response cards in
their rooms, asking guests, “How was your stay?’ Only a small percentage of guests ever
bother to complete those cards. Not surprisingly, those who do respond probably had a
bad experience. For this reason, marketers can find it difficult to judge the true level of
customer satisfaction. By reviewing survey data over time, however, they may discover
important trends or changes. If complaints suddenly rise, for example, that may consti-
tute early warning of a decline in quality or service. (See number of complaints in the
following section.)
Sample selection may distort satisfaction ratings in other ways as well. Because only cus-
tomers are surveyed for customer satisfaction, a firm’s ratings may rise artificially as
deeply dissatisfied customers take their business elsewhere. Also, some populations may
be more frank than others, or more prone to complain. These normative differences can
affect perceived satisfaction levels. In analyzing satisfaction data, a firm might interpret
rating differences as a sign that one market is receiving better service than another,
when the true difference lies only in the standards that customers apply. To correct for
this issue, marketers are advised to review satisfaction measures over time within the
same market.
A final caution: Because many firms define customer satisfaction as “meeting or exceed-
ing expectations,” this metric may fall simply because expectations have risen. Thus, in
interpreting ratings data, managers may come to believe that the quality of their offer-
ing has declined when that is not the case. Of course, the reverse is also true. A firm
might boost satisfaction by lowering expectations. In so doing, however, it might suffer
a decline in sales as its product or service comes to appear unattractive.

Trade Satisfaction: Founded upon the same principles as consumer satisfaction,
trade satisfaction measures the attitudes of trade customers.
Number of Complaints: The number of complaints lodged by customers in a
given time period.


2.9 Net Promoter6
Net promoter is a measure of the degree to which current customers will recommend
a product, service, or company.
Net Promoter Score (I) = Percentage of Promoters (%) – Percentage of Detractors (%)
Net promoter is claimed to be a particularly useful measure of customer satisfac-
tion and/or loyalty.

Purpose: To measure how well the brand or company is succeeding in
creating satisfied, loyal customers.
Net Promoter Score7 (NPS) is a registered trademark of Frederick R. Reichheld, Bain &
Company, and Satmetrix that is a particularly simple measure of the satisfaction/loyalty
of current customers. Customers are surveyed and asked (on a ten-point scale) how
likely they are to recommend the company or brand to a friend or colleague. Based on
their answers to this single question, customers are divided into
■ Promoters: Customers who are willing to recommend the company to others
(who gave the company a rating of 9 or 10).
■ Passives: Satisfied but unenthusiastic customers (ratings of 7 or 8).
■ Detractors: Customers who are unwilling to recommend the company to others
(ratings of 0 to 6).
High NPSs generally mean that a company is doing a good job of securing their cus-
tomers’ loyalty and active evangelism. Low and negative Net Promoter Scores are
important early warning signals for the firm. Because the metric is simple and easy to
understand, it provides a stable measure companies use to motivate employees and
monitor progress.

Construction
The Net Promoter Score (NPS) is created by subtracting the percentage of detractors
among current customers from the percentage of promoters among current customers.
Net Promoter Score (I) = Percentage of Promoters (%) – Percentage of Detractors (%)
For example if a survey of a company’s customers reports that there were 20% pro-
moters, 70% passives, and 10% detractors, the company would have a Net Promoter
Score of 20–10 =10.


Data Sources, Complications and Cautions
Although the trademarked NPS asks a specific question, uses a 10-point scale, and
defines promoters, passives, and detractors in a particular way (detractors are those giv-
ing ratings of 0 through 6), it is easy to imagine other versions of NPS that differ with
respect to the wording of the question, the scale used (1 through 5 rather than 0 through
10), and the definitions (and labels) of the resulting groups of responders. The defining
features of NPS are that it is constructed from responses to a question about willingness
to recommend and is a net measure found by subtracting the fraction unwilling to rec-
ommend from the fraction willing to recommend and leaving out those in the middle.
The same NPS score can indicate different business circumstances. For instance, a Net
Promoter Score of zero can indicate highly polarized customers, 50% promoters, 50%
detractors, or a totally ambivalent customer base, 100% passives. Getting the NPS score
may be a good way of starting a discussion about customer perceptions of the brand. As
it is an average of current customers’ responses, managers must drill down to the data
to understand the precise situation their business faces.
This score in specific circumstances can generate results that could mislead a manager
who is not being careful. For example, consider a company whose current customers are
30% promoters, 30% detractors, and 40% passives. This company’s NPS is an unim-
pressive zero, or 30%-30%.
Suppose next that a new competitor steals two-thirds of the company’s detractors, and
because these detractors immediately defect to the new competitor, they cease to be cus-
tomers of the company. The NPS is remeasured.
Promoters are now 30% / (100% – 20% = 80%) = 37.5% of the customers that remain.
Passives are now 40% / (100% – 20% = 80%) = 50% of the customers that remain.
Detractors are now only (30% – 20% = 10%) / (100% – 20% = 80%) = 12.5% of the
customers that remain.
The NPS is now 37.5% – 12.5% = a very healthy looking + 25.
The defection of the most vulnerable and unhappy customers led directly to an increase
in NPS. Managers should make sure they fully understand what has happened.
While benchmarking is often a useful exercise, it is inappropriate to directly apply this
measure across categories. Some products are in categories that are more likely to gain
engagement both positive and negative than others.
A high Net Promoter Score while generally desirable does beg the question whether the
company is properly monetizing the value they are providing to the consumer. The eas-
iest way to develop a high Net Promoter Score is to provide a highly valued product free
to customers. Why wouldn’t they be happy to recommend you? While there may be


strategic reasons for situations like this to be acceptable to the company in the short or
medium term, this probably won’t be a viable long-term strategy.
The Net Promoter Score is calculated from survey data. As such it may suffer from the
problems common to most surveys, and the results should be interpreted in light of
other data, such as sales trends. Is increased customer satisfaction leading to increased
sales? If so, fine; if not, why not?
Although the Net Promoter Score has received much attention and relatively rapid
adoption, it has also been the target of a recent award-winning article. Consultant
Timothy Keiningham and his co-authors claim the benefits of the measure have been
overstated relative to other measures of loyalty and satisfaction.8

2.10 Willingness to Search
Although many metrics explore brand loyalty, one has been called the “acid test.”
That is,
Willingness to Search (%) Percentage of Customers Willing to Delay Purchases,
Change Stores, or Reduce Purchase Quantities to
Avoid Switching Brands
This metric can tell a company much about the attitudes of its customers and
whether its position in the market is likely to be defensible against sustained pressure
from a competitor.

Purpose: To assess the commitment of a firm’s or a brand’s
customer base.
Brand or company loyalty is a key marketing asset. Marketers evaluate aspects of it
through a number of metrics, including repurchase rate, share of requirements, willing-
ness to pay a price premium, and other AAU measures. Perhaps the most fundamental
test of loyalty, however, can be captured in a simple question: When faced with a situa-
tion in which a brand is not available, will its customers search further or substitute the
best available option?
When a brand enjoys loyalty at this level, its provider can generate powerful leverage in
trade negotiations. Often, such loyalty will also give providers time to respond to a com-
petitive threat. Customers will stay with them while they address the threat.
Loyalty is grounded in a number of factors, including
■ Satisfied and influential customers who are willing to recommend the brand.


■ Hidden values or emotional benefits, which are effectively communicated.
■ A strong image for the product, the user, or the usage experience.
Purchase-based loyalty metrics are also affected by whether a product is broadly
and conveniently available for purchase, and whether customers enjoy other options in
its category.

Construction
Willingness to Search: The likelihood that customers will settle for a second-
choice product if their first choice is not available. Also called “accept no
substitutes.”
Willingness to search represents the percentage of customers who are willing to leave a
store without a product if their favorite brand is unavailable. Those willing to substitute
constitute the balance of the population.

Loyalty has multiple dimensions. Consumers who are loyal to a brand in the sense of
rarely switching may or may not be willing to pay a price premium for that brand or
recommend it to their friends. Behavioral loyalty may also be difficult to distinguish
from inertia or habit. When asked about loyalty, consumers often don’t know what they
will do in new circumstances. They may not have accurate recall about past behavior,
especially in regard to items with which they feel relatively low involvement.
Furthermore, different products generate different levels of loyalty. Few customers will
be as loyal to a brand of matches, for example, as to a brand of baby formula.
Consequently, marketers should exercise caution in comparing loyalty rates across
products. Rather, they should look for category-specific norms.
Degrees of loyalty also differ between demographic groups. Older consumers have been
shown to demonstrate the highest loyalty rates.
Even with these complexities, however, customer loyalty remains one of the most
important metrics to monitor. Marketers should understand the worth of their brands
in the eyes of the customer––and of the retailer.


3
MARGINS AND PROFITS

Introduction

Margins Marketing Spending—Total, Fixed, and
Selling Prices and Channel Margins Variable

Average Price per Unit and Price Break-Even Analysis and Contribution
per Statistical Unit Analysis

Variable Costs and Fixed Costs Target Volume

Peter Drucker has written that the purpose of a business is to create a customer. As mar-
keters, we agree. But we also recognize that a business can’t survive unless it makes a
margin as well as a customer. At one level, margins are simply the difference between a
product’s price and its cost. This calculation becomes more complicated, however, when
multiple variations of a product are sold at multiple prices, through multiple channels,
incurring different costs along the way. For example, a recent Business Week article noted
that less “than two-thirds of GM’s sales are retail. The rest go to rental-car agencies or to
company employees and their families—sales that provide lower gross margins.”1
Although it is still the case that a business can’t survive unless it earns a positive margin,
it can be a challenge to determine precisely what margin the firm actually does earn.
In the first section of this chapter, we’ll explain the basic computation of unit and
percentage margins, and we’ll introduce the practice of calculating margins as a
percentage of selling price.
Next, we’ll show how to “chain” this calculation through two or more levels in a
distribution channel and how to calculate end-user purchase price on the basis of a

65

marketer’s selling price. We’ll explain how to combine sales through different channels
to calculate average margins and how to compare the economics of different distribution
channels.
In the third section, we’ll discuss the use of “statistical” and standard units in tracking
price changes over time.
We’ll then turn our attention to measuring product costs, with particular emphasis on
the distinction between fixed and variable costs. The margin between a product’s unit
price and its variable cost per unit represents a key calculation. It tells us how much the
sale of each unit of that product will contribute to covering a firm’s fixed costs.
“Contribution margin” on sales is one of the most useful marketing concepts. It
requires, however, that we separate fixed from variable costs, and that is often a chal-
lenge. Frequently, marketers must take “as a given” which of their firm’s operating and
production costs are fixed and which are variable. They are likely, however, to be respon-
sible for making these fixed versus variable distinctions for marketing costs. That is the
subject of the fifth section of this chapter.
In the sixth section, we’ll discuss the use of fixed- and variable-cost estimates in calcu-
lating the break-even levels of sales and contribution. Finally, we’ll extend our calcula-
tion of break-even points, showing how to identify sales and profit targets that are
mutually consistent.

3.1 Unit Margin Unit price less the What are the stan- Determine value of
unit cost. dard units in the incremental sales.
industry? May not Guide pricing and
reflect contribution promotion.
margin if some
fixed costs are
allocated.
3.1 Margin (%) Unit margin as a May not reflect Compare margins
percentage of unit contribution margin across different
price. if some fixed costs are products/sizes/
allocated. forms of product.
Determine value of
incremental sales.
Guide pricing
and promotion
decisions.


3.2 Channel Channel profits as Distinguish margin Evaluate channel
Margins percentage of chan- on sales (usual) from value added in
nel selling price. markup on cost (also context of selling
encountered). price. Calculate
effect of price
changes at one
level of channel on
prices and margins
at other levels in
the same channel
(supply chain).

3.3 Average Price Can be calculated Some units may have Understand how
per Unit as total revenue greater relevance average prices are
divided by total from producers’ per- affected by shifts in
unit sales. spective than con- pricing and prod-
sumers’ (e.g., ounces uct mix.
of shampoo vs. bot-
tles). Changes may
not be result of pric-
ing decisions.
3.3 Price per SKU prices Percentage SKU mix Isolate effect of
Statistical weighted by rele- should correspond price changes from
Unit vant percentage of over medium-term to mix changes by
each SKU in a sta- actual mix of sales. standardizing the
tistical unit. SKU mix of a stan-
dard unit.

3.4 Variable and Divide costs into Variable costs may Understand how
Fixed Costs two categories: include production, costs are affected
those that vary marketing, and sell- by changes in sales
with volume (vari- ing expenses. Some volume.
able) and those that variable costs depend
do not (fixed). on units sold; others
depend on revenue.

3.5 Marketing Analyze costs that Can be divided into Understand how
Spending comprise market- fixed and variable marketing spend-
ing spending. marketing costs. ing changes with
sales.

Continues

Chapter 3 Margins and Profits 67

3.6 Contribution per Unit price less Ensure that mar- Understand profit
Unit unit variable cost. keting variable impact of changes
costs have not in volume.
already been Calculate break-
deducted from even level of sales.
price.

3.6 Contribution Contribution per Ensure that Same as above,
Margin (%) unit divided by variable costs are but applies to
unit price. consistently based dollar sales.
on units or
revenue, as
appropriate.

3.6 Break-Even Sales For unit break- Variable and fixed Rough
Level even, divide fixed cost estimates may indicator of
costs by contribu- be valid only over project attractive-
tion per unit. For certain ranges ness and ability to
revenue break- of sales and earn profit.
even, divide fixed production.
costs by contribu-
tion margin (%).

3.7 Target Volume Adjust break-even Variable market- Ensure that unit
calculation to ing costs must be sales objectives
include profit reflected in contri- will enable firm to
target. bution margins. achieve financial
Sales increases hurdle rates for
often require profit, ROS, or
increased invest- ROI.
ment or working
capital.
3.7 Target Revenues Convert target Same as above. Same as above,
volume to target applied to revenue
revenues by using objectives.
average prices per
unit. Alternatively,
combine cost and
target data with
knowledge of con-
tribution margins.


3.1 Margins
Margin (on sales) is the difference between selling price and cost. This difference is
typically expressed either as a percentage of selling price or on a per-unit basis.
Unit Margin ($) = Selling Price per Unit ($) Cost per Unit ($)
Unit Margin ($)
Margin (%) =
Selling Price per Unit ($)
Managers need to know margins for almost all marketing decisions. Margins repre-
sent a key factor in pricing, return on marketing spending, earnings forecasts, and
analyses of customer profitability.

Purpose: To determine the value of incremental sales,
and to guide pricing and promotion decisions.
Margin on sales represents a key factor behind many of the most fundamental business
considerations, including budgets and forecasts. All managers should, and generally do,
know their approximate business margins. Managers differ widely, however, in the
assumptions they use in calculating margins and in the ways they analyze and commu-
nicate these important figures.
Percentage Margins and Unit Margins: A fundamental variation in the way people talk
about margins lies in the difference between percentage margins and unit margins on
sales. The difference is easy to reconcile, and managers should be able to switch back and
forth between the two.
What is a unit? Every business has its own notion of a “unit,” ranging from a ton of mar-
garine, to 64 ounces of cola, to a bucket of plaster. Many industries work with multiple
units and calculate margin accordingly. The cigarette industry, for example, sells “sticks,”
“packs,” “cartons,” and 12M “cases” (which hold 1,200 individual cigarettes). Banks cal-
culate margin on the basis of accounts, customers, loans, transactions, households, and
branch offices. Marketers must be prepared to shift between such varying perspectives
with little effort because decisions can be grounded in any of these perspectives.

Construction
Unit Margin ($) = Selling Price per Unit ($) Cost per Unit ($)
Unit Margin ($)
Margin (%) =


Percentage margins can also be calculated using total sales revenue and total costs.

[Total Sales Revenue ($) Total Cost ($)]
Margin (%) =
Total Sales Revenue ($)
When working with either percentage or unit margins, marketers can perform a simple
check by verifying that the individual parts sum to the total.
To Verify a Unit Margin ($): Selling Price per Unit = Unit Margin Cost per Unit
To Verify a Margin (%): Cost as % of Sales = 100% Margin %

EXAMPLE: A company markets sailcloth by the lineal yard. Its cost basis and selling
price for standard cloth are as follows:

Unit Selling Price (Selling Price per Unit) = $24 per Lineal Yard
Unit Cost (Cost per Unit) = $18 per Lineal Yard
To calculate unit margin, we subtract the cost from the selling price:
Unit Margin = $24 per Yard $18 per Yard
= $6 per Yard
To calculate the percentage margin, we divide the unit margin by the selling price:
($24 $18) per Yard
Margin (%) =
$24
$6
= = 25%
$24
Let’s verify that our calculations are correct:
Unit Selling Price = Unit Margin Unit Cost
$24 per Yard = $6 per Yard $18 per Yard correct
A similar check can be made on our calculations of percentage margin:
100% Margin on Sales (%) = Cost as % of Selling Price
$18
100% 25% =
$24
75% = 75% correct


When considering multiple products with different revenues and costs, we can calculate
overall margin (%) on either of two bases:
■ Total revenue and total costs for all products, or
■ The dollar-weighted average of the percentage margins of the different
products

EXAMPLE: The sailcloth company produces a new line of deluxe cloth, which sells for $64
per lineal yard and costs $32 per yard to produce. The margin on this item is 50%.

Unit Margin ($) = $64 per Yard $32 per Yard
= $32 per Yard
($64 $32)
Margin (%) =
$64
$32
=
$64
= 50%
Because the company now sells two different products, its average margin can only be calcu-
lated when we know the volume of each type of goods sold. It would not be accurate to take a
simple average of the 25% margin on standard cloth and the 50% margin on deluxe cloth,
unless the company sells the same dollar volume of both products.
If, one day, the company sells 20 yards of standard cloth and two yards of deluxe cloth, we
can calculate its margins for that day as follows (see also Table 3.1):

Total Sales = 20 Yards at $24, and 2 Yards at $64
= $608
Total Costs = 20 Yards at $18, and 2 Yards at $32
= $424
Margin ($) = $184
Margin ($184)
Margin (%) =
Total Sales ($608)
= 30%
Because dollar sales differ between the two products, the company margin of 30% is not
a simple average of the margins of those products.


Table 3.1 Sales, Costs, and Margins

Standard Deluxe Total

Sales in Yards 20 2 22

Selling Price per Yard $24.00 $64.00
Total Sales $ $480.00 $128.00 $608.00

Cost per Yard $18.00 $32.00
Total Costs $ $360.00 $64.00 $424.00

Total Dollar Margin ($) $120.00 $64.00 $184.00
Unit Margin $6.00 $32.00 $8.36

Margin (%) 25% 50% 30%

After you determine which units to use, you need two inputs to determine margins: unit
costs and unit selling prices.
Selling prices can be defined before or after various “charges” are taken: Rebates, cus-
tomer discounts, brokers’ fees, and commissions can be reported to management either
as costs or as deductions from the selling price. Furthermore, external reporting can
vary from management reporting because accounting standards might dictate a treat-
ment that differs from internal practices. Reported margins can vary widely, depending
on the calculation technique used. This can result in deep organizational confusion on
as fundamental a question as what the price of a product actually is.
Please see Section 8.4 on price waterfalls for cautions on deducting certain discounts
and allowances in calculating “net prices.” Often, there is considerable latitude on
whether certain items are subtracted from list price to calculate a net price or are added
to costs. One example is the retail practice of providing gift certificates to customers
who purchase certain amounts of goods. It is not easy to account for these in a way that
avoids confusion among prices, marketing costs, and margins. In this context, two
points are relevant: (1) Certain items can be treated either as deductions from prices or
as increments to cost, but not both. (2) The treatment of such an item will not affect the
unit margin, but will affect the percentage margin.
Margin as a percentage of costs: Some industries, particularly retail, calculate margin as
a percentage of costs, not of selling prices. Using this technique in the previous example,
the percentage margin on a yard of standard sailcloth would be reckoned as the


Table 3.2 Relationship Between Margins and Markups

Price Cost Margin Markup

$10 $9.00 10% 11%

$10 $7.50 25% 33%

$10 $6.67 33.3% 50%

$10 $5.00 50% 100%

$10 $4.00 60% 150%

$10 $3.33 66.7% 200%

$10 $2.50 75% 300%

$6.00 unit margin divided by the $18.00 unit cost, or 33%. This can lead to confusion.
Marketers must become familiar with the practices in their industry and stand ready to
shift between them as needed.
Markup or margin? Although some people use the terms “margin” and “markup” inter-
changeably, this is not appropriate. The term “markup” commonly refers to the practice
of adding a percentage to costs in order to calculate selling prices.
To get a better idea of the relationship between margin and markup, let’s calculate a few.
For example, a 50% markup on a variable cost of $10 would be $5, yielding a retail price
of $15. By contrast, the margin on an item that sells at a retail price of $15 and that car-
ries a variable cost of $10 would be $5/$15, or 33.3%. Table 3.2 shows some common
margin/markup relationships.
One of the peculiarities that can occur in retail is that prices are “marked up” as a per-
centage of a store’s purchase price (its variable cost for an item) but “marked down”
during sales events as a percentage of retail price. Most customers understand that a
50% “sale” means that retail prices have been marked down by 50%.

EXAMPLE: An apparel retailer buys t-shirts for $10 and sells them at a 50% markup.
As noted previously, a 50% markup on a variable cost of $10 yields a retail price of $15.
Unfortunately, the goods don’t sell, and the store owner wants to sell them at cost to clear
shelf space. He carelessly asks a sales assistant to mark the goods down by 50%. This 50%
markdown, however, reduces the retail price to $7.50. Thus, a 50% markup followed by a
50% markdown results in a loss of $2.50 on each unit sold.


It is easy to see how confusion can occur. We generally prefer to use the term margin to
refer to margin on sales. We recommend, however, that all managers clarify with their
colleagues what is meant by this important term.

EXAMPLE: A wireless provider sells a handset for $100. The handset costs $50 to
manufacture and includes a $20 mail-in rebate. The provider’s internal reports add this
rebate to the cost of goods sold. Its margin calculations therefore run as follows:
Unit Margin ($) = Selling Price Cost of Goods Sold and Rebate
= $100 ($50 + $20) = $30
$30
Margin (%) = = 30%
$100

Accounting standards mandate, however, that external reports deduct rebates from sales
revenue (see Table 3.3). Under this construction, the company’s margin calculations run
differently and yield a different percentage margin:
Unit Margin ($) Selling Price, Net of Rebate Cost of Goods Sold
($100 $20) $50 = $30
$30
Margin (%) =
($100 $20)
$30
= = 37.5%
$80

Table 3.3 Internal and External Reporting May Vary

Internal Reporting External Reporting

Dollars Received from Customer $100 $100
Rebate — $20
Sales $100 $80

Manufacturing Cost $50 $50
Rebate $20 —
Cost of Goods Sold $70 $50

Unit Margin ($) $30 $30
Margin (%) 30.0% 37.5%

In this example, managers add the rebate to cost of goods sold for the sake of internal
reports. In contrast, accounting regulations require that the rebate be deducted from


sales for the purpose of external reports. This means that the percentage margin varies
between the internal and external reports. This can cause considerable angst within the
company when quoting a percentage margin.

As a general principle, we recommend that internal margins follow formats mandated
for external reporting in order to limit confusion.
Various costs may or may not be included: The inclusion or exclusion of costs generally
depends on the intended purpose of the relevant margin calculations. We’ll return to this
issue several times. At one extreme, if all costs are included, then margin and net profit
will be equivalent. On the other hand, a marketer may choose to work with “contribution
margin” (deducting only variable costs), “operating margin,” or “margin before market-
ing.” By using certain metrics, marketers can distinguish fixed from variable costs and can
isolate particular costs of an operation or of a department from the overall business.

Gross Margin: This is the difference between revenue and cost before accounting
for certain other costs. Generally, it is calculated as the selling price of an item, less
the cost of goods sold (production or acquisition costs, essentially). Gross margin
can be expressed as a percentage or in total dollar terms. If the latter, it can be
reported on a per-unit basis or on a per-period basis for a company.

3.2 Prices and Channel Margins
Channel margins can be expressed on a per-unit basis or as a percentage of selling
price. In “chaining” the margins of sequential distribution channels, the selling price
of one channel member becomes the “cost” of the channel member for which it
serves as a supplier.
Supplier Selling Price ($) = Customer Selling Price ($) Customer Margin ($)
Supplier Selling Price ($)
Customer Selling Price ($) =
[1 Customer Margin (%)]
When there are several levels in a distribution chain—including a manufacturer, dis-
tributor, and retailer, for example—one must not simply add all channel margins as
reported in order to calculate “total” channel margin. Instead, use the selling prices at
the beginning and end of the distribution chain (that is, at the levels of the manufac-
turer and the retailer) to calculate total channel margin. Marketers should be able to
work forward from their own selling price to the consumer’s purchase price and
should understand channel margins at each step.


Purpose: To calculate selling prices at each level in the
distribution channel.
Marketing often involves selling through a series of “value-added” resellers. Sometimes,
a product changes form through this progression. At other times, its price is simply
“marked up” along its journey through the distribution channel (see Figure 3.1).
In some industries, such as imported beer, there may be as many as four or five channel
members that sequentially apply their own margins before a product reaches the con-
sumer. In such cases, it is particularly important to understand channel margins and
pricing practices in order to evaluate the effects of price changes.

Buys Raw
Materials for
$0.50
Buys from
Manufacturer Manufacturer
for $1.00
Sells to Buys from
Distributor Distributor Distributor
for $1.00 for $2.00
Sells to Buys from
Wholesaler Wholesaler Wholesaler
for $2.00 for $3.00
Buys
Sells to from
Retailer Retailer Retailer
for $3.00 for $5.00
Sells to
Consumer Consumer
for $5.00

Unit
Margin $0.50 $1.00 $1.00 $2.00 $5.00

Margin 50% 50% 33.3% 40%
%

Margin ($) for entire chain $4.50
Margin (%) 90%

Figure 3.1 Example of a Distribution Channel

Remember: Selling Price = Cost + Margin


Construction
First, decide whether you want to work “backward,” from customer selling prices to sup-
plier selling prices, or “forward.” We provide two equations to use in working backward,
one for dollar margins and the other for percentage margins:
Supplier Selling Price ($) = Customer Selling Price ($) Customer Margin ($)
Supplier Selling Price ($) = Customer Selling Price ($) * [1 Customer Margin (%)]

EXAMPLE: Aaron owns a small furniture store. He buys BookCo brand bookcases
from a local distributor for $200 per unit. Aaron is considering buying directly from
BookCo, and he wants to calculate what he would pay if he received the same price that
BookCo charges his distributor. Aaron knows that the distributor’s percentage margin
is 30%.
The manufacturer supplies the distributor. That is, in this link of the chain, the manu-
facturer is the supplier, and the distributor is the customer. Thus, because we know the
customer’s percentage margin, in order to calculate the manufacturer’s price to Aaron’s
distributor, we can use the second of the two previous equations.
Supplier Selling Price ($) = Customer Selling Price ($) * [1 Customer Margin (%)]
= $200 * 70% = $140
Aaron’s distributor buys each bookcase for $140 and sells it for $200, earning a margin of
$60 (30%).

Although the previous example may be the most intuitive version of this formula, by
rearranging the equation, we can also work forward in the chain, from supplier prices to
customer selling prices. In a forward-looking construction, we can solve for the cus-
tomer selling price, that is, the price charged to the next level of the chain, moving
toward the end consumer.2
Supplier Selling Price ($)
Customer Selling Price ($) =
[1 Customer Margin (%)]
Customer Selling Price ($) = Supplier Selling Price ($) + Customer Margin ($)

EXAMPLE: Clyde’s Concrete sells 100 cubic yards of concrete for $300 to a road con-
struction contractor. The contractor wants to include this in her bill of materials, to be
charged to a local government (see Figure 3.2). Further, she wants to earn a 25% margin.
What is the contractor’s selling price for the concrete?


Supplier to Supplier to

Clyde The Contractor Local Government

Customer of Customer of

Figure 3.2 Customer Relationships

This question focuses on the link between Clyde’s Concrete (supplier) and the contractor
(customer). We know the supplier’s selling price is $300 and the customer’s intended
margin is 25%. With this information, we can use the first of the two previous equations.
Supplier Selling Price
Customer Selling Price =
(1 Customer Margin %)
$300
=
(1 25%)
$300
= = $400
75%
To verify our calculations, we can determine the contractor’s percentage margin, based
on a selling price of $400 and a cost of $300.
(Customer Selling Price Supplier Selling Price)
Customer Margin =
Customer Selling Price
($400 $300)
=
$400
$100
= = 25%
$400

First Channel Member’s Selling Price: Equipped with these equations and with knowl-
edge of all the margins in a chain of distribution, we can work all the way back to the
selling price of the first channel member in the chain.
First Channel Member’s Selling Price ($) = Last Channel Member’s Selling Price ($) * [1 Last
Channel Margin (%)] * [1 Next-to-last Channel
Margin (%)] * [1 Next-to-next-to-last Channel
Margin (%)] . . . and so on


EXAMPLE: The following margins are received at various steps along the chain of dis-
tribution for a jar of pasta sauce that sells for a retail price of $5.00 (see Table 3.4).
What does it cost the manufacturer to produce a jar of pasta sauce? The retail selling
price ($5.00), multiplied by 1 less the retailer margin, will yield the wholesaler selling
price. The wholesaler selling price can also be viewed as the cost to the retailer. The cost
to the wholesaler (distributor selling price) can be found by multiplying the wholesaler
selling price by 1 less the wholesaler margin, and so forth. Alternatively, one might follow
the next procedure, using a channel member’s percentage margin to calculate its dollar
margin, and then subtracting that figure from the channel member’s selling price to
obtain its cost (see Table 3.5).
Thus, a jar of pasta that sells for $5.00 at retail actually costs the manufacturer 50 cents
to make.

Table 3.4 Example—Pasta Sauce Distribution Margins

Distribution Stage Margin

Manufacturer 50%

Distributor 50%

Wholesaler 33%

Retailer 40%

Table 3.5 Cost (Purchase Price) of Retailer

Stage Margin % $

Cost to Consumer $5.00
Retailer Margin 40% $2.00

Cost to Retailer $3.00
Wholesaler Margin 33% $1.00

Cost to Wholesaler $2.00
Distributor Margin 50% $1.00

Cost to Distributor $1.00
Manufacturer Margin 50% $0.50

Manufacturer’s Cost $0.50


The margins taken at multiple levels of a distribution process can have a dramatic
effect on the price paid by consumers. To work backward in analyzing these, many
people find it easier to convert markups to margins. Working forward does not require
this conversion.

EXAMPLE: To show that margins and markups are two sides of the same coin, let’s
demonstrate that we can obtain the same sequence of prices by using the markup method
here. Let’s look at how the pasta sauce is marked up to arrive at a final consumer price
of $5.00.
As noted previously, the manufacturer’s cost is $0.50. The manufacturer’s percentage markup
is 100%. Thus, we can calculate its dollar markup as $0.50 * 100% = $0.50. Adding the manu-
facturer’s markup to its cost, we arrive at its selling price: $0.50 (cost) + $0.50 (markup) = $1.00.
The manufacturer sells the sauce for $1.00 to a distributor. The distributor applies a markup
of 100%, taking the price to $2.00, and sells the sauce to a wholesaler. The wholesaler applies
a markup of 50% and sells the sauce to a retailer for $3.00. Finally, the retailer applies a
markup of 66.7% and sells the pasta sauce to a consumer for $5.00. In Table 3.6, we track these
markups to show the pasta sauce’s journey from a manufacturer’s cost of $0.50 to a retail price
(consumer’s cost) of $5.00.

Table 3.6 Markups Along the Distribution Channel

Stage Markup % $ Margin

Manufacturer’s Cost $0.50
Manufacturer Markup 100% $0.50 50%

Cost to Distributor $1.00
Distributor Markup 100% $1.00 50%

Cost to Wholesaler $2.00
Wholesaler Markup 50% $1.00 33.3%

Cost to Retailer $3.00
Retailer Markup 67% $2.00 40%

Cost to Consumer $5.00


The information needed to calculate channel margins is the same as for basic margins.
Complications arise, however, because of the layers involved. In this structure, the sell-
ing price for one layer in the chain becomes the cost to the next layer. This is clearly vis-
ible in consumer goods industries, where there are often multiple levels of distribution
between the manufacturer and the consumer, and each channel member requires its
own margin.
Cost and selling price depend on location within the chain. One must always ask,
“Whose cost is this?” and “Who sells at this price?” The process of “chaining” a sequence
of margins is not difficult. One need only clarify who sells to whom. In tracking this, it
can help first to draw a horizontal line, labeling all the channel members along the
chain, with the manufacturer at the far left and the retailer on the right. For example, if
a beer exporter in Germany sells to an importer in the U.S., and that importer sells to a
distributor in Virginia, who sells the beer to a retailer, then four distinct selling prices
and three channel margins will intervene between the exporter and retail store cus-
tomer. In this scenario, the exporter is the first supplier. The importer is the first cus-
tomer. To avoid confusion, we recommend mapping out the channel and calculating
margins, purchase prices, and selling prices at each level.
Throughout this section, we’ve assumed that all margins are “gross margins,” calculated
as selling price minus cost of goods sold. Of course, channel members will incur other
costs in the process of “adding value.” If a wholesaler pays his salespeople a commission
on sales, for example, that would be a cost of doing business. But it would not be a part
of the cost of goods sold, and so it is not factored into gross margin.

HYBRID (MIXED) CHANNEL MARGINS
Hybrid Channel: The use of multiple distribution systems to reach the same
market. A company might approach consumers through stores, the Web, and
telemarketing, for example. Margins often differ among such channels. Hybrid
channels may also be known as mixed channels.
Increasingly, businesses “go to market” in more than one way. An insurance company,
for example, might sell policies through independent agents, toll-free telephone lines,
and the Web. Multiple channels often generate different channel margins and cause a
supplier to incur different support costs. As business migrates from one channel to
another, marketers must adjust pricing and support in economically sensible ways. To
make appropriate decisions, they must recognize the more profitable channels in their
mix and develop programs and strategies to fit these.


When selling through multiple channels with different margins, it is important to per-
form analyses on the basis of weighted average channel margins, as opposed to a simple
average. Using a simple average can lead to confusion and poor decision-making.
As an example of the variations that can occur, let’s suppose that a company sells
10 units of its product through six channels. It sells five units through one channel
at a 20% margin, and one unit through each of the other five channels at a 50%
margin. Calculating its average margin on a weighted basis, we arrive at the following
figure:

[(5 * 20%) (5 * 50%)]
Percentage Margin (%) 35%
10
By contrast, if we calculate the average margin among this firm’s six channels on a
simple basis, we arrive at a very different figure:

[(1 * 20%) (5 * 50%)]
Percentage Margin (%) 45%
6
This difference in margin could significantly blur management decision-making.

AVERAGE MARGIN
When assessing margins in dollar terms, use percentage of unit sales.
Average Margin ($) [Percentage of Unit Sales through Channel 1 (%) * Margin
Earned in Channel 1 ($)] [Percentage of Unit Sales
through Channel 2 (%) * Margin Earned in Channel 2 ($)]
Continued to Last Channel
When assessing margin in percentage terms, use percentage of dollar sales.
Average Margin (%) [Percentage of Dollar Sales through Channel 1 (%) * Margin
Earned in Channel 1 (%)] [Percentage of Dollar Sales
through Channel 2 (%) * Margin Earned in Channel 2 (%)]
Continued to Last Channel

EXAMPLE: Gael’s Glass sells through three channels: phone, online, and store. These
channels generate the following margins: 50%, 40%, and 30%, respectively. When Gael’s
wife asks what his average margin is, he initially calculates a simple margin and says it’s
40%. Gael’s wife investigates further, however, and learns that her husband answered too
quickly. Gael’s company sells a total of 10 units. It sells one unit by phone at a 50% mar-
gin, four units online at a 40% margin, and five units in the store at a 30% margin.


To determine the company’s average margin among these channels, the margin in each
must be weighted by its relative sales volume. On this basis, Gael’s wife calculates the
weighted average margin as follows:
Average Channel Margin = (Percentage of Unit Sales by Phone * Phone Channel Margin)
(Percentage of Unit Sales Online * Online Channel Margin)
(Percentage of Unit Sales through Store * Store Channel
Margin)
= (1/10 * 50%) (4/10 * 40%) (5/10 * 30%)
= 5% 16% 15%
Average Channel Margin = 36%

EXAMPLE: Sadetta, Inc. has two channels—online and retail—which generate the
following results:
One customer orders online, paying $10 for one unit of goods that costs the company $5.
This generates a 50% margin for Sadetta. A second customer shops at the store, buying
two units of product for $12 each. Each costs $9. Thus, Sadetta earns a 25% margin on
these sales. Summarizing:
Online Margin (1) 50%. Selling Price (1) $10. Supplier Selling Price (1) $5.
Store Margin (2) 25%. Selling Price (2) $12. Supplier Selling Price (2) $9.
In this scenario, the relative weightings are easy to establish. In unit terms, Sadetta sells a
total of three units: one unit (33.3%) online, and two (66.6%) in the store. In dollar
terms, Sadetta generates a total of $34 in sales: $10 (29.4%) online, and $24 (70.6%) in
the store.
Thus, Sadetta’s average unit margin ($) can be calculated as follows: The online channel
generates a $5.00 margin, while the store generates a $3.00 margin. The relative weight-
ings are online 33.3% and store 66.6%.
Average Unit Margin ($) = [Percentage Unit Sales Online (%) * Unit Margin Online ($)]
[Percentage Unit Sales in Store (%) * Unit Margin in Store ($)]
= 33.3% * $5.00 66.6% * $3.00
= $1.67 $2.00
= $3.67
Sadetta’s average margin (%) can be calculated as follows: The online channel generates
a 50% margin, while the store generates a 25% margin. The relative weightings are
online 29.4% and store 70.6%.


Average Margin (%) = [Percentage Dollar Sales Online (%) * Margin Online (%)]
[Percentage Dollar Sales in Store (%) * Margin in Store (%)]
= 29.4% * 50% 70.6% * 25%
= 14.70% 17.65%
= 32.35%
Average margins can also be calculated directly from company totals. Sadetta, Inc. generated a
total gross margin of $11 by selling three units of product. Its average unit margin was thus
$11/3, or $3.67. Similarly, we can derive Sadetta’s average percentage margin by dividing its
total margin by its total revenue. This yields a result that matches our weighted previous calcu-
lations: $11/$34 = 32.35%.

The same weighting process is needed to calculate average selling prices.
Average Selling Price ($) = [Percentage Unit Sales through Channel 1 (%)
* Selling Price in Channel 1 ($)] [Percentage Unit Sales
through Channel 2 (%) * Selling Price in Channel 2 ($)]
Continued to [Percentage Unit Sales through the Last
Channel (%) * the Last Channel’s Selling Price ($)]

EXAMPLE: Continuing the previous example, we can see how Sadetta, Inc. calculates
its average selling price.
Sadetta’s online customer pays $10 per item. Its store customer pays $12 per item. Weighting
each channel by unit sales, we can derive Sadetta’s average selling price as follows:
Average Selling Price ($) = [Percentage Unit Sales Online (%) * Selling Price Online ($)]
[Percentage Unit Sales in Store (%) * Selling Price in Store ($)]
= 33.3% * $10 + 66.7% * $12
= $3.33 $8
= $11.33

The calculation of average supplier selling price is conceptually similar.
Average Supplier Selling Price ($) = [Percentage Unit Sales through Channel 1 (%)
* Supplier Selling Price in Channel 1 ($)] [Percentage
Unit Sales through Channel 2 (%) * Supplier Selling
Price in Channel 2 ($)] Continued to [Percentage Unit
Sales through the Last Channel (%) * the Last Channel
Supplier’s Selling Price ($)]


EXAMPLE: Now, let’s consider how Sadetta, Inc. calculates its average supplier selling
price.
Sadetta’s online merchandise cost the company $5 per unit. Its in-store merchandise cost
$9 per unit. Thus:
Average Supplier Selling Price ($) = [Percentage Unit Sales Online (%) * Supplier Selling
Price Online ($)] [Percentage Unit Sales through
Store (%) * Supplier Selling Price in Store ($)]
= 33.3% * $5 + 66.7% * $9
= $1.67 $6 = $7.67
With all these pieces of the puzzle, we now have much greater insight into Sadetta, Inc.’s
business (see Table 3.7).

Table 3.7 Sadetta’s Channel Measures

Online In Store Average/Total

Selling Price (SP) $10.00 $12.00
Supplier Selling Price (SSP) $5.00 $9.00
Unit Margin ($) $5.00 $3.00
Margin (%) 50% 25%
Units Sold 1 2 3
% Unit Sales 33.3% 66.7%
Dollar Sales $10.00 $24.00 $34.00
% Dollar Sales 29.4% 70.6%
Total Margin $5.00 $6.00 $11.00
Average Unit Margin ($) $3.67
Average Margin (%) 32.4%
Average Selling Price $11.33
Average Supplier Selling Price $7.67

3.3 Average Price per Unit and Price per Statistical Unit
Average prices represent, quite simply, total sales revenue divided by total units sold.
Many products, however, are sold in multiple variants, such as bottle sizes. In these
cases, managers face a challenge: They must determine “comparable” units.


Average prices can be calculated by weighting different unit selling prices by the per-
centage of unit sales (mix) for each product variant. If we use a standard, rather than
an actual mix of sizes and product varieties, the result is price per statistical unit.
Statistical units are also known as equivalent units.
Revenue ($)
Average Price per Unit ($) =
Units Sold (#)
or
[Price of SKU 1 ($) * SKU 1 Percentage of Sales (%)]
[Price of SKU 2 ($) * SKU 2 Percentage of Sales (%)]
Price per Statistical Unit ($) Total Price of a Bundle of SKUs Comprising
a Statistical Unit ($)
Price per Statistical Unit ($)
Unit Price per Statistical Unit ($) =
Total Units in the Bundle of SKUs
Comprising that Statistical Unit (#)

Average price per unit and prices per statistical unit are needed by marketers who
sell the same product in different packages, sizes, forms, or configurations at a
variety of different prices. As in analyses of different channels, these product and
price variations must be reflected accurately in overall average prices. If they are
not, marketers may lose sight of what is happening to prices and why. If the price of
each product variant remained unchanged, for example, but there was a shift in the
mix of volume sold, then the average price per unit would change, but the price per
statistical unit would not. Both of these metrics have value in identifying market
movements.

Purpose: To calculate meaningful average selling prices within a product
line that includes items of different sizes.
Many brands or product lines include multiple models, versions, flavors, colors, sizes,
or—more generally—stock keeping units (SKUs). Brita water filters, for example, are
sold in a number of SKUs. They are sold in single-filter packs, double-filter packs, and
special banded packs that may be restricted to club stores. They are sold on a standalone
basis and in combination with pitchers. These various packages and product forms may
be known as SKUs, models, items, and so on.
Stock Keeping Unit (SKU): A term used by retailers to identify individual items
that are carried or “stocked” within an assortment. This is the most detailed level
at which the inventory and sales of individual products are recorded.


Marketers often want to know both their own average prices and those of retailers. By
reckoning in terms of SKUs, they can calculate an average price per unit at any level in
the distribution chain. Two of the most useful of these averages are

1. A unit price average that includes all sales of all SKUs, expressed as an average
price per defined unit. In the water filter industry, for example, these might
include such figures as $2.23/filter, $0.03/filtered ounce, and so on.
2. A price per statistical unit that consists of a fixed bundle (number) of individ-
ual SKUs. This bundle is often constructed so as to reflect the actual mix of
sales of the various SKUs.
The average price per unit will change when there is a shift in the percentage of sales
represented by SKUs with different unit prices. It will also change when the prices of the
individual SKUs are modified. This contrasts with price per statistical unit, which,
by definition, has a fixed proportion of each SKU. Consequently, a price per statistical
unit will change only when there is a change in the price of one or more of the SKUs
included in it.
The information gleaned from a price per statistical unit can be helpful in considering
price movements within a market. Price per statistical unit, in combination with unit
price averages, provides insight into the degree to which the average prices in a market
are changing as a result of shifts in “mix”—proportions of sales generated by differently
priced SKUs—versus price changes for individual items. Alterations in mix—such as a
relative increase in the sale of larger versus smaller ice cream tubs at retail grocers, for
example—will affect average unit price, but not price per statistical unit. Pricing changes
in the SKUs that make up a statistical unit, however, will be reflected by a change in the
price of that statistical unit.

Construction
As with other marketing averages, average price per unit can be calculated either from
company totals or from the prices and shares of individual SKUs.
Revenue ($)
Average Price per Unit ($)
Units Sales (#)
or
[Unit Price of SKU 1 ($) * SKU 1 Percentage of Sales (%)]
[Unit Price of SKU 2 ($) * SKU 2 Percentage of Sales (%)]
and so forth
The average price per unit depends on both unit prices and unit sales of individual
SKUs. The average price per unit can be driven upward by a rise in unit prices, or by an
increase in the unit shares of higher-priced SKUs, or by a combination of the two.


An “average” price metric that is not sensitive to changes in SKU shares is the price per
statistical unit.

Price per Statistical Unit
Procter & Gamble and other companies face a challenge in monitoring prices for a wide
variety of product sizes, package types, and product formulations. There are as many as
25 to 30 different SKUs for some brands, and each SKU has its own price. In these situ-
ations, how do marketers determine a brand’s overall price level in order to compare it
to competitive offerings or to track whether prices are rising or falling? One solution is
the “statistical unit,” also known as the “statistical case” or—in volumetric or weight
measures—the statistical liter or statistical ton. A statistical case of 288 ounces of liquid
detergent, for example, might be defined as comprising
Four 4-oz bottles 16 oz
Twelve 12-oz bottles 144 oz
Two 32-oz bottles 64 oz
One 64-oz bottle 64 oz
Note that the contents of this statistical case were carefully chosen so that it contains the
same number of ounces as a standard case of 24 12-ounce bottles. In this way, the sta-
tistical case is comparable in size to a standard case. The advantage of a statistical case is
that its contents can approximate the mix of SKUs the company actually sells.
Whereas a statistical case of liquid detergent will be filled with whole bottles, in other
instances a statistical unit might contain fractions of certain packaging sizes in order for
its total contents to match a required volumetric or weight total.
Statistical units are composed of fixed proportions of different SKUs. These fixed pro-
portions ensure that changes in the prices of the statistical unit reflect only changes in
the prices of the SKUs that comprise it.
The price of a statistical unit can be expressed either as a total price for the bundle of
SKUs comprising it, or in terms of that total price divided by the total volume of its con-
tents. The former might be called the “price per statistical unit”; the latter, the “unit
price per statistical unit.”

EXAMPLE: Carl’s Coffee Creamer (CCC) is sold in three sizes: a one-liter economy
size, a half-liter “fridge-friendly” package, and a 0.05-liter single serving. Carl defines a
12-liter statistical case of CCC as

Two units of the economy size 2 liters (2 * 1.0 liter)
19 units of the fridge-friendly package 9.5 liters (19 * 0.5 liter)
Ten single servings 0.5 liter (10 * .05)


Prices for each size and the calculation of total price for the statistical unit are shown in
the following table:

Number Liters in
Price of in Statistical Statistical Total
SKU Names Size Item Case Case Price

Economy 1 Liter $8.00 2 2.0 $16.00

Fridge-Friendly 0.5 Liter $6.00 19 9.5 $114.00

Single Serving 0.05 Liter $1.00 10 0.5 $10.00

TOTAL 12 $140.00

Thus, the total price of the 12-liter statistical case of CCC is $140. The per-liter price
within the statistical case is $11.67.
Note that the $140 price of the statistical case is higher than the $96 price of a case of
12 economy packs. This higher price reflects the fact that smaller packages of CCC com-
mand a higher price per liter. If the proportions of the SKUs in the statistical case exactly
match the actual proportions sold, then the per-liter price of the statistical case will match
the per-liter price of the actual liters sold.

EXAMPLE: Carl sells 10,000 one-liter economy packs of CCC, 80,000 fridge-friendly
half liters, and 40,000 single servings. What was his average price per liter?
Revenue ($)
Average Price per Unit ($)
Unit Sales (#)

($8 * 10k $6 * 80k $1 * 40k)
(1 * 10k 0.5 * 80k 0.05 * 40k)
$600k
$11.54
52k

Note that Carl’s average price per liter, at $11.54, is less than the per-liter price in his sta-
tistical case. The reason is straightforward: Whereas fridge-friendly packs outnumber
economy packs by almost ten to one in the statistical case, the actual sales ratio of these
SKUs was only eight to one. Similarly, whereas the ratio of single-serving items to econ-
omy items in the statistical case is five to one, their actual sales ratio was only four to one.
Carl’s company sold a smaller percentage of the higher (per liter) priced items than was
represented in its statistical case. Consequently, its actual average price per liter was less
than the per-liter price within its statistical unit.


In the following table, we illustrate the calculation of the average price per unit as the
weighted average of the unit prices and unit shares of the three SKUs of Carl’s Coffee
Creamer. Unit prices and unit (per-liter) shares are provided.

SKUs Units Sold Unit Price Unit
SKU Name Size Price Sold (Liters) (per Liter) Share
Economy 1 Liter $8 10k 10k $8 19.23%
Fridge-Friendly 0.5 Liter $6 80k 40k $12 76.92%
Single Serving 0.05 Liter $1 40k 2k $20 3.85%

TOTAL 130k 52k 100%

On this basis, the average price per unit ($) = ($8 * 0.1923) + ($12 * 0.7692) + ($20 *
0.0385) = $11.54.

With complex and changing product lines, and with different selling prices charged by
different retailers, marketers need to understand a number of methodologies for calcu-
lating average prices. Merely determining how many units of a product are sold, and at
what price, throughout the market is a major challenge. As a standard method of track-
ing prices, marketers use statistical units, which are based on constant proportions of
sales of different SKUs in a product line.
Typically, the proportions of SKUs in a statistical unit correspond—at least
approximately—to historical market sales. Sales patterns can change, however. In conse-
quence, these proportions need to be monitored carefully in evolving markets and
changing product lines.
Calculating a meaningful average price is complicated by the need to differentiate
between changes in sales mix and changes in the prices of statistical units. In some
industries, it is difficult to construct appropriate units for analyzing price and sales data.
In the chemical industry, for example, an herbicide might be sold in a variety of differ-
ent sizes, applicators, and concentration levels. When we factor in the complexity of dif-
ferent prices and different assortments offered by competing retail outlets, calculating
and tracking average prices becomes a non-trivial exercise.
Similar challenges arise in estimating inflation. Economists calculate inflation by using
a basket of goods. Their estimates might vary considerably, depending on the goods
included. It is also difficult to capture quality improvements in inflation figures. Is a
2009 car, for example, truly comparable to a car built 30 years earlier?
In evaluating price increases, marketers are advised to bear in mind that a consumer
who shops for large quantities at discount stores may view such increases very


differently from a pensioner who buys small quantities at local stores. Establishing a
“standard” basket for such different consumers requires astute judgment. In seeking to
summarize the aggregate of such price increases throughout an economy, economists
may view inflation as, in effect, a statistical unit price measure for that economy.

3.4 Variable Costs and Fixed Costs
Variable costs can be aggregated into a “total” or expressed on a “per-unit” basis. Fixed
costs, by definition, do not change with the number of units sold or produced.
Variable costs are assumed to be relatively constant on a per-unit basis. Total variable
costs increase directly and predictably with unit sales volume. Fixed costs, on the other
hand, do not change as a direct result of short-term unit sales increases or decreases.
Total Costs ($) Fixed Costs ($) Total Variable Costs ($)
Total Variable Costs ($) Unit Volume (#) * Variable Cost per Unit ($)
Marketers need to have an idea of how costs divide between variable and fixed. This
distinction is crucial in forecasting the earnings generated by various changes in unit
sales and thus the financial impact of proposed marketing campaigns. It is also fun-
damental to an understanding of price and volume trade-offs.

Purpose: To understand how costs change with volume.
At first glance, this appears to be an easy subject to master. If a marketing campaign will
generate 10,000 units of additional sales, we need only know how much it will cost to
supply that additional volume.
The problem, of course, is that no one really knows how changes in quantity will affect a
firm’s total costs—in part because the workings of a firm can be so complex. Companies
simply can’t afford to employ armies of accountants to answer every possible expense
question precisely. Instead, we often use a simple model of cost behavior that is good
enough for most purposes.

Construction
The standard linear equation, Y mX b, helps explain the relationship between total
costs and unit volume. In this application, Y will represent a company’s total cost, m will
be its variable cost per unit, X will represent the quantity of products sold (or pro-
duced), and b will represent the fixed cost (see Figure 3.3).

Total Cost ($) Variable Cost per Unit ($) * Quantity (#) Fixed Cost ($)


Fixed and Variable Costs
Total Cost $ (Y)

Effect of adding 1 unit -
variable cost (m)

} Fixed Cost (b)

Quantity (X)

Figure 3.3 Fixed and Variable Costs

On this basis, to determine a company’s total cost for any given quantity of products, we
need only multiply its variable cost per unit by that quantity and add its fixed cost.
To communicate fully the implications of fixed costs and variable costs, it may help to
separate this graph into two parts (see Figure 3.4).
By definition, fixed costs remain the same, regardless of volume. Consequently, they are
represented by a horizontal line across the graph in Figure 3.4. Fixed costs do not
increase vertically—that is, they do not add to the total cost—as quantity rises.
The result of multiplying variable cost per unit by quantity is often called the total vari-
able cost. Variable costs differ from fixed costs in that, when there is no production, their
total is zero. Their total increases in a steadily rising line, however, as quantity increases.
We can represent this model of cost behavior in a simple equation.
Total Cost ($) Total Variable Cost ($) Fixed Cost ($)


Fixed Costs

Total Cost $ (Y)

Fixed Cost (b)

Quantity (X)

Variable Costs
Total Cost $ (Y)

Total Variable Cost (m*X)

Quantity (X)

Figure 3.4 Total Cost Consists of Fixed and Variable Costs


To use this model, of course, we must place each of a firm’s costs into one or the other
of these two categories. If an expense does not change with volume (rent, for example),
then it is part of fixed costs and will remain the same, regardless of how many units the
firm produces or sells. If a cost does change with volume (sales commissions, for exam-
ple), then it is a variable cost.
Total Variable Costs ($) = Unit Volume (#) * Variable Cost per Unit ($)
Total Cost per Unit: It is also possible to express the total cost for a given quantity on a
per-unit basis. The result might be called total cost per unit, unit total cost, average cost,
full cost, or even fully loaded cost. For our simple linear cost model, the total cost per
unit can be calculated in either of two ways. The most obvious would be to divide the
total cost by the number of units.
Total Cost ($)
Total Cost per Unit ($) =
Quantity (#)

This can be plotted graphically, and it tells an interesting tale (see Figure 3.5). As the
quantity rises, the total cost per unit (average cost per unit) declines. The shape of this

Effects of Fixed and Variable Costs on Cost per Unit
Unit Total Cost ($)

Quantity

Figure 3.5 Total Cost per Unit Falls with Volume (Typical Assumptions)


curve will vary among firms with different cost structures, but wherever there are both
fixed and variable costs, the total cost per unit will decline as fixed costs are spread
across an increasing quantity of units.
The apportionment of fixed costs across units produced leads us to another common
formula for the total cost per unit.

Total Cost per Unit ($) = Variable Cost per Unit ($) + [Fixed Cost ($)/Quantity (#)]

As the quantity increases—that is, as fixed costs are spread over an increasing number of
units—the total cost per unit declines in a non-linear way.3

EXAMPLE: As a company’s unit sales increase, its fixed costs hold steady at $500. The
variable cost per unit remains constant at $10 per unit. Total variable costs increase with
each unit sold. The total cost per unit (also known as average total cost) decreases as
incremental units are sold and as fixed costs are spread across this rising quantity.
Eventually, as more and more units are produced and sold, the company’s total cost per
unit approaches its variable cost per unit (see Table 3.8).

Table 3.8 Fixed and Variable Costs at Increasing Volume Levels

Units Sold 1 10 100 1,000
Fixed Costs $500 $500 $500 $500
Variable Costs $10 $100 $1,000 $10,000
Total Costs $510 $600 $1,500 $10,500

Total Cost per Unit $510.00 $60.00 $15.00 $10.50
Variable Cost per Unit $10 $10 $10 $10

In summary, the simplest model of cost behavior is to assume total costs increase lin-
early with quantity supplied. Total costs are composed of fixed and variable costs. Total
cost per unit decreases in a non-linear way with rising quantity supplied.

Total cost is typically assumed to be a linear function of quantity supplied. That is, the
graph of total cost versus quantity will be a straight line. Because some costs are fixed,
total cost starts at a level above zero, even when no units are produced. This is because


fixed costs include such expenses as factory rent and salaries for full-time employees,
which must be paid regardless of whether any goods are produced and sold. Total
variable costs, by contrast, rise and fall with quantity. Within our model, however, vari-
able cost per unit is assumed to hold constant—at $10 per unit for example—regardless
of whether one unit or 1,000 units are produced. This is a useful model. In using it,
however, marketers must recognize that it fails to account for certain complexities.
The linear cost model does not fit every situation: Quantity discounts, expectations of
future process improvements, and capacity limitations, for example, introduce dynamics
that will limit the usefulness of the fundamental linear cost equation: Total Cost Fixed
Cost Variable Cost per Unit * Quantity. Even the notion that quantity determines the
total cost can be questioned. Although firms pay for inputs, such as raw materials and
labor, marketers want to know the cost of the firm’s outputs, that is, finished goods sold.
This distinction is clear in theory. In practice, however, it can be difficult to uncover the
precise relationship between a quantity of outputs and the total cost of the wide array of
inputs that go into it.
The classification of costs as fixed or variable depends on context: Even though the
linear model may not work in all situations, it does provide a reasonable approximation
for cost behavior in many contexts. Some marketers have trouble, however, with the fact
that certain costs can be considered fixed in some contexts and variable in others. In
general, for shorter time frames and modest changes in quantity, many costs are fixed.
For longer time frames and larger changes in quantity, most costs are variable. Let’s con-
sider rent, for example. Small changes in quantity do not require a change in workspace
or business location. In such cases, rent should be regarded as a fixed cost. A major
change in quantity, however, would require more or less workspace. Rent, therefore,
would become variable over that range of quantity.
Don’t confuse Total Cost per Unit with Variable Cost per Unit: In our linear cost
equation, the variable cost per unit is the amount by which total costs increase if the
firm increases its quantity by one unit. This number should not be confused with the
total cost per unit, calculated as Variable Cost per Unit (Fixed Cost/Quantity). If a firm
has fixed costs, then its total cost per unit will always be greater than the variable cost
per unit. Total cost per unit represents the firm’s average cost per unit at the current
quantity—and only at the current quantity. Do not make the mistake of thinking of
total cost per unit as a figure that applies to changing quantities. Total cost per unit only
applies at the volume at which it was calculated.
A related misunderstanding may arise at times from the fact that total cost per unit gen-
erally decreases with rising quantity. Some marketers use this fact to argue for aggres-
sively increasing quantity in order to “bring our costs down” and improve profitability.
Total cost, by contrast with total cost per unit, almost always increases with quantity.
Only with certain quantity discounts or rebates that “kick in” when target volumes are
reached can total cost decrease as volume increases.


3.5 Marketing Spending—Total, Fixed, and Variable
To predict how selling costs change with sales, a firm must distinguish between fixed
selling costs and variable selling costs.

Total Selling (Marketing) Costs ($) = Total Fixed Selling Costs ($)
Total Variable Selling Costs ($)
Total Variable Selling Costs ($) = Revenue ($) * Variable Selling Cost (%)
Recognizing the difference between fixed and variable selling costs can help firms
account for the relative risks associated with alternative sales strategies. In general,
strategies that incur variable selling costs are less risky because variable selling costs
will remain lower in the event that sales fail to meet expectations.

Purpose: To forecast marketing spending and assess budgeting risk.
Marketing Spending: Total expenditure on marketing activities. This typically
includes advertising and non-price promotion. It sometimes includes sales force
spending and may also include price promotions.
Marketing costs are often a major part of a firm’s overall discretionary expenditures. As
such, they are important determinants of short-term profits. Of course, marketing and
selling budgets can also be viewed as investments in acquiring and maintaining cus-
tomers. From either perspective, however, it is useful to distinguish between fixed mar-
keting costs and variable marketing costs. That is, managers must recognize which
marketing costs will hold steady, and which will change with sales. Generally, this classi-
fication will require a “line-item by line-item” review of the entire marketing budget.
In prior sections, we have viewed total variable costs as expenses that vary with unit
sales volume. With respect to selling costs, we’ll need a slightly different conception.
Rather than varying with unit sales, total variable selling costs are more likely to vary
directly with the monetary value of the units sold—that is, with revenue. Thus, it is
more likely that variable selling costs will be expressed as a percentage of revenue, rather
than a certain monetary amount per unit.
The classification of selling costs as fixed or variable will depend on an organization’s
structure and on the specific decisions of management. A number of items, however,
typically fall into one category or the other—with the proviso that their status as fixed
or variable can be time-specific. In the long run, all costs eventually become variable.
Over typical planning periods of a quarter or a year, fixed marketing costs might include
■ Sales force salaries and support.
■ Major advertising campaigns, including production costs.


■ Marketing staff.
■ Sales promotion material, such as point-of-purchase sales aids, coupon produc-
tion, and distribution costs.
■ Cooperative advertising allowances based on prior-period sales.
Variable marketing costs might include
■ Sales commissions paid to sales force, brokers, or manufacturer representatives.
■ Sales bonuses contingent on reaching sales goals.
■ Off-invoice and performance allowances to trade, which are tied to current
volume.
■ Early payment terms (if included in sales promotion budgets).
■ Coupon face-value payments and rebates, including processing fees.
■ Bill-backs for local campaigns, which are conducted by retailers but reimbursed
by national brand and cooperative advertising allowances, based on current
period sales.
Marketers often don’t consider their budgets in fixed and variable terms, but they can
derive at least two benefits by doing so.
First, if marketing spending is in fact variable, then budgeting in this way is more accurate.
Some marketers budget a fixed amount and then face an end-of-period discrepancy or
“variance” if sales miss their declared targets. By contrast, a flexible budget—that is, one
that takes account of its genuinely variable components—will reflect actual results, regard-
less of where sales end up.
Second, the short-term risks associated with fixed marketing costs are greater than those
associated with variable marketing costs. If marketers expect revenues to be sensitive to
factors outside their control—such as competitive actions or production shortages—
they can reduce risk by including more variable and less fixed spending in their budgets.
A classic decision that hinges on fixed marketing costs versus variable marketing costs is
the choice between engaging third-party contract sales representatives versus an in-
house sales force. Hiring a salaried—or predominantly salaried—sales force entails
more risk than the alternative because salaries must be paid even if the firm fails to
achieve its revenue targets. By contrast, when a firm uses third-party brokers to sell its
goods on commission, its selling costs decline when sales targets are not met.

Construction
Total Selling (Marketing) Costs ($) Total Fixed Selling Costs ($)
Total Variable Selling Costs ($)
Total Variable Selling Costs ($) Revenue ($) * Variable Selling Cost (%)


Commissioned Sales Costs: Sales commissions represent one example of selling costs
that vary in proportion to revenue. Consequently, any sales commissions should be
included in variable selling costs.

EXAMPLE: Henry’s Catsup spends $10 million a year to maintain a sales force that
calls on grocery chains and wholesalers. A broker offers to perform the same selling tasks
for a 5% commission.
At $100 million in revenue,
Total Variable Selling Cost $100 million * 5% $5 million
If revenues run less than $200 million, the broker will cost less than the in-house sales
force. At $200 million in revenue, the broker will cost the same as the sales force. At rev-
enue levels greater than $200 million, the broker will cost more.
Of course, the transition from a salaried sales staff to a broker may itself cause a change
in revenues. Calculating the revenue level at which selling costs are equal is only a starting
point for analysis. But it is an important first step in understanding the trade-offs.

There are many types of variable selling costs. For example, selling costs could be based
upon a complicated formula, specified in a firm’s contracts with its brokers and dealers.
Selling costs might include incentives to local dealers, which are tied to the achievement
of specific sales targets. They might include promises to reimburse retailers for spend-
ing on cooperative advertising. By contrast, payments to a Web site for a fixed number
of impressions or click-throughs, in a contract that calls for specific dollar compensa-
tion, would more likely be classified as fixed costs. On the other hand, payments for con-
versions (sales) would be classified as variable marketing costs.

EXAMPLE: A small manufacturer of a regional food delicacy must select a budget for
a television advertising campaign that it plans to launch. Under one plan, it might pay to
create a commercial and air it in a certain number of time slots. Its spending level would
thus be fixed. It would be selected ahead of time and would not vary with the results of
the campaign.
Under an alternative plan, the company could produce the advertisement—still a fixed
cost—but ask retailers to air it in their local markets and pay the required media fees to


television stations as part of a cooperative advertising arrangement. In return for paying
the media fees, local stores would receive a discount (a bill-back) on every unit of the
company’s product that they sell.
Under the latter plan, the product discount would be a variable cost, as its total amount
would depend on the number of units sold. By undertaking such a cooperative advertising
campaign, the manufacturer would make its marketing budget a mix of fixed and variable
costs. Is such cooperative advertising a good idea? To decide this, the company must
determine its expected sales under both arrangements, as well as the consequent
economics and its tolerance for risk.

Fixed costs are often easier to measure than variable costs. Typically, fixed costs might be
assembled from payroll records, lease documents, or financial records. For variable
costs, it is necessary to measure the rate at which they increase as a function of activity
level. Although variable selling costs often represent a predefined percentage of revenue,
they may alternatively vary with the number of units sold (as in a dollar-per-case dis-
count). An additional complication arises if some variable selling costs apply to only a
portion of total sales. This can happen, for example, when some dealers qualify for cash
discounts or full-truckload rates and some do not.
In a further complication, some expenses may appear to be fixed when they are actually
stepped. That is, they are fixed to a point, but they trigger further expenditures beyond
that point. For example, a firm may contract with an advertising agency for up to three
campaigns per year. If it decides to buy more than three campaigns, it would incur an
incremental cost. Typically, stepped costs can be treated as fixed—provided that the
boundaries of analysis are well understood.
Stepped payments can be difficult to model. Rebates for customers whose purchases
exceed a certain level, or bonuses for salespeople who exceed quota, can be challenging
functions to describe. Creativity is important in designing marketing discounts. But this
creativity can be difficult to reflect in a framework of fixed and variable costs.
In developing their marketing budgets, firms must decide which costs to expense in the
current period and which to amortize over several periods. The latter course is appropri-
ate for expenditures that are correctly viewed as investments. One example of such an
investment would be a special allowance for financing receivables from new distributors.
Rather than adding such an allowance to the current period’s budget, it would be better
viewed as a marketing item that increases the firm’s investment in working capital. By
contrast, advertising that is projected to generate long-term impact may be loosely called
an investment, but it would be better treated as a marketing expense. Although there may
be a valid theoretical case for amortizing advertising, that discussion is beyond the scope
of this book.


Levels of marketing spending are often used to compare companies and to demonstrate
how heavily they “invest” in this area. For this purpose, marketing spending is generally
viewed as a percentage of sales.
Marketing As a Percentage of Sales: The level of marketing spending as a fraction
of sales. This figure provides an indication of how heavily a company is marketing.
The appropriate level for this figure varies among products, strategies, and markets.
Marketing Spending ($)
Marketing As a Percentage of Sales (%) =
Revenue ($)

Variants on this metric are used to examine components of marketing in comparison
with sales. Examples include trade promotion as a percentage of sales, or sales force as a
percentage of sales. One particularly common example is:
Advertising As a Percentage of Sales: Advertising expenditures as a fraction of
sales. Generally, this is a subset of marketing as a percentage of sales.
Before using such metrics, marketers are advised to determine whether certain market-
ing costs have already been subtracted in the calculation of sales revenue. Trade
allowances, for example, are often deducted from “gross sales” to calculate “net sales.”
Slotting Allowances: These are a particular form of selling costs encountered when new
items are introduced to retailers or distributors. Essentially, they represent a charge
made by retailers for making a “slot” available for a new item in their stores and
warehouses. This charge may take the form of a one-time cash payment, free goods, or
a special discount. The exact terms of the slotting allowance will determine whether it
constitutes a fixed or a variable selling cost, or a mix of the two.

3.6 Break-Even Analysis and Contribution Analysis
The break-even level represents the sales amount—in either unit or revenue terms—
that is required to cover total costs (both fixed and variable). Profit at break-even is
zero. Break-even is only possible if a firm’s prices are higher than its variable costs
per unit. If so, then each unit of product sold will generate some “contribution”
toward covering fixed costs. The difference between price per unit and variable cost
per unit is defined as Contribution per Unit.
Contribution per Unit ($) Selling Price per Unit ($) Variable Cost per Unit ($)4
Contribution per Unit ($)
Contribution Margin (%)


Fixed Costs ($)
Break-Even Volume (#) =

Break-Even Revenue ($) = Break-Even Volume (Units) (#) * Price per Unit ($)

or
Fixed Costs ($)
=

Break-even analysis is the Swiss Army knife of marketing economics. It is useful in a
variety of situations and is often used to evaluate the likely profitability of marketing
actions that affect fixed costs, prices, or variable costs per unit. Break-even is often
derived in a “back-of-the-envelope” calculation that determines whether a more
detailed analysis is warranted.

Purpose: To provide a rough indicator of the earnings impact
of a marketing activity.
The break-even point for any business activity is defined as the level of sales at which nei-
ther a profit nor a loss is made on that activity—that is, where Total Revenues Total
Costs. Provided that a company sells its goods at a price per unit that is greater than its
variable cost per unit, the sale of each unit will make a “contribution” toward covering
some portion of fixed costs. That contribution can be calculated as the difference
between price per unit (revenue) and variable cost per unit. On this basis, break-even
constitutes the minimum level of sales at which total contribution fully covers fixed costs.

Construction
To determine the break-even point for a business program, one must first calculate the
fixed costs of engaging in that program. For this purpose, managers do not need to esti-
mate projected volumes. Fixed costs are constant, regardless of activity level. Managers
do, however, need to calculate the difference between revenue per unit and variable costs
per unit. This difference represents contribution per unit ($). Contribution rates can
also be expressed as a percentage of selling price.

EXAMPLE: Apprentice Mousetraps wants to know how many units of its “Magic
Mouse Trapper” it must sell to break even. The product sells for $20. It costs $5 per unit
to make. The company’s fixed costs are $30,000. Break-even will be reached when total
contribution equals fixed costs.


Fixed Costs
Break-Even Volume
Contribution per Unit
Contribution per Unit Sale Price per Unit Variable Cost per Unit
$20 $5 $15
$30,000
Break-Even Volume 2,000 mousetraps
$15

This dynamic can be summarized in a graph that shows fixed costs, variable costs, total
costs, and total revenue (see Figure 3.6). Below the break-even point, total costs exceed
total revenue, creating a loss. Above the break-even point, a company generates profits.
Break-Even: Break-even occurs when the total contribution equals the fixed costs.
Profits and losses at this point equal zero.
One of the key building blocks of break-even analysis is the concept of contribution.
Contribution represents the portion of sales revenue that is not consumed by variable
costs and so contributes to the coverage of fixed costs.
Contribution per Unit ($) Selling Price per Unit ($) Variable Cost per Unit ($)

The Break-Even Point Profit is earned when
Total Revenue exceeds
Total Cost

Break-Even

Below Break-Even, Total Costs
Total Costs exceed
$

Total Revenue
Total Revenue

Fixed Costs
Total Variable
Costs

Units

Figure 3.6 At Break-Even, Total Costs Total Revenues


Contribution can also be expressed in percentage terms, quantifying the fraction of the
sales price that contributes toward covering fixed costs. This percentage is often called
the contribution margin.
Contribution Margin (%) =

Formulas for total contribution include the following:
Total Contribution ($) Units Sold (#) * Contribution per Unit ($)
Total Contribution ($) Total Revenues ($) Total Variable Costs ($)
As previously noted,
Total Variable Costs Variable Costs per Unit * Units Sold
Total Revenues Selling Price per Unit * Units Sold
Break-Even Volume: The number of units that must be sold to cover fixed costs.
Fixed Costs ($)
Break-Even Volume (#) =

Break-even will occur when an enterprise sells enough units to cover its fixed costs. If
the fixed costs are $10 and the contribution per unit is $2, then a firm must sell five
units to break even.
Break-Even Revenue: The level of dollar sales required to break even.
Break-Even Revenue ($) Break-Even Volume (Units) (#) * Price per Unit ($)
This formula is the simple conversion of volume in units to the revenues generated by
that volume.

EXAMPLE: Apprentice Mousetraps wants to know how many dollars’ worth of its
“Deluxe Mighty Mouse Trapper” it must sell to break even. The product sells for $40 per
unit. It costs $10 per unit to make. The company’s fixed costs are $30,000.
With fixed costs of $30,000, and a contribution per unit of $30, Apprentice must sell
$30,000/$30 1,000 deluxe mousetraps to break even. At $40 per trap, this corresponds
to revenues of 1,000 * $40 $40,000.
Break-Even Revenue ($) Break-Even Volume (#) * Price per Unit ($)
1,000 * $40 $40,000
Break-even in dollar terms can also be calculated by dividing fixed costs by the fraction
of the selling price that represents contribution.


Fixed Costs
Break-Even Revenue
[(Selling Price Variable Costs)/Selling Price]
$30,000
[($40 $10)/$40]
$30,000
$40,000
75%

BREAK-EVEN ON INCREMENTAL INVESTMENT
Break-even on incremental investment is a common form of break-even analysis. It
examines the additional investment needed to pursue a marketing plan, and it calculates
the additional sales required to cover that expenditure. Any costs or revenues that would
have occurred regardless of the investment decision are excluded from this analysis.

EXAMPLE: John’s Clothing Store employs three salespeople. It generates annual sales
of $1 million and an average contribution margin of 30%. Rent is $50,000. Each sales
person costs $50,000 per year in salary and benefits. How much would sales have to
increase for John to break even on hiring an additional salesperson?
If the additional “investment” in a salesperson is $50,000, then break-even on the new
hire will be reached when sales increase by $50,000 / 30%, or $166,666.67.

To calculate a break-even sales level, one must know the revenues per unit, the variable
costs per unit, and the fixed costs. To establish these figures, one must classify all costs as
either fixed (those that do not change with volume) or variable (those that increase lin-
early with volume).
The time scale of the analysis can influence this classification. Indeed, one’s managerial
intent can be reflected in the classification. (Will the company fire employees and sub-
let factory space if sales turn down?) As a general rule, all costs become variable in the
long term. Firms generally view rent, for example, as a fixed cost. But in the long term,
even rent becomes variable as a company may move into larger quarters when sales
grow beyond a certain point.
Before agonizing over these judgments, managers are urged to remember that the most
useful application of the break-even exercise is to make a rough judgment about
whether more detailed analyses are likely to be worth the effort. The break-even calcu-
lation enables managers to judge various options and proposals quickly. It is not, how-
ever, a substitute for more detailed analyses, including projections of target profits
(Section 3.7), risk, and the time value of money (Sections 5.3 and 10.4).


Payback Period: The period of time required to recoup the funds expended in an
investment. The payback period is the time required for an investment to reach
break-even (see previous sections).

3.7 Profit-Based Sales Targets
In launching a program, managers often start with an idea of the dollar profit they
desire and ask what sales levels will be required to reach it. Target volume (#) is the
unit sales quantity required to meet an earnings goal. Target revenue ($) is the corre-
sponding figure for dollar sales. Both of these metrics can be viewed as extensions of
break-even analysis.
[Fixed Costs ($) Target Profits ($)]
Target Volume (#)
Target Revenue ($) Target Volume (#) * Selling Price per Unit ($)
or
Increasingly, marketers are expected to generate volumes that meet the target profits of
their firm. This will often require them to revise sales targets as prices and costs change.

Purpose: To ensure that marketing and sales objectives mesh
with profit targets.
In the previous section, we explored the concept of break-even, the point at which a
company sells enough to cover its fixed costs. In target volume and target revenue cal-
culations, managers take the next step. They determine the level of unit sales or revenues
needed not only to cover a firm’s costs but also to attain its profit targets.

Construction
Target Volume: The volume of sales necessary to generate the profits specified in a
company’s plans.
The formula for target volume will be familiar to those who have performed break-even
analysis. The only change is to add the required profit target to the fixed costs. From
another perspective, the break-even volume equation can be viewed as a special case of


the general target volume calculation—one in which the profit target is zero, and a com-
pany seeks only to cover its fixed costs. In target volume calculations, the company
broadens this objective to solve for a desired profit.
[Fixed Costs ($) + Target Profits ($)]
Target Volume (#)

EXAMPLE: Mohan, an artist, wants to know how many caricatures he must sell to
realize a yearly profit objective of $30,000. Each caricature sells for $20 and costs $5 in
materials to make. The fixed costs for Mohan’s studio are $30,000 per year:
(Fixed Costs Target Profits)
Target Volume
(Sales Price Variable Costs)

($30,000 $30,000)
($20 $5)

4,000 caricatures per year
It is quite simple to convert unit target volume to target revenues. One need only multiply
the volume figure by an item’s price per unit. Continuing the example of Mohan’s studio,
Target Revenue ($) Target Volume (#) * Selling Price ($)
4,000 * $20 $80,000
Alternatively, we can use a second formula:
Target Revenue

($30,000 $30,000)
($15/$20)
$60,000
$80,000
0.75

The information needed to perform a target volume calculation is essentially the same
as that required for break-even analysis—fixed costs, selling price, and variable costs. Of
course, before determining target volume, one must also set a profit target.
The major assumption here is the same as in break-even analysis: Costs are linear with
respect to unit volume over the range explored in the calculation.


Target Volumes not based on Target Profit: In this section, we have assumed that a firm
starts with a profit target and seeks to determine the volume required to meet it. In cer-
tain instances, however, a firm might set a volume target for reasons other than short-
term profit. For example, firms sometimes adopt top-line growth as a goal. Please do
not confuse this use of target volume with the profit-based target volumes calculated in
this section.
Returns and Targets: Companies often set hurdle rates for return on sales and return on
investment and require that projections achieve these before any plan can be approved.
Given these targets, we can calculate the sales volume required for the necessary return.
(See Section 10.2 for more details.)

EXAMPLE: Niesha runs business development at Gird, a company that has estab-
lished a return on sales target of 15%. That is, Gird requires that all programs generate
profits equivalent to 15% of sales revenues. Niesha is evaluating a program that will add
$1,000,000 to fixed costs. Under this program, each unit of product will be sold for $100
and will generate a contribution margin of 25%. To reach break-even on this program,
Gird must sell $1,000,000/$25 40,000 units of product. How much must Gird sell to
reach its target return on sales (ROS) of 15%?
To determine the revenue level required to achieve a 15% ROS, Niesha can use either a
spreadsheet model and trial and error, or the following formula:
Fixed Costs ($)
Target Revenue
[Contribution Margin (%) Target ROS (%)]

$1,000,000
(0.25 0.15)

$1,000,000
$10,000,000
0.1

Thus, Gird will achieve its 15% ROS target if it generates $10,000,000 in sales. At a selling
price of $100 per unit, this is equivalent to unit sales of 100,000.


4
PRODUCT AND PORTFOLIO
MANAGEMENT

Introduction

Trial, Repeat, Penetration, Conjoint Utilities and Consumer
and Volume Projections Preference
Growth: Percentage and CAGR Segmentation and Conjoint
Utilities
Cannibalization Rate and Fair
Share Draw Rate Conjoint Utilities and Volume
Projection
Brand Equity Metrics

Effective marketing comes from customer knowledge and an understanding of how a
product fits customers’ needs. In this chapter, we’ll describe metrics used in product
strategy and planning. These metrics address the following questions: What volumes
can marketers expect from a new product? How will sales of existing products be
affected by the launch of a new offering? Is brand equity increasing or decreasing? What
do customers really want, and what are they willing to sacrifice to obtain it?
We’ll start with a section on trial and repeat rates, explaining how these metrics are
determined and how they’re used to generate sales forecasts for new products. Because
forecasts involve growth projections, we’ll then discuss the difference between year-
on-year growth and compound annual growth rates (CAGR). Because growth of one
product sometimes comes at the expense of an existing product line, it is important to

109

understand cannibalization metrics. These reflect the impact of new products on a port-
folio of existing products.
Next, we’ll cover selected metrics associated with brand equity—a central focus of mar-
keting. Indeed, many of the metrics throughout this book can be useful in evaluating
brand equity. Certain metrics, however, have been developed specifically to measure the
“health” of brands. This chapter will discuss them.
Although branding strategy is a major aspect of a product offering, there are others, and
managers must be prepared to make trade-offs among them, informed by a sense of the
“worth” of various features. Conjoint analysis helps identify customers’ valuation of
specific product attributes. Increasingly, this technique is used to improve products and
to help marketers evaluate and segment new or rapidly growing markets. In the final
sections of this chapter, we’ll discuss conjoint analysis from multiple perspectives.


4.1 Trial First-time users Distinguish “ever- Over time, sales
as a percentage tried” from “new” should rely less on
of the target triers in current trial and more on
population. period. repeat purchasers.
4.1 Repeat Volume Repeat buyers, Depending on Measure of the
multiplied by the when trial was stability of a
number of prod- achieved, not all brand franchise.
ucts they buy in triers will have an
each purchase, equal opportunity
multiplied by the to make repeat
number of times purchases.
they purchase per
period.
4.1 Penetration Users in the previ- The length of the Measure of the
ous period, multi- period will affect population buying
plied by repeat norms, that is, in the current
rate for the cur- more customers period.
rent period, plus buy in a year than
new triers in the in a month.
current period.
4.1 Volume Combine trial vol- Adjust trial and Plan production
Projections ume and repeat repeat rates for and inventories
volume. time frame. Not for both trade
all triers will have sales and con-
time or opportu- sumer off-take.
nity to repeat.


4.2 Year-on-Year Percentage Distinguish unit Plan production
Growth change from one and dollar growth and budgeting.
year to the next. rates.
4.2 Compound Ending value May not reflect Useful for averag-
Annual Growth divided by start- individual year- ing growth rates
Rate (CAGR) ing value to the on-year growth over long periods.
power of 1/N, in rates.
which N is the
number of
periods.
4.3 Cannibalization Percentage of new Market expansion Useful to account
Rate product sales effects should also for the fact that
taken from exist- be considered. new products
ing product line. often reduce the
sales of existing
products.
4.3 Fair Share Draw Assumption that May not be a rea- Useful to generate
new entrants in a sonable assump- an estimate of
market capture tion if there are sales and shares
sales from estab- significant differ- after entry of new
lished competitors ences among competitor.
in proportion to competing
established mar- brands.
ket shares.
4.4 Brand Equity Numerous Metrics tracking Monitor health of
Metrics measures, for essence of brand a brand. Diagnose
example, Conjoint may not track weaknesses, as
utility attributed health and value. needed.
to brand.
4.5 Conjoint Utilities Regression coeffi- May be function Indicates the rela-
cients for attrib- of number, level, tive values that
ute levels derived and type of attrib- customers place
from conjoint utes in study. on attributes of
analysis. which product
offerings are
composed.
Continues

Chapter 4 Product and Portfolio Management 111


4.6 Segment Utilities Clustering of May be function Uses customer
individuals into of number, level, valuations of
market segments and type of product attributes
on the basis of attributes in con- to help define
sum-of-squares joint study. market segments.
distance between Assumes homo-
regression coeffi- geneity within
cients drawn segments.
from conjoint
analysis.
4.7 Conjoint Utilities Used within Assumes aware- Forecast sales
and Volume conjoint simula- ness and distribu- for alternative
Projection tor to estimate tion levels are products,
volume. known or can be designs, prices,
estimated. and branding
strategies.

4.1 Trial, Repeat, Penetration, and Volume Projections
Test markets and volume projections enable marketers to forecast sales by sampling
customer intentions through surveys and market studies. By estimating how many
customers will try a new product, and how often they’ll make repeat purchases,
marketers can establish the basis for such projections.
First-time Triers in Period t (#)
Trial Rate (%)

First-time Triers in Period t (#) = Total Population (#) * Trial Rate (%)
Penetration t (#) = [Penetration in t-1 (#) * Repeat Rate Period t (%)]
Projection of Sales t (#) Penetration t (#) * Average Frequency of Purchase (#)
* Average Units per Purchase (#)
Projections from customer surveys are especially useful in the early stages of product
development and in setting the timing for product launch. Through such projec-
tions, customer response can be estimated without the expense of a full product
launch.


Purpose: To understand volume projections.
When projecting sales for relatively new products, marketers typically use a system of
trial and repeat calculations to anticipate sales in future periods. This works on the prin-
ciple that everyone buying the product will either be a new customer (a “trier”) or a
repeat customer. By adding new and repeat customers in any period, we can establish
the penetration of a product in the marketplace.
It is challenging, however, to project sales to a large population on the basis of simulated
test markets, or even full-fledged regional rollouts. Marketers have developed various
solutions to increase the speed and reduce the cost of test marketing, such as stocking a
store with products (or mock-ups of new products) or giving customers money to buy
the products of their choice. These simulate real shopping conditions but require spe-
cific models to estimate full-market volume on the basis of test results. To illustrate the
conceptual underpinnings of this process, we offer a general model for making volume
projection on the basis of test market results.

Construction
The penetration of a product in a future period can be estimated on the basis of popu-
lation size, trial rates, and repeat rates.
Trial Rate (%): The percentage of a defined population that purchases or uses a
product for the first time in a given period.

EXAMPLE: A cable TV company keeps careful records of the names and addresses of its
customers. The firm’s vice president of marketing notes that 150 households made first-time
use of his company’s services in March 2009. The company has access to 30,000 households. To
calculate the trial rate for March, we can divide 150 by 30,000, yielding 0.5%.

First-time Triers in Period t (#): The number of customers who purchase or use a
product or brand for the first time in a given period.
Penetration t (#) [Penetration in t-1 (#) * Repeat Rate Period t (%)]

EXAMPLE: A cable TV company started selling a monthly sports package in January.
The company typically has an 80% repeat rate and anticipates that this will continue for
the new offering. The company sold 10,000 sports packages in January. In February, it
expects to add 3,000 customers for the package. On this basis, we can calculate expected
penetration for the sports package in February.


Penetration in February (Penetration January * Repeat Rate)
First-time Triers in February
(10,000 * 80%) 3,000 11,000
Later that year, in September, the company has 20,000 subscribers. Its repeat rate remains
80%. The company had 18,000 subscribers in August. Management wants to know how
many new customers the firm added for its sports package in September:
First-time Triers Penetration Repeat Customers
20,000 (18,000 * 80%) = 5,600

From penetration, it is a short step to projections of sales.
Projection of Sales (#) Penetration (#) * Frequency of Purchase (#)
* Units per Purchase (#)

Simulated Test Market Results and Volume Projections
TRIAL VOLUME
Trial rates are often estimated on the basis of surveys of potential customers. Typically,
these surveys ask respondents whether they will “definitely” or “probably” buy a prod-
uct. As these are the strongest of several possible responses to questions of purchase
intentions, they are sometimes referred to as the “top two boxes.” The less favorable
responses in a standard five-choice survey include “may or may not buy,” “probably
won’t buy,” and “definitely won’t buy.” (Refer to Section 2.7 for more on intention to
purchase.)
Because not all respondents follow through on their declared purchase intentions, firms
often make adjustments to the percentages in the top two boxes in developing sales pro-
jections. For example, some marketers estimate that 80% of respondents who say they’ll
“definitely buy” and 30% of those who say that they’ll “probably buy” will in fact pur-
chase a product when given the opportunity.1 (The adjustment for customers following
through is used in the following model.) Although some respondents in the bottom
three boxes might buy a product, their number is assumed to be insignificant. By reduc-
ing the score for the top two boxes, marketers derive a more realistic estimate of the
number of potential customers who will try a product, given the right circumstances.
Those circumstances are often shaped by product awareness and availability.
Awareness: Sales projection models include an adjustment for lack of awareness of a
product within the target market (see Figure 4.1). Lack of awareness reduces the trial
rate because it excludes some potential customers who might try the product but don’t


First-Time Use Repeat Use

Customer Survey

“Deﬁnitely Buy” “Probably Buy”

Adjustment for Customers
Following Through

Adjustment for Awareness
and Distribution

Trial Population Estimated Repeat
Repeat
Rate
Volume Each
Purchase
Volume Each
Purchase Frequency of
Repeat
Purchase

Trial Volume + Repeat Volume

= Total Volume

Figure 4.1 Schematic of Simulated Test Market Volume Projection

know about it. By contrast, if awareness is 100%, then all potential customers know
about the product, and no potential sales are lost due to lack of awareness.
Distribution: Another adjustment to test market trial rates is usually applied—
accounting for the estimated availability of the new product. Even survey respondents
who say they’ll “definitely” try a product are unlikely to do so if they can’t find it easily.
In making this adjustment, companies typically use an estimated distribution, a per-
centage of total stores that will stock the new product, such as ACV % distribution. (See
Section 6.6 for further detail.)
Adjusted Trial Rate (%) Trial Rate (%) * Awareness (%) * ACV (%)


After making these modifications, marketers can calculate the number of customers
who are expected to try the product, simply by applying the adjusted trial rate to the
target population.
Trial Population (#) Target Population (#) * Adjusted Trial Rate (%)
Estimated in this way, trial population (#) is identical to penetration (#) in the trial period.
To forecast trial volume, multiply trial population by the projected average number of
units of a product that will be bought in each trial purchase. This is often assumed to be
one unit because most people will experiment with a single unit of a new product before
buying larger quantities.
Trial Volume (#) Trial Population (#) * Units per Purchase (#)
Combining all these calculations, the entire formula for trial volume is
Trial Volume (#) Target Population (#) * [(80% * Definitely Buy (#))
(30% * Probably Buy (#)) * Awareness (%) * ACV (%)]
* Units per Purchase (#)

EXAMPLE: The marketing team of an office supply manufacturer has a great idea for
a new product—a safety stapler. To sell the idea internally, they want to project the vol-
ume of sales they can expect over the stapler’s first year. Their customer survey yields the
following results (see Table 4.1).

Table 4.1 Customer Survey Responses

% of Customers Responding

Definitely Will Buy 20%

Probably Will Buy 50%
May/May Not Buy 15%
Probably Won’t Buy 10%
Definitely Won’t Buy 5%
Total 100%

On this basis, the company estimates a trial rate for the new stapler by applying the
industry-standard expectation that 80% of “definites” and 30% of “probables” will in fact
buy the product if given the opportunity.


Trial Rate 80% of “Definites” 30% of “Probables”
(80% * 20%) (30% * 50%)
31%
Thus, 31% of the population is expected to try the product if they are aware of it and if
it is available in stores. The company has a strong advertising presence and a solid
distribution network. Consequently, its marketers believe they can obtain an ACV of
approximately 60% for the stapler and that they can generate awareness at a similar level.
On this basis, they project an adjusted trial rate of 11.16% of the population:
Adjusted Trial Rate Trial Rate * Awareness * ACV
31% * 60% * 60% 11.16%

The target population comprises 20 million people. The trial population can be calcu-
lated by multiplying this figure by the adjusted trial rate.
Trial Population Target Population * Adjusted Trial Rate
20 million * 11.16% 2.232 million
Assuming that each person buys one unit when trying the product, the trial volume will
total 2.232 million units.
We can also calculate the trial volume by using the full formula:
Trial Volume Target Population
* [((80% * Definites) (30% * Probables)) * Awareness * ACV]
* Units per purchase
20m * [((80% * 20%) (30% * 50%)) * 60% * 60%)] * 1
2.232 million

REPEAT VOLUME
The second part of projected volume concerns the fraction of people who try a product
and then repeat their purchase decision. The model for this dynamic uses a single esti-
mated repeat rate to yield the number of customers who are expected to purchase again
after their initial trial. In reality, initial repeat rates are often lower than subsequent repeat
rates. For example, it is not uncommon for 50% of trial purchasers to make a first repeat
purchase, but for 80% of those who purchase a second time to go on to purchase a
third time.

Repeat Buyers (#) Trial Population (#) * Repeat Rate (%)

To calculate the repeat volume, the repeat buyers figure can then be multiplied by
an expected volume per purchase among repeat customers and by the number of


times these customers are expected to repeat their purchases within the period under
consideration.

Repeat Volume (#) Repeat Buyers (#) * Repeat Unit Volume per Customer (#)
* Repeat Occasions (#)

This calculation yields the total volume that a new product is expected to generate
among repeat customers over a specified introductory period. The full formula can be
written as

Repeat Volume (#) [Trial Population (#) * Repeat Rate (%)]
* Repeat Unit Volume per Customer (#) * Repeat Occasions (#)

EXAMPLE: Continuing the previous office supplies example, the safety stapler has a
trial population of 2.232 million. Marketers expect the product to be of sufficient quality
to generate a 10% repeat rate in its first year. This will yield 223,200 repeat buyers:
Repeat Buyers Trial Population * Repeat Rate
2.232 million * 10%
223,200
On average, the company expects each repeat buyer to purchase on four occasions during
the first year. On average, each purchase is expected to comprise two units.
Repeat Volume Repeat Buyers * Repeat Unit Volume per Customer * Repeat Occasions
223,200 * 2 * 4
1,785,600 units
This can be represented in the full formula:
Repeat Volume (#) [Repeat Rate (%) * Trial Population (#)]
* Repeat Volume per Customer (#)
* Repeat Occasions (#)
(10% * 2,232,000) * 2 * 4
1,785,600 units

TOTAL VOLUME
Total volume is the sum of trial volume and repeat volume, as all volume must be sold
to either new customers or returning customers.
Total Volume (#) Trial Volume (#) Repeat Volume (#)


To capture total volume in its fully detailed form, we need only combine the previous
formulas.
Total Volume (#) [Target Population * ((0.8 * Definitely Buy 0.3 * Probably Buy)
* Awareness * ACV) * Units per Trial Purchase]
[(Trial Population * Repeat Rate)
* Repeat Volume per Customer * Repeat Occasions]

Example: Total volume in year one for the stapler is the sum of trial volume and
repeat volume.

Total Volume Trial Volume Repeat Volume
2,232,000 1,785,600
4,017,600 Units
A full calculation of this figure and a template for a spreadsheet calculation are presented
in Table 4.2.

Table 4.2 Volume Projection Spreadsheet

Preliminary Data Source

Definitely Will Buy Customer Survey 20%

Probably Will Buy Customer Survey 50%

Likely Buyers
Likely Buyers from Definites Definitely Buy * 80% 16%
Likely Buyers from Probables Probably Buy * 30% 15%
Trial Rate (%) Total of Likely Buyers 31%

Marketing Adjustments
Awareness Estimated from Marketing Plan 60%
ACV Estimated from Marketing Plan 60%
Adjusted Trial Rate (%) Trial Rate * Awareness * ACV 11.2%
Target Population (#) (thousands) Marketing Plan Data 20,000
Trial Population (#) (thousands) Target Population * Adjusted 2,232
Trial Rate
Continues


Table 4.2 Continued

Preliminary Data Source
Unit Volume Purchased per Trial (#) Estimated from Marketing Plan 1
Trial Volume (#) (Thousands) Trial Population * Volume per 2,232
Trier

Repeat Rate (%) Estimated from Marketing Plan 10%
Repeat Buyers (#) Repeat Rate * Trial Population 223,200
Avg. Volume per Repeat Estimated from Marketing Plan 2
Purchase (#)
Repeat Purchase Frequency ** (#) Estimated from Marketing Plan 4
Repeat Volume (Thousands) Frequency Repeat Buyers * Repeat Volume 1,786
per Purchase * Repeat Purchase
Total Volume (Thousands) 4,018
**Note: The average frequency of repeat purchases per repeat purchaser should be adjusted to reflect the
time available for first-time triers to repeat, the purchase cycle (frequency) for the category, and availability.
For example, if trial rates are constant over the year, the number of repeat purchases would be about 50%
of what it would have been if all had tried on day 1 of the period.

Sales projections based on test markets will always require the inclusion of key assump-
tions. In setting these assumptions, marketers face tempting opportunities to make the
assumptions fit the desired outcome. Marketers must guard against that temptation and
perform sensitivity analysis to establish a range of predictions.
Relatively simple metrics such as trial and repeat rates can be difficult to capture
in practice. Although strides have been made in gaining customer data—through cus-
tomer loyalty cards, for example—it will often be difficult to determine whether
customers are new or repeat buyers.
Regarding awareness and distribution: Assumptions concerning the level of public
awareness to be generated by launch advertising are fraught with uncertainty. Marketers
are advised to ask: What sort of awareness does the product need? What complementary
promotions can aid the launch?


Trial and repeat rates are both important. Some products generate strong results in the
trial stage but fail to maintain ongoing sales. Consider the following example.

EXAMPLE: Let’s compare the safety stapler with a new product, such as an enhanced
envelope sealer. The envelope sealer generates less marketing buzz than the stapler but
enjoys a greater repeat rate. To predict results for the envelope sealer, we have adapted the
data from the safety stapler by reducing the top two box responses by half (reflecting its
lower initial enthusiasm) and raising the repeat rate from 10% to 33% (showing stronger
product response after use).
At the six-month mark, sales results for the safety stapler (Product A) are superior to
those for the envelope sealer (Product B). After one year, sales results for the two prod-
ucts are equal. On a three-year time scale, however, the envelope sealer—with its loyal
base of customers—emerges as the clear winner in sales volume (see Figure 4.2).
The data for the graph is derived as shown in Table 4.3.

Volume Over Time: High Interest Versus Loyalty Generating Products
Volume of Sales Generated (Thousands)

7,000

6,000

5,000

4,000

3,000

2,000
6 Months 12 Months 18 Months 2 Years

Product A 3,125 4,018 4,910 5,803

Product B 2,589 4,062 5,535 7,008

Time from Product Launch

Figure 4.2 Time Horizon Influences Perceived Results


Table 4.3 High Initial Interest or Long-Term Loyalty—Results over Time
122

6 Months 12 Months 18 Months 2 Years
MARKETING METRICS

Preliminary Data Source Prod A Prod B Prod A Prod B Prod A Prod B Prod A Prod B

Definitely Will Buy Customer Survey 20% 10% 20% 10% 20% 10% 20% 10%

Probably Will Buy Customer Survey 50% 25% 50% 25% 50% 25% 50% 25%

Differences Highlighted in
Yellow
Likely Buyers

Likely Buyers from Definites Definitely Buy * 16% 8% 16% 8% 16% 8% 16% 8%
80%

Likely Buyers from Probably Buy * 15% 8% 15% 8% 15% 8% 15% 8%
Probables 30%

Trial Rate Total of Likely 31% 16% 31% 16% 31% 16% 31% 16%
Buyers
Marketing Adjustments

Awareness Estimated from 60% 60% 60% 60% 60% 60% 60% 60%
Marketing Plan

ACV Estimated from 60% 60% 60% 60% 60% 60% 60% 60%
Marketing Plan
Adjusted Trial Rate Trial Rate * 11.2% 5.6% 11.2% 5.6% 11.2% 5.6% 11.2% 5.6%
Awareness * ACV
Target Population Marketing Plan 20,000 20,000 20,000 20,000 20,000 20,000 20,000 20,000
(Thousands) Data


Trial Population Target 2,232 1,116 2,232 1,116 2,232 1,116 2,232 1,116
(Thousands) Population *
Adjusted Trial Rate
Unit Volume Purchased Estimated from 1 1 1 1 1 1 1 1
at Trial Marketing Plan
Trial Volume (Thousands) Trial Population 2,232 1,116 2,232 1,116 2,232 1,116 2,232 1,116
* Volume bought

Repeat Rate Estimated from 10% 33% 10% 33% 10% 33% 10% 33%
Marketing Plan
Repeat Buyers Repeat Rate * 223.20 368.28 223.20 368.28 223.20 368.28 223.20 368.28
Trial Population
Chapter 4 Product and Portfolio Management

Repeat Purchase Unit Estimated from 2 2 2 2 2 2 2 2
Volume Marketing Plan
Number of Repeat Estimated from 2 2 4 4 6 6 8 8
Purchases Marketing Plan
Repeat Volume (Thousands) Repeat Buyers * 893 1,473 1,786 2,946 2,678 4,419 3,571 5,892
Repeat Volume *
Number of Repeat
Purchases
Total Volume 3,125 2,589 4,018 4,062 4,910 5,535 5,803 7,008
123


Repeating and Trying: Some models assume that customers, after they stop repeating
purchases, are lost and do not return. However, customers may be acquired, lost, reac-
quired, and lost again. In general, the trial-repeat model is best suited to projecting sales
over the first few periods. Other means of predicting volume include share of require-
ments and penetration metrics (refer to Sections 2.4 and 2.5). Those approaches may be
preferable for products that lack reliable repeat rates.

Heavy
Penetration Share of Usage Market
Market Size Share Requirements Index Share Units Sold

New
Product 1,000,000 5% 80% 1.2 4.8% 48,000

Source Estimated Estimated Estimated Estimated Penetration Share *
Share * Market Size
Share of
Requirements
* Heavy
Usage Index

Ever-Tried: This is slightly different from trial in that it measures the percentage of the
target population that has “ever” (in any previous period) purchased or consumed the
product under study. Ever-tried is a cumulative measure and can never add up to more
than 100%. Trial, by contrast, is an incremental measure. It indicates the percentage of
the population that tries the product for the first time in a given period. Even here, how-
ever, there is potential for confusion. If a customer stops buying a product but tries it
again six months later, some marketers will categorize that individual as a returning
purchaser, others as a new customer. By the latter definition, if individuals can “try” a
product more than once, then the sum of all “triers” could equal more than the total
population. To avoid confusion, when reviewing a set of data, it’s best to clarify the
definitions behind it.
Variations on Trial: Certain scenarios reduce the barriers to trial but entail a lower
commitment by the customer than a standard purchase.
■ Forced Trial: No other similar product is available. For example, many people
who prefer Pepsi-Cola have “tried” Coca-Cola in restaurants that only serve the
latter, and vice versa.
■ Discounted Trial: Consumers buy a new product but at a substantially reduced
price.


Forced and discounted trials are usually associated with lower repeat rates than trials
made through volitional purchase.
Evoked Set: The set of brands that consumers name in response to questions about
which brands they consider (or might consider) when making a purchase in a specific
category. Evoked Sets for breakfast cereals, for example, are often quite large, while those
for coffee may be smaller.
Number of New Products: The number of products introduced for the first time in a
specific time period.
Revenue from New Products: Usually expressed as the percentage of sales generated by
products introduced in the current period or, at times, in the most recent three to five
periods.
Margin on New Products: The dollar or percentage profit margin on new products.
This can be measured separately but does not differ mathematically from margin
calculations.
Company Profit from New Products: The percentage of company profits that is derived
from new products. In working with this figure, it is important to understand how “new
product” is defined.
Target Market Fit: Of customers purchasing a product, target market fit represents the
percentage who belong in the demographic, psychographic, or other descriptor set for that
item. Target market fit is useful in evaluating marketing strategies. If a large percentage of
customers for a product belongs to groups that have not previously been targeted, mar-
keters may reconsider their targets—and their allocation of marketing spending.

4.2 Growth: Percentage and CAGR
There are two common measures of growth. Year-on-year percentage growth uses
the prior year as a base for expressing percentage change from one year to the next.
Over longer periods of time, compound annual growth rate (CAGR) is a generally
accepted metric for average growth rates.
Value ($,#,%) t Value ($,#,%) t 1
Year-on-Year Growth (%)
Value ($,#,%) t 1
Compound Annual Growth Rate, or = {[Ending Value ($,#,%)/Starting Value ($,#,%)]
CAGR (%) ^ [1/Number of Years (#)]} 1

Same stores growth Growth calculated only on the basis of stores that were fully
established in both the prior and current periods.


Purpose: To measure growth.
Growth is the aim of virtually all businesses. Indeed, perceptions of the success or fail-
ure of many enterprises are based on assessments of their growth. Measures of year-on-
year growth, however, are complicated by two factors:

1. Changes over time in the base from which growth is measured. Such changes
might include increases in the number of stores, markets, or salespeople gener-
ating sales. This issue is addressed by using “same store” measures (or corollary
measures for markets, sales personnel, and so on).
2. Compounding of growth over multiple periods. For example, if a company
achieves 30% growth in one year, but its results remain unchanged over the
two subsequent years, this would not be the same as 10% growth in each of
three years. CAGR, the compound annual growth rate, is a metric that
addresses this issue.

Construction
Percentage growth is the central plank of year-on-year analysis. It addresses the ques-
tion: What has the company achieved this year, compared to last year? Dividing the
results for the current period by the results for the prior period will yield a comparative
figure. Subtracting one from the other will highlight the increase or decrease between
periods. When evaluating comparatives, one might say that results in Year 2 were, for
example, 110% of those in Year 1. To convert this figure to a growth rate, one need only
subtract 100%.
The periods considered are often years, but any time frame can be chosen.
Value ($,#,%) t Value ($,#,%) t 1
Year-on-Year Growth (%)
Value ($,#,%) t 1

EXAMPLE: Ed’s is a small deli, which has had great success in its second year of oper-
ation. Revenues in Year 2 are $570,000, compared with $380,000 in Year 1. Ed calculates
his second-year sales results to be 150% of first-year revenues, indicating a growth rate
of 50%.
$570,000 $380,000
Year-on-Year Sales Growth 50%
$380,000

Same Stores Growth: This metric is at the heart of retail analysis. It enables mar-
keters to analyze results from stores that have been in operation for the entire period


under consideration. The logic is to eliminate the stores that have not been open for the
full period to ensure comparability. Thus, same stores growth sheds light on the effec-
tiveness with which equivalent resources were used in the period under study versus the
prior period. In retail, modest same stores growth and high general growth rates would
indicate a rapidly expanding organization, in which growth is driven by investment.
When both same stores growth and general growth are strong, a company can be viewed
as effectively using its existing base of stores.

EXAMPLE: A small retail chain in Bavaria posts impressive percentage growth figures,
moving from €58 million to €107 million in sales (84% growth) from one year to the
next. Despite this dynamic growth, however, analysts cast doubt on the firm’s business
model, warning that its same stores growth measure suggests that its concept is failing
(see Table 4.4).

Table 4.4 Revenue of a Bavarian Chain Store

Store Opened Revenue First Year (m) Revenue Second Year (m)
A Year 1 €10 €9
B Year 1 €19 €20
C Year 1 €20 €15
D Year 1 €9 €11
E Year 2 n/a €52

€58 €107

Same stores growth excludes stores that were not open at the beginning of the first year
under consideration. For simplicity, we assume that stores in this example were opened
on the first day of Years 1 and 2, as appropriate. On this basis, same stores revenue in
Year 2 would be €55 million—that is, the €107 million total for the year, less the €52 mil-
lion generated by the newly opened Store E. This adjusted figure can be entered into the
same stores growth formula:
(Stores A-D Sales Year 2) (Stores A-D Sales Year 1)
Same Stores Growth
$Stores A-D Sales Year 1

€55m €58m
5%
€58


As demonstrated by its negative same stores growth figure, sales growth at this firm has
been fueled entirely by a major investment in a new store. This suggests serious doubts
about its existing store concept. It also raises a question: Did the new store “cannibalize”
existing store sales? (See the next section for cannibalization metrics.)

Compounding Growth, Value at Future Period: By compounding, managers adjust
growth figures to account for the iterative effect of improvement. For example, 10%
growth in each of two successive years would not be the same as a total of 20% growth
over the two-year period. The reason: Growth in the second year is built upon the ele-
vated base achieved in the first. Thus, if sales run $100,000 in Year 0 and rise by 10% in
Year 1, then Year 1 sales come to $110,000. If sales rise by a further 10% in Year 2,
however, then Year 2 sales do not total $120,000. Rather, they total $110,000 (10% *
$110,000) = $121,000.
The compounding effect can be easily modeled in spreadsheet packages, which enable
you to work through the compounding calculations one year at a time. To calculate a
value in Year 1, multiply the corresponding Year 0 value by one plus the growth rate.
Then use the value in Year 1 as a new base and multiply it by one plus the growth rate
to determine the corresponding value for Year 2. Repeat this process through the
required number of years.

EXAMPLE: Over a three-year period, $100 dollars, compounded at a 10% growth
rate, yields $133.10.

Year 0 to Year 1 $100 10% Growth (that is, $10) $110
Year 1 to Year 2 $110 10% Growth ($11) $121
Year 2 to Year 3 $121 10% Growth ($12.10) $133.10

There is a mathematical formula that generates this effect. It multiplies the value at the
beginning—that is, in Year 0—by one plus the growth rate to the power of the number
of years over which that growth rate applies.
Value in Future Period ($,#,%) Current Value ($,#,%) * [(1 CAGR (%)) ^ Number of
Periods (#)]

EXAMPLE: Using the formula, we can calculate the impact of 10% annual growth
over a period of three years. The value in Year 0 is $100. The number of years is 3. The
growth rate is 10%.


Value in Future Period Value in Year 0 * (1 Growth Rate) ^ Number of Years
$100 * (100% 10%) ^ 3
$100 * 133.1% $133.10

Compound Annual Growth Rate (CAGR): The CAGR is a constant year-on-year
growth rate applied over a period of time. Given starting and ending values, and the
length of the period involved, it can be calculated as follows:
CAGR (%) {[Ending Value ($,#)/Starting Value ($,#)] ^ 1/Number of Periods (#)} 1

EXAMPLE: Let’s assume we have the results of the compounding growth observed in
the previous example, but we don’t know what the growth rate was. We know that the
starting value was $100, the ending value was $133.10, and the number of years was 3. We
can simply enter these numbers into the CAGR formula to derive the CAGR.

CAGR [(Ending Value/Starting Value) ^ (1/Number of Years)] 1
[($133.10/$100) ^ 1/3] 1
[1.331 (The Increase) ^ 1/3 (Cube Root)] 1 = 1.1 1 = 10%
Thus, we determine that the growth rate was 10%.

Percentage growth is a useful measure as part of a package of metrics. It can be deceiv-
ing, however, if not adjusted for the addition of such factors as stores, salespeople, or
products, or for expansion into new markets. “Same store” sales, and similar adjust-
ments for other factors, tell us how effectively a company uses comparable resources.
These very adjustments, however, are limited by their deliberate omission of factors that
weren’t in operation for the full period under study. Adjusted figures must be reviewed
in tandem with measures of total growth.

Life Cycle: Marketers view products as passing through four stages of development:

■ Introductory: Small markets not yet growing fast.
■ Growth: Larger markets with faster growth rates.
■ Mature: Largest markets but little or no growth.
■ Decline: Variable size markets with negative growth rates.
This is a rough classification. No generally accepted rules exist for making these
classifications.


4.3 Cannibalization Rates and Fair Share Draw
Cannibalization is the reduction in sales (units or dollars) of a firm’s existing prod-
ucts due to the introduction of a new product. The cannibalization rate is generally
calculated as the percentage of a new product’s sales that represents a loss of sales
(attributable to the introduction of the new entrant) of a specific existing product
or products.
Sales Lost from Existing Products (#,$)
Cannibalization Rate (%)
Sales of New Product (#,$)

Cannibalization rates represent an important factor in the assessment of new prod-
uct strategies.
Fair share draw constitutes an assumption or expectation that a new product will
capture sales (in unit or dollar terms) from existing products in proportion to the
market shares of those existing products.

Cannibalization is a familiar business dynamic. A company with a successful product
that has strong market share is faced by two conflicting ideas. The first is that it wants to
maximize profits on its existing product line, concentrating on the current strengths
that promise success in the short term. The second idea is that this company—or its
competitors—may identify opportunities for new products that better fit the needs of
certain segments. If the company introduces a new product in this field, however, it may
“cannibalize” the sales of its existing products. That is, it may weaken the sales of
its proven, already successful product line. If the company declines to introduce the new
product, however, it will leave itself vulnerable to competitors who will launch such
a product, and may thereby capture sales and market share from the company. Often,
when new segments are emerging and there are advantages to being early to market, the
key factor becomes timing. If a company launches its new product too early, it may lose
too much income on its existing line; if it launches too late, it may miss the new oppor-
tunity altogether.
Cannibalization: A market phenomenon in which sales of one product are
achieved at the expense of some of a firm’s other products.
The cannibalization rate is the percentage of sales of a new product that come from a
specific set of existing products.
Sales Lost from Existing Products (#,$)
Cannibalization Rate (%)
Sales of New Product (#,$)


EXAMPLE: A company has a single product that sold 10 units in the previous period.
The company plans to introduce a new product that will sell 5 units with a cannibaliza-
tion rate of 40%. Thus 40% of the sales of the new product (40% * 5 units 2 units)
come at the expense of the old product. Therefore, after cannibalization, the company
can expect to sell 8 units of the old product and 5 of the new product, or 13 units in total.

Any company considering the introduction of a new product should confront the
potential for cannibalization. A firm would do well to ensure that the amount of canni-
balization is estimated beforehand to provide an idea of how the product line’s contri-
bution as a whole will change. If performed properly, this analysis will tell a company
whether overall profits can be expected to increase or decrease with the introduction of
the new product line.

EXAMPLE: Lois sells umbrellas on a small beach where she is the only provider. Her
financials for last month were as follows:
Umbrella Sales Price: $20
Variable Cost per Umbrella: $10
Umbrella Contribution per Unit: $10
Total Unit Sales per Month: 100
Total Monthly Contribution: $1,000
Next month, Lois plans to introduce a bigger, lighter-weight umbrella called the “Big
Block.” Projected financials for the Big Block are as follows:
Big Block Sales Price: $30
Variable Cost per Big Block: $15
Big Block Contribution per Unit: $15
Total Unit Sales per Month (Big Block): 50
Total Monthly Contribution (Big Block): $750
If there is no cannibalization, Lois thus expects her total monthly contribution will be
$1,000 $750, or $1,750. Upon reflection, however, Lois thinks that the unit
cannibalization rate for Big Block will be 60%. Her projected financials after accounting
for cannibalization are therefore as follows:
Big Block Unit Sales: 50
Cannibalization Rate: 60%
Regular Umbrella Sales Lost: 50 * 60% 30


New Regular Umbrella Sales: 100 30 70
New Total Contribution (Regular): 70 Units * $10 Contribution per Unit $700
Big Block Total Contribution: 50 Units * $15 Contribution per Unit $750
Lois’ Total Monthly Contribution: $1,450
Under these projections, total umbrella sales will increase from 100 to 120, and total con-
tribution will increase from $1,000 to $1,450. Lois will replace 30 regular sales with 30
Big Block sales and gain an extra $5 unit contribution on each. She will also sell 20 more
umbrellas than she sold last month and gain $15 unit contribution on each.
In this scenario, Lois was in the enviable position of being able to cannibalize a lower-
margin product with a higher-margin one. Sometimes, however, new products carry unit
contributions lower than those of existing products. In these instances, cannibalization
reduces overall profits for the firm.

An alternative way to account for cannibalization is to use a weighted contribution
margin. In the previous example, the weighted contribution margin would be the unit
margin Lois receives for Big Block after accounting for cannibalization. Because each
Big Block contributes $15 directly and cannibalizes the $10 contribution generated
by regular umbrellas at a 60% rate, Big Block’s weighted contribution margin is $15
(0.6 * $10), or $9 per unit. Because Lois expects to sell 50 Big Blocks, her total contribu-
tion is projected to increase by 50 * $9, or $450. This is consistent with our previous
calculations.
If the introduction of Big Block requires some fixed marketing expenditure, then the
$9 weighted margin can be used to find the break-even number of Big Block sales
required to justify that expenditure. For example, if the launch of Big Block requires
$360 in one-time marketing costs, then Lois needs to sell $360/$9, or 40 Big Blocks to
break even on that expenditure.
If a new product has a margin lower than that of the existing product that it cannibal-
izes, and if its cannibalization rate is high enough, then its weighted contribution mar-
gin might be negative. In that case, company earnings will decrease with each unit of the
new product sold.
Cannibalization refers to a dynamic in which one product of a firm takes share from
one or more other products of the same firm. When a product takes sales from a
competitor’s product, that is not cannibalization . . . though managers sometimes incor-
rectly state that their new products are “cannibalizing” sales of a competitor’s goods.
Though it is not cannibalization, the impact of a new product on the sales of competing
goods is an important consideration in a product launch. One simple assumption about
how the introduction of a new product might affect the sales of existing products is
called “fair share draw.”


Fair Share Draw: The assumption that a new product will capture sales (in unit
or dollar terms) from existing products in direct proportion to the market shares
held by those existing products.

EXAMPLE: Three rivals compete in the youth fashion market in a small town. Their
sales and market shares for last year appear in the following table.

Firm Sales Share
Threadbare $500,000 50%
Too Cool for School $300,000 30%
Tommy Hitchhiker $200,000 20%
Total $1,000,000 100%

A new entrant is expected to enter the market in the coming year and to generate
$300,000 in sales. Two-thirds of those sales are expected to come at the expense of the
three established competitors. Under an assumption of fair share draw, how much will
each firm sell next year?
If the new firm takes two-thirds of its sales from existing competitors, then this “capture”
of sales will total (2/3) * $300,000, or $200,000. Under fair share draw, the breakdown of
that $200,000 will be proportional to the shares of the current competitors. Thus 50% of
the $200,000 will come from Threadbare, 30% from Too Cool, and 20% from Tommy.
The following table shows the projected sales and market shares next year of the four
competitors under the fair share draw assumption:

Firm Sales Share
Threadbare $400,000 36.36%
Too Cool for School $240,000 21.82%
Tommy Hitchhiker $160,000 14.55%
New Entrant $300,000 27.27%
Total $1,100,000 100%


Notice that the new entrant expands the market by $100,000, an amount equal to the
sales of the new entrant that do not come at the expense of existing competitors. Notice
also that under fair share draw, the relative shares of the existing competitors remain
unchanged. For example, Threadbare’s share, relative to the total of the original three
competitors, is 36.36/(36.36 21.82 14.55), or 50%—equal to its share before the
entry of the new competitor.

As noted previously, in cannibalization, one of a firm’s products takes sales from one or
more of that firm’s other products. Sales taken from the products of competitors are not
“cannibalized” sales, though some managers label them as such.
Cannibalization rates depend on how the features, pricing, promotion, and distribution
of the new product compare to those of a firm’s existing products. The greater the
similarity of their respective marketing strategies, the higher the cannibalization rate is
likely to be.
Although cannibalization is always an issue when a firm launches a new product that
competes with its established line, this dynamic is particularly damaging to the firm’s
profitability when a low-margin entrant captures sales from the firm’s higher-margin
offerings. In such cases, the new product’s weighted contribution margin can be nega-
tive. Even when cannibalization rates are significant, however, and even if the net effect
on the bottom line is negative, it may be wise for a firm to proceed with a new product
if management believes that the original line is losing its competitive strength. The
following example is illustrative.

EXAMPLE: A producer of powdered-milk formula has an opportunity to introduce a
new, improved formula. The new formula has certain attributes not found in the firm’s
existing products. Due to higher costs, however, it will carry a contribution margin of
only $8, compared with the $10 margin of the established formula. Analysis suggests that
the unit cannibalization rate of the new formula will be 90% in its initial year. If the firm
expects to sell 300 units of the new formula in its first year, should it proceed with the
introduction?
Analysis shows that the new formula will generate $8 * 300, or $2,400 in direct contribu-
tion. Cannibalization, however, will reduce contribution from the established line by
$10 * 0.9 * 300, or $2,700. Thus, the company’s overall contribution will decline by
$300 with the introduction of the new formula. (Note also that the weighted unit margin


for the new product is $1.) This simple analysis suggests that the new formula should
not be introduced.
The following table, however, contains the results of a more detailed four-year analysis.
Reflected in this table are management’s beliefs that without the new formula, sales of the
regular formula will decline to 700 units in Year 4. In addition, unit sales of the new formula
are expected to increase to 600 in Year 4, while cannibalization rates decline to 60%.

Year 1 Year 2 Year 3 Year 4 Total
Unit Sales of Regular Formula 1,000 900 800 700 3,400
Without New Product Launch

— —
Unit Sales of New Formula 300 400 500 600 1,800
Cannibalization Rate 90% 80% 70% 60% —
Unit Sales of Regular Formula 730 580 450 340 2,100
with New Product Launch

Without the new formula, total four-year contribution is projected as $10 * 3,400,
or $34,000. With the new formula, total contribution is projected as ($8 * 1,800)
($10 * 2,100), or $35,400. Although forecast contribution is lower in Year 1 with
the new formula than without it, total four-year contribution is projected to be higher
with the new product due to increases in new-formula sales and decreases in the
cannibalization rate.

4.4 Brand Equity Metrics
Brand equity is strategically crucial, but famously difficult to quantify. Many experts
have developed tools to analyze this asset, but there’s no universally accepted way to
measure it. In this section, we’ll consider the following techniques to gain insight in
this area:
Brand Equity Ten (Aaker)
Brand Asset® Valuator (Young & Rubicam)
Brand Equity Index (Moran)
Brand Valuation Model (Interbrand)


Purpose: To measure the value of a brand.
A brand encompasses the name, logo, image, and perceptions that identify a product,
service, or provider in the minds of customers. It takes shape in advertising, packaging,
and other marketing communications, and becomes a focus of the relationship with
consumers. In time, a brand comes to embody a promise about the goods it identifies—
a promise about quality, performance, or other dimensions of value, which can influ-
ence consumers’ choices among competing products. When consumers trust a brand
and find it relevant, they may select the offerings associated with that brand over those
of competitors, even at a premium price. When a brand’s promise extends beyond a par-
ticular product, its owner may leverage it to enter new markets. For all these reasons, a
brand can hold tremendous value, known as brand equity.
Yet this value can be remarkably difficult to measure. At a corporate level, when one
company buys another, marketers might analyze the goodwill component of the pur-
chase price to shed light on the value of the brands acquired. As goodwill represents the
excess paid for a firm—beyond the value of its tangible, measurable assets, and as a
company’s brands constitute important intangible assets—the goodwill figure may pro-
vide a useful indicator of the value of a portfolio of brands. Of course, a company’s
brands are rarely the only intangible assets acquired in such a transaction. Goodwill
more frequently encompasses intellectual property and other intangibles in addition to
brand. The value of intangibles, as estimated by firm valuations (sales or share prices),
is also subject to economic cycles, investor “exuberance,” and other influences that are
difficult to separate from the intrinsic value of the brand.
From a consumer’s perspective, the value of a brand might be the amount she would be
willing to pay for merchandise that carries the brand’s name, over and above the price
she’d pay for identical unbranded goods.2 Marketers strive to estimate this premium in
order to gain insight into brand equity. Here again, however, they encounter daunting
complexities, as individuals vary not only in their awareness of different brands, but in
the criteria by which they judge them, the evaluations they make, and the degree to
which those opinions guide their purchase behavior.
Theoretically, a marketer might aggregate these preferences across an entire population
to estimate the total premium its members would pay for goods of a certain brand. Even
that, however, wouldn’t fully capture brand equity. What’s more, the value of a brand
encompasses not only the premium a customer will pay for each unit of merchandise
associated with that brand, but also the incremental volume it generates. A successful
brand will shift outward the demand curve for its goods or services; that is, it not only
will enable a provider to charge a higher price (P’ rather than P, as seen in Figure 4.3),
but it will also sell an increased quantity (Q’ rather than Q). Thus, brand equity in this
example can be viewed as the difference between the revenue with the brand (P’ × Q’)
and the revenue without the brand (P × Q)—depicted as the shaded area in Figure 4.3.


(Of course, this example focuses on revenue, when, in fact, it is profit or present value of
profits that matters more.)

Price

P’
P High Brand Equity
Low Brand Equity
Q Q’
Quantity

Figure 4.3 Brand Equity—Outward Shift of Demand Curve

In practice, of course, it’s difficult to measure a demand curve, and few marketers do so.
Because brands are crucial assets, however, both marketers and academic researchers
have devised means to contemplate their value. David Aaker, for example, tracks 10
attributes of a brand to assess its strength. Bill Moran has formulated a brand equity
index that can be calculated as the product of effective market share, relative price, and
customer retention. Kusum Ailawadi and her colleagues have refined this calculation,
suggesting that a truer estimate of a brand’s value might be derived by multiplying the
Moran index by the dollar volume of the market in which it competes. Young &
Rubicam, a marketing communications agency, has developed a tool called the Brand
Asset Valuator©, which measures a brand’s power on the basis of differentiation, rele-
vance, esteem, and knowledge. An even more theoretical conceptualization of brand
equity is the difference of the firm value with and without the brand. If you find it dif-
ficult to imagine the firm without its brand, then you can appreciate how difficult it is
to quantify brand equity. Interbrand, a brand strategy agency, draws upon its own
model to separate tangible product value from intangible brand value and uses the lat-
ter to rank the top 100 global brands each year. Finally, conjoint analysis can shed light
on a brand’s value because it enables marketers to measure the impact of that brand on
customer preference, treating it as one among many attributes that consumers trade off
in making purchase decisions (see section 4.5).

Construction
Brand Equity Ten (Aaker): David Aaker, a marketing professor and brand consultant,
highlights 10 attributes of a brand that can be used to assess its strength. These include
Differentiation, Satisfaction or Loyalty, Perceived Quality, Leadership or Popularity,
Perceived Value, Brand Personality, Organizational Associations, Brand Awareness,


Market Share, and Market Price and Distribution Coverage. Aaker doesn’t weight the
attributes or combine them in an overall score, as he believes any weighting would be
arbitrary and would vary among brands and categories. Rather, he recommends track-
ing each attribute separately.
Brand Equity Index (Moran): Marketing executive Bill Moran has derived an index of
brand equity as the product of three factors: Effective Market Share, Relative Price, and
Durability.
Brand Equity Index (I) = Effective Market Share (%) * Relative Price (I) * Durability (%)
Effective Market Share is a weighted average. It represents the sum of a brand’s market
shares in all segments in which it competes, weighted by each segment’s proportion of
that brand’s total sales. Thus, if a brand made 70% of its sales in Segment A, in which it
had a 50% share of the market, and 30% of its sales in Segment B, in which it had a 20%
share, its Effective Market Share would be (0.7 * 0.5) + (0.3 * 0.2) = 0.35 + 0.06 = 0.41,
or 41%.
Relative Price is a ratio. It represents the price of goods sold under a given brand,
divided by the average price of comparable goods in the market. For example, if goods
associated with the brand under study sold for $2.50 per unit, while competing goods
sold for an average of $2.00, that brand’s Relative Price would be 1.25, and it would be
said to command a price premium. Conversely, if the brand’s goods sold for $1.50, ver-
sus $2.00 for the competition, its Relative Price would be 0.75, placing it at a discount to
the market. Note that this measure of relative price is not the same as dividing the brand
price by the market average price. It does have the advantage that, unlike the latter, the
calculated value is not affected by the market share of the firm or its competitors.
Durability is a measure of customer retention or loyalty. It represents the percentage of
a brand’s customers who will continue to buy goods under that brand in the following
year.

EXAMPLE: ILLI is a tonic drink that focuses on two geographic markets—eastern and
western U.S. metropolitan areas. In the western market, which accounts for 60% of ILLI’s
sales, the drink has a 30% share of the market. In the East, where ILLI makes the remain-
ing 40% of its sales, it has a 50% share of the market.
Effective Market Share is equal to the sum of ILLI’s shares of the segments, weighted by
the percentage of total brand sales represented by each.
West = 30% * 60% = 0.18
East = 50% * 40% = 0.20
Effective Market Share = 0.38


The average price for tonic drinks is $2.00, but ILLI enjoys a premium. It generally sells
for $2.50, yielding a Relative Price of $2.50 / $2.00, or 1.25.
Half of the people who purchase ILLI this year are expected to repeat next year, generat-
ing a Durability figure of 0.5. (See section 4.1 for a definition of repeat rates.)
With this information, ILLI’s Brand Equity Index can be calculated as follows:
Brand Equity = Effective Market Share * Relative Price * Durability = 0.38 * 1.25 * 0.5
= 0.2375

Clearly, marketers can expect to encounter interactions among the three factors behind
a Brand Equity Index. If they raise the price of a brand’s goods, for example, they may
increase its Relative Price but reduce its Effective Market Share and Durability. Would
the overall effect be positive for the brand? By estimating the Brand Equity Index before
and after the price increase under consideration, marketers may gain insight into that
question.
Notice that two of the factors behind this index, Effective Market Share and Relative
Price, draw upon the axes of a demand curve (quantity and price). In constructing his
index, Moran has taken those two factors and combined them, through year-to-year
retention, with the dimension of time.
Ailawadi, et al suggested that the equity index of a brand can be enhanced by multiply-
ing it by the dollar volume of the market in which the brand competes, generating a bet-
ter estimate of its value. Ailawadi also contends that the equity of a brand is better
captured by its overall revenue premium (relative to generic goods) rather than its price
per unit alone, as the revenue figure incorporates both price and quantity and so reflects
a jump from one demand curve to another rather than a movement along a single
curve.
Brand Asset Valuator (Young & Rubicam): Young & Rubicam, a marketing communi-
cations agency, has developed the Brand Asset Valuator, a tool to diagnose the power
and value of a brand. In using it, the agency surveys consumers’ perspectives along four
dimensions:
■ Differentiation: The defining characteristics of the brand and its distinctiveness
relative to competitors.
■ Relevance: The appropriateness and connection of the brand to a given
consumer.
■ Esteem: Consumers’ respect for and attraction to the brand.
■ Knowledge: Consumers’ awareness of the brand and understanding of what it
represents.


Young & Rubicam maintains that these criteria reveal important factors behind brand
strength and market dynamics. For example, although powerful brands score high on all
four dimensions, growing brands may earn higher grades for Differentiation and
Relevance, relative to Knowledge and Esteem. Fading brands often show the reverse pat-
tern, as they’re widely known and respected but may be declining toward commoditiza-
tion or irrelevance (see Figure 4.4).

Growth Brand Strong Brand

Di Re Es Kn Di Re Es Kn
ffe l te o ffe l te o
r en e v a n em w l e r en evan em w l e
tia ce dg tia ce dg
tio e tio e
n n

Declining Brand

Weak Brand

Di Re Es Kn Di Re Es Kn
ffe l te o ffe l te o
r en e v a n em w l e r en e v a n em w l e
tia dg tia dg
tio c e e tio c e e
n n

Figure 4.4 Young & Rubicam Brand Asset Valuator Patterns of Brand Equity

The Brand Asset Valuator is a proprietary tool, but the concepts behind it have broad
appeal. Many marketers apply these concepts by conducting independent research and
exercising judgment about their own brands relative to the competition. Leon
Ramsellar3 of Philips Consumer Electronics, for example, has reported using four key
measures in evaluating brand equity and offered sample questions for assessing them.
■ Uniqueness: Does this product offer something new to me?
■ Relevance: Is this product relevant for me?
■ Attractiveness: Do I want this product?
■ Credibility: Do I believe in the product?


Clearly Ramsellar’s list is not the same as Y&R’s BAV, but the similarity of the first two
factors is hard to miss.
Brand Valuation Model (Interbrand): Interbrand, a brand strategy agency, draws upon
financial results and projections in its own model for brand valuation. It reviews a com-
pany’s financial statements, analyzes its market dynamics and the role of brand in
income generation, and separates those earnings attributable to tangible assets (capital,
product, packaging, and so on) from the residual that can be ascribed to a brand. It then
forecasts future earnings and discounts these on the basis of brand strength and risk.
The agency estimates brand value on this basis and tabulates a yearly list of the 100 most
valuable global brands.
Conjoint Analysis: Marketers use conjoint analysis to measure consumers’ preference
for various attributes of a product, service, or provider, such as features, design, price, or
location (see section 4.5). By including brand and price as two of the attributes under
consideration, they can gain insight into consumers’ valuation of a brand—that is, their
willingness to pay a premium for it.

The methods described previously represent experts’ best attempts to place a value on a
complex and intangible entity. Almost all of the metrics in this book are relevant to
brand equity along one dimension or another.

Brand strategy is a broad field and includes several concepts that at first may appear to
be measurable. Strictly speaking, however, brand strategy is not a metric.
Brand Identity: This is the marketer’s vision of an ideal brand—the company’s goal for
perception of that brand by its target market. All physical, emotional, visual, and verbal
messages should be directed toward realization of that goal, including name, logo, sig-
nature, and other marketing communications. Brand Identity, however, is not stated in
quantifiable terms.
Brand Position and Brand Image: These refer to consumers’ actual perceptions of a
brand, often relative to its competition. Brand Position is frequently measured along
product dimensions that can be mapped in multi-dimensional space. If measured con-
sistently over time, these dimensions may be viewed as metrics—as coordinates on a
perceptual map. (See Section 2.7 for a discussion of attitude, usage measures, and the
hierarchy of effects.)


Product Differentiation: This is one of the most frequently used terms in marketing,
but it has no universally agreed-upon definition. More than mere “difference,” it gener-
ally refers to distinctive attributes of a product that generate increased customer prefer-
ence or demand. These are often difficult to view quantitatively because they may be
actual or perceived, as well as non-monotonic. In other words, although certain attrib-
utes such as price can be quantified and follow a linear preference model (that is, either
more or less is always better), others can’t be analyzed numerically or may fall into a
sweet spot, outside of which neither more nor less would be preferred (the spiciness of
a food, for example). For all these reasons, Product Differentiation is hard to analyze as
a metric and has been criticized as a “meaningless term.”

Additional Citations
Simon, Julian, “Product Differentiation”: A Meaningless Term and an Impossible
Concept, Ethics, Vol. 79, No. 2 (Jan., 1969), pp. 131-138. Published by The University of
Chicago Press.

4.5 Conjoint Utilities and Consumer Preference
Conjoint utilities measure consumer preference for an attribute level and then—
by combining the valuations of multiple attributes—measure preference for an
overall choice. Measures are generally made on an individual basis, although this
analysis can also be performed on a segment level. In the frozen pizza market, for
example, conjoint utilities can be used to determine how much a customer values
superior taste (one attribute) versus paying extra for premium cheese (a second
attribute).
Conjoint utilities can also play a role in analyzing compensatory and non-
compensatory decisions. Weaknesses in compensatory factors can be made up in
other attributes. A weakness in a non-compensatory factor cannot be overcome by
other strengths.
Conjoint analysis can be useful in determining what customers really want and—
when price is included as an attribute—what they’ll pay for it. In launching new
products, marketers find such analyses useful for achieving a deeper understanding
of the values that customers place on various product attributes. Throughout prod-
uct management, conjoint utilities can help marketers focus their efforts on the
attributes of greatest importance to customers.


Purpose: To understand what customers want.
Conjoint analysis is a method used to estimate customers’ preferences, based on how
customers weight the attributes on which a choice is made. The premise of conjoint
analysis is that a customer’s preference between product options can be broken into a set
of attributes that are weighted to form an overall evaluation. Rather than asking people
directly what they want and why, in conjoint analysis, marketers ask people about their
overall preferences for a set of choices described on their attributes and then decompose
those into the component dimensions and weights underlying them. A model can be
developed to compare sets of attributes to determine which represents the most appeal-
ing bundle of attributes for customers.
Conjoint analysis is a technique commonly used to assess the attributes of a product or
service that are important to targeted customers and to assist in the following:
■ Product design
■ Advertising copy
■ Pricing
■ Segmentation
■ Forecasting

Construction
Conjoint Analysis: A method of estimating customers by assessing the overall
preferences customers assign to alternative choices.
An individual’s preference can be expressed as the total of his or her baseline preferences
for any choice, plus the partworths (relative values) for that choice expressed by the
individual.
In linear form, this can be represented by the following formula:
Conjoint Preference Linear Form (I) [Partworth of Attribute1 to Individual (I)
* Attribute Level (1)] [Partworth of
Attribute2 to Individual (I) * Attribute Level
(2)] [Partworth of Attribute3 to Individual
(I) * Attribute Level (3)] etc.

EXAMPLE: Two attributes of a cell phone, its price and its size, are ranked through
conjoint analysis, yielding the results shown in Table 4.5.


This could be read as follows:

Table 4.5 Conjoint Analysis: Price and Size of a Cell Phone

Attribute Rank Partworth
Price $100 0.9
Price $200 0.1
Price $300 1
Size Small 0.7
Size Medium 0.1
Size Large 0.6

A small phone for $100 has a partworth to customers of 1.6 (derived as 0.9 0.7). This
is the highest result observed in this exercise. A small but expensive ($300) phone is rated
as 0.3 (that is, 1 0.7). The desirability of this small phone is offset by its price. A
large, expensive phone is least desirable to customers, generating a partworth of 1.6
(that is, ( 1) ( 0.6)).
On this basis, we determine that the customer whose views are analyzed here would pre-
fer a medium-size phone at $200 (utility 0) to a small phone at $300 (utility 0.3).
Such information would be instrumental to decisions concerning the trade-offs between
product design and price.
This analysis also demonstrates that, within the ranges examined, price is more impor-
tant than size from the perspective of this consumer. Price generates a range of effects
from 0.9 to 1 (that is, a total spread of 1.9), while the effects generated by the most and
least desirable sizes span a range only from 0.7 to 0.6 (total spread 1.3).

COMPENSATORY VERSUS NON-COMPENSATORY CONSUMER DECISIONS
A compensatory decision process is one in which a customer evaluates choices with
the perspective that strengths along one or more dimensions can compensate for weak-
nesses along others.
In a non-compensatory decision process, by contrast, if certain attributes of a product
are weak, no compensation is possible, even if the product possesses strengths along
other dimensions. In the previous cell phone example, for instance, some customers may
feel that if a phone were greater than a certain size, no price would make it attractive.


In another example, most people choose a grocery store on the basis of proximity. Any
store within a certain radius of home or work may be considered. Beyond that distance,
however, all stores will be excluded from consideration, and there is nothing a store can
do to overcome this. Even if it posts extraordinarily low prices, offers a stunningly wide
assortment, creates great displays, and stocks the freshest foods, for example, a store will
not entice consumers to travel 400 miles to buy their groceries.
Although this example is extreme to the point of absurdity, it illustrates an important
point: When consumers make a choice on a non-compensatory basis, marketers need to
define the dimensions along which certain attributes must be delivered, simply to qual-
ify for consideration of their overall offering.
One form of non-compensatory decision-making is elimination-by-aspect. In this
approach, consumers look at an entire set of choices and then eliminate those that do
not meet their expectations in the order of the importance of the attributes. In the selec-
tion of a grocery store, for example, this process might run as follows:
■ Which stores are within 5 miles of my home?
■ Which ones are open after 8 p.m.?
■ Which carry the spicy mustard that I like?
■ Which carry fresh flowers?
The process continues until only one choice is left.
In the ideal situation, in analyzing customers’ decision processes, marketers would have
access to information on an individual level, revealing

■ Whether the decision for each customer is compensatory or not
■ The priority order of the attributes
■ The “cut-off ” levels for each attribute
■ The relative importance weight of each attribute if the decision follows a com-
pensatory process
More frequently, however, marketers have access only to past behavior, helping them
make inferences regarding these items.
In the absence of detailed, individual information for customers throughout a market,
conjoint analysis provides a means to gain insight into the decision-making processes of
a sampling of customers. In conjoint analysis, we generally assume a compensatory
process. That is, we assume utilities are additive. Under this assumption, if a choice is
weak along one dimension (for example, if a store does not carry spicy mustard), it can
compensate for this with strength along another (for example, it does carry fresh-cut


flowers) at least in part. Conjoint analyses can approximate a non-compensatory model
by assigning non-linear weighting to an attribute across certain levels of its value.
For example, the weightings for distance to a grocery store might run as follows:

Within 1 mile: 0.9
1-5 miles away: 0.8
5-10 miles away: 0.8
More than 10 miles away: 0.9
In this example, stores outside a 5-mile radius cannot practically make up the loss of
utility they incur as a result of distance. Distance becomes, in effect, a non-compensatory
dimension.
By studying customers’ decision-making processes, marketers gain insight into the
attributes needed to meet consumer expectations. They learn, for example, whether
certain attributes are compensatory or non-compensatory. A strong understanding of
customers’ valuation of different attributes also enables marketers to tailor products and
allocate resources effectively.
Several potential complications arise in considering compensatory versus non-
compensatory decisions. Customers often don’t know whether an attribute is compen-
satory or not, and they may not be readily able to explain their decisions. Therefore, it is
often necessary either to infer a customer’s decision-making process or to determine
that process through an evaluation of choices, rather than a description of the process.
It is possible, however, to uncover non-compensatory elements through conjoint analy-
sis. Any attribute for which the valuation spread is so high that it cannot practically be
made up by other features is, in effect, a non-compensatory attribute.

EXAMPLE: Among grocery stores, Juan prefers the Acme market because it’s close to his
home, despite the fact that Acme’s prices are generally higher than those at the local Shoprite
store. A third store, Vernon’s, is located in Juan’s apartment complex. But Juan avoids it
because Vernon’s doesn’t carry his favorite soda.
From this information, we know that Juan’s shopping choice is influenced by at least three fac-
tors: price, distance from his home, and whether a store carries his favorite soda. In Juan’s deci-
sion process, price and distance seem to be compensating factors. He trades price for distance.
Whether the soda is stocked seems to be a non-compensatory factor. If a store doesn’t carry
Juan’s favorite soda, it will not win his business, regardless of how well it scores on price and
location.


Prior to conducting a conjoint study, it is necessary to identify the attributes of impor-
tance to a customer. Focus groups are commonly used for this purpose. After attributes
and levels are determined, a typical approach to Conjoint Analysis is to use a fractional
factorial orthogonal design, which is a partial sample of all possible combinations of
attributes. This is to reduce the total number of choice evaluations required by the
respondent. With an orthogonal design, the attributes remain independent of one
another, and the test doesn’t weigh one attribute disproportionately to another.
There are multiple ways to gather data, but a straightforward approach would be to
present respondents with choices and to ask them to rate those choices according to
their preferences. These preferences then become the dependent variable in a regression,
in which attribute levels serve as the independent variables, as in the previous equation.
Conjoint utilities constitute the weights determined to best capture the preference rat-
ings provided by the respondent.
Often, certain attributes work in tandem to influence customer choice. For example, a
fast and sleek sports car may provide greater value to a customer than would be sug-
gested by the sum of the fast and sleek attributes. Such relationships between attributes
are not captured by a simple conjoint model, unless one accounts for interactions.
Ideally, conjoint analysis is performed on an individual level because attributes can be
weighted differently across individuals. Marketers can also create a more balanced view
by performing the analysis across a sample of individuals. It is appropriate to perform
the analysis within consumer segments that have similar weights. Conjoint analysis can
be viewed as a snapshot in time of a customer’s desires. It will not necessarily translate
indefinitely into the future.
It is vital to use the correct attributes in any conjoint study. People can only tell you their
preferences within the parameters you set. If the correct attributes are not included in a
study, while it may be possible to determine the relative importance of those attributes
that are included, and it may technically be possible to form segments on the basis of the
resulting data, the analytic results may not be valid for forming useful segments. For exam-
ple, in a conjoint analysis of consumer preferences regarding colors and styles of cars, one
may correctly group customers as to their feelings about these attributes. But if con-
sumers really care most about engine size, then those segmentations will be of little value.

4.6 Segmentation Using Conjoint Utilities
Understanding customers’ desires is a vital goal of marketing. Segmenting, or cluster-
ing similar customers into groups, can help managers recognize useful patterns and
identify attractive subsets within a larger market. With that understanding, managers


can select target markets, develop appropriate offerings for each, determine the most
effective ways to reach the targeted segments, and allocate resources accordingly.
Conjoint analysis can be highly useful in this exercise.

Purpose: To identify segments based on conjoint utilities.
As described in the previous section, conjoint analysis is used to determine customers’
preferences on the basis of the attribute weightings that they reveal in their decision-
making processes. These weights, or utilities, are generally evaluated on an individual
level.
Segmentation entails the grouping of customers who demonstrate similar patterns of
preference and weighting with regard to certain product attributes, distinct from the
patterns exhibited by other groups. Using segmentation, a company can decide which
group(s) to target and can determine an approach to appeal to the segment’s members.
After segments have been formed, a company can set strategy based on their attractive-
ness (size, growth, purchase rate, diversity) and on its own capability to serve these seg-
ments, relative to competitors.

Construction
To complete a segmentation based on conjoint utilities, one must first determine utility
scores at an individual customer level. Next, one must cluster these customers into seg-
ments of like-minded individuals. This is generally done through a methodology known
as cluster analysis.
Cluster Analysis: A technique that calculates the distances between customer and
forms groups by minimizing the differences within each group and maximizing the
differences between groups.
Cluster analysis operates by calculating a “distance” (a sum of squares) between individ-
uals and, in a hierarchical fashion, starts pairing those individuals together. The process
of pairing minimizes the “distance” within a group and creates a manageable number of
segments within a larger population.

EXAMPLE: The Samson-Finn Company has three customers. In order to help manage
its marketing efforts, Samson-Finn wants to organize like-minded customers into seg-
ments. Toward that end, it performs a conjoint analysis in which it measures its cus-
tomers’ preferences among products that are either reliable or very reliable, either fast or
very fast (see Table 4.6). It then considers the conjoint utilities of each of its customers to
see which of them demonstrate similar wants. When clustering on conjoint data, the dis-
tances would be calculated on the partworths.


Table 4.6 Customer Conjoint Utilities

Very Reliable Reliable Very Fast Fast
Bob 0.4 0.3 0.6 0.2
Erin 0.9 0.1 0.2 0.7
Yogesh 0.3 0.3 0.5 0.2

The analysis looks at the difference between Bob’s view and Erin’s view on the impor-
tance of reliability on their choice. Bob’s score is 0.4 and Erin’s is 0.9. We can square the
difference between these to derive the “distance” between Bob and Erin.
Using this methodology, the distance between each pair of Samson-Finn’s customers can
be calculated as follows:
Distances Very Reliable Reliable Very Fast Fast
Bob and Erin: (0.4 0.9)2 (0.3 0.1)2 (0.6 0.2)2 (0.2 0.7)2
0.25 0.04 0.16 0.25
0.7
Bob and Yogesh: (0.4 0.3)2 (0.3 0.3)2 (0.6 0.5)2 (0.2 0.2)2
0.01 0.0 0.01 0.0
0.02
Erin and Yogesh: (0.9 0.3)2 (0.1 0.3)2 (0.2 0.5)2 (0.7 0.2)2
= 0.36 0.04 0.09 0.25
0.74
On this basis, Bob and Yogesh appear to be very close to each other because their sum of
squares is 0.02. As a result, they should be considered part of the same segment.
Conversely, in light of the high sum-of-squares distance established by her preferences,
Erin should not be considered a part of the same segment with either Bob or Yogesh.
Of course, most segmentation analyses are performed on large customer bases. This
example merely illustrates the process involved in the cluster analysis calculations.

As noted previously, a customer’s utilities may not be stable, and the segment to which a
customer belongs can shift over time or across occasions. An individual might belong to
one segment for personal air travel, in which price might be a major factor, and another
for business travel, in which convenience might become more important. Such a cus-
tomer’s conjoint weights (utilities) would differ depending on the purchase occasion.


Determining the appropriate number of segments for an analysis can be somewhat arbi-
trary. There is no generally accepted statistical means for determining the “correct”
number of segments. Ideally, marketers look for a segment structure that fulfills the fol-
lowing qualifications:
■ Each segment constitutes a homogeneous group, within which there is relatively
little variance between attribute utilities of different individuals.
■ Groupings are heterogeneous across segments; that is, there is a wide variance
of attribute utilities between segments.

4.7 Conjoint Utilities and Volume Projection
The conjoint utilities of products and services can be used to forecast the market
share that each will achieve and the volume that each will sell. Marketers can project
market share for a given product or service on the basis of the proportion of individ-
uals who select it from a relevant choice set, as well as its overall utility.

Purpose: To use conjoint analysis to project the market share
and the sales volume that will be achieved by a product or service.
Conjoint analysis is used to measure the utilities for a product. The combination
of these utilities, generally additive, represents a scoring of sorts for the expected
popularity of that product. These scores can be used to rank products. However, further
information is needed to estimate market share. One can anticipate that the top-ranked
product in a selection set will have a greater probability of being chosen by an individ-
ual than products ranked lower for that individual. Adding the number of customers
who rank the brand first should allow the calculation of customer share.

To complete a sales volume projection, it is necessary to have a full conjoint analysis.
This analysis must include all the important features according to which consumers
make their choice. Defining the “market” is clearly crucial to a meaningful result.
To define a market, it is important to identify all the choices in that market. Calculating
the percentage of “first choice” selections for each alternative merely provides a “share of
preferences.” To extend this to market share, one must estimate (1) the volume of sales
per customer, (2) the level of distribution or availability for each choice, and (3) the per-
centage of customers who will defer their purchase until they can find their first choice.


The greatest potential error in this process would be to exclude meaningful attributes
from the conjoint analysis.
Network effects can also distort a conjoint analysis. In some instances, customers do
not make purchase decisions purely on the basis of a product’s attributes but are also
affected by its level of acceptance in the marketplace. Such network effects, and the
importance of harnessing or overcoming them, are especially evident during shifts in
technology industries.

Aaker, D.A. (1991). Managing Brand Equity: Capitalizing on the Value of a Brand Name, New
York: Free Press; Toronto; New York: Maxwell Macmillan; Canada: Maxwell Macmillan
International.
Aaker, D.A. (1996). Building Strong Brands, New York: Free Press.
Aaker, D.A., and J.M. Carman. (1982). “Are You Overadvertising?” Journal of Advertising Research,
22(4), 57–70.
Aaker, D.A., and K.L. Keller. (1990). “Consumer Evaluations of Brand Extensions,” Journal of
Marketing, 54(1), 27–41.
Ailawadi, Kusum, and Kevin Keller. (2004). “Understanding Retail Branding: Conceptual Insights
and Research Priorities,” Journal of Retailing, Vol. 80, Issue 4, Winter, 331–342.
Ailawadi, Kusum, Donald Lehman, and Scott Neslin. (2003). “Revenue Premium As an Outcome
Measure of Brand Equity,” Journal of Marketing, Vol. 67, No. 4, 1–17.
Burno, Hernan A., Unmish Parthasarathi, and Nisha Singh, eds. (2005). “The Changing Face of
Measurement Tools Across the Product Lifecycle,” Does Marketing Measure Up? Performance
Metrics: Practices and Impact, Marketing Science Institute, No. 05-301.
Harvard Business School Case: Nestlé Refrigerated Foods Contadina Pasta & Pizza (A)
9-595-035. Rev Jan 30 1997.
Moran, Bill. Personal communication with Paul Farris.


5
CUSTOMER PROFITABILITY

Introduction

Customers, Recency, and Retention Prospect Value Versus Customer Value
Customer Profit Acquisition Versus Retention Spending
Customer Lifetime Value

Chapter 2, “Share of Hearts, Minds, and Markets,” presented metrics designed to measure
how well the firm is doing with its customers as a whole. Previously discussed metrics
were summaries of firm performance with respect to customers for entire markets or
market segments. In this chapter, we cover metrics that measure the performance of indi-
vidual customer relationships. We start with metrics designed to simply count how many
customers the firm serves. As this chapter will illustrate, it is far easier to count the num-
ber of units sold than to count the number of people or businesses buying those units.
Section 5.2 introduces the concept of customer profit. Just as some brands are more prof-
itable than others, so too are some customer relationships. Whereas customer profit is
a metric that summarizes the past financial performance of a customer relationship,
customer lifetime value looks forward in an attempt to value existing customer relation-
ships. Section 5.3 discusses how to calculate and interpret customer lifetime value. One
of the more important uses of customer lifetime value is to inform prospecting decisions.
Section 5.4 explains how this can be accomplished and draws the careful distinction
between prospect and customer value. Section 5.5 discusses acquisition and retention
spending—two metrics firms track in order to monitor the performance of these two
important kinds of marketing spending—spending designed to acquire new customers
and spending designed to retain and profit from existing customers.

153

5.1 Customers The number of Avoid double Measure how well
people (busi- counting people the firm is attract-
nesses) who who bought more ing and retaining
bought from the than one product. customers.
firm during a Carefully define
specified time customer as
period. individual/
household/
screen-name/
division who
bought/ordered/
registered.

5.1 Recency The length of time In non- Track changes in
since a customer’s contractual situa- number of active
last purchase. tions, the firm will customers.
want to track the
recency of its
customers.
5.1 Retention Rate The ratio of cus- Not to be Track changes in
tomers retained to confused with the ability of the
the number at growth (decline) firm to retain
risk. in customer customers.
counts. Retention
refers only to
existing customers
in contractual
situations.
5.2 Customer Profit The difference Requires assigning Allows the firm to
between the revenues and costs identify which
revenues earned to individual customers are
from and the costs customers. profitable and
associated with which are not . . .
the customer rela- as a precursor to
tionship during a differential treat-
specified period. ment designed to
improve firm
profitability.



5.3 Customer The present value Requires a Customer rela-
Lifetime Value of the future cash projection of tionship manage-
flows attributed to future cash flows ment decisions
the customer from a customer should be made
relationship. relationship. with the objective
This will be easier of improving CLV.
to do in a contrac- Acquisition budg-
tual situation. eting should be
Formulations of based on CLV.
CLV differ with
respect to the
treatment of the
initial margin and
acquisition
spending.
5.4 Prospect Lifetime The response rate There are a variety To guide the firm’s
Value times the sum of of equivalent ways prospecting deci-
the initial margin to do the calcula- sions. Prospecting
and the CLV of tions necessary to is beneficial only
the acquired see whether a if the expected
customer minus prospecting effort prospect lifetime
the cost of the is worthwhile. value is positive.
prospecting effort.
5.5 Average The ratio of It is often difficult To track the cost
Acquisition Cost acquisition spend- to isolate acquisi- of acquiring new
ing to the number tion spending customers and to
of new customers from total mar- compare that cost
acquired. keting spending. to the value of the
newly acquired
customers.
5.5 Average Retention The ratio of reten- It is often difficult To monitor reten-
Cost tion spending to to isolate retention spending on
the number of tion spending a per-customer
customers from total mar- basis.
retained. keting spending.
The average reten-
tion cost number
is not very useful
to help make
retention budget-
ing decisions.

Chapter 5 Customer Profitability 155

5.1 Customers, Recency, and Retention
These three metrics are used to count customers and track customer activity irre-
spective of the number of transactions (or dollar value of those transactions) made
by each customer.
A customer is a person or business that buys from the firm.

■ Customer Counts: These are the number of customers of a firm for a specified time
period.
■ Recency: This refers to the length of time since a customer’s last purchase. A six-month
customer is someone who purchased from the firm at least once within the last six
months.
■ Retention Rate: This is the ratio of the number of retained customers to the number
at risk.
In contractual situations, it makes sense to talk about the number of customers cur-
rently under contract and the percentage retained when the contract period runs out.
In non-contractual situations (such as catalog sales), it makes less sense to talk about
the current number of customers, but instead to count the number of customers of a
specified recency.

Purpose: To monitor firm performance in attracting and
retaining customers.
Only recently have most marketers worried about developing metrics that focus on
individual customers. In order to begin to think about managing individual customer
relationships, the firm must first be able to count its customers. Although consistency
in counting customers is probably more important than formulating a precise defini-
tion, a definition is needed nonetheless. In particular, we think the definition of and
the counting of customers will be different in contractual versus non-contractual
situations.

Construction
COUNTING CUSTOMERS
In contractual situations, it should be fairly easy to count how many customers are cur-
rently under contract at any point in time. For instance, Vodafone Australia,1 a global
mobile phone company, was able to report 2.6 million direct customers at the end of the
December quarter.


One complication in counting customers in contractual situations is the handling
of contracts that cover two or more individuals. Does a family plan that includes
five phones but one bill count as one or five? Does a business-to-business contract with
one base fee and charges for each of 1,000 phones in use count as one or 1,000 cus-
tomers? Does the answer to the previous question depend on whether the individual
users pay Vodafone, pay their company, or pay nothing? In situations such as these, the
firm must select some standard definition of a customer (policy holder, member) and
implement it consistently.
A second complication in counting customers in contractual situations is the treatment
of customers with multiple contracts with a single firm. USAA, a global insurance and
diversified financial services association, provides insurance and financial services to the
U.S. military community and their families. Each customer is considered a member,
complete with a unique membership number. This allows USAA to know exactly how
many members it has at any time—more than five million at the end of 2004—most of
whom avail themselves of a variety of member services.
For other financial services companies, however, counts are often listed separately
for each line of business. The 2003 annual report for State Farm Insurance, for exam-
ple, lists a total of 73.9 million policies and accounts with a pie chart showing the
percentage breakdown among auto, homeowners, life, annuities, and so on. Clearly the
73.9 million is a count of policies and not customers. Presumably because some cus-
tomers use State Farm for auto, home, and life insurance, they get double and even
triple counted in the 73.9 million number. Because State Farm knows the names and
addresses of all their policyholders, it seems feasible that they could count how many
individual customers they serve. The fact that State Farm counts policies and not
customers suggests an emphasis on selling policies rather than managing customer
relationships.
Finally, we offer an example of a natural gas company that went out of its way to dou-
ble count customers—defining a customer to be “a consumer of natural gas distributed
in any one billing period at one location through one meter. An entity using gas at sep-
arate locations is considered a separate customer at each location.” For this natural gas
company, customers were synonymous with meters. This is probably a great way to view
things if your job is to install and service meters. It is not such a great way to view things
if your job is to market natural gas.
In non-contractual situations, the ability of the firm to count customers depends on
whether individual customers are identifiable. If customers are not identifiable, firms
can only count visits or transactions. Because Wal-Mart does not identify its shoppers,
its customer counts are nothing more than the number of transactions that go
through the cash registers in a day, week, or year. These “traffic” counts are akin to
turnstile numbers at sporting events and visits to a Web site. In one sense they count


people, but when summed over several periods, they no longer measure separate
individuals. So whereas home attendance at Atlanta Braves games in 19932 was
3,884,720, the number of people attending one or more Braves games that year was
some smaller number.
In non-contractual situations with identifiable customers (direct mail, retailers with fre-
quent shopper cards, warehouse clubs, purchases of rental cars and lodging that require
registration), a complication is that customer purchase activity is sporadic. Whereas the
New York Times knows exactly how many current customers (subscribers) it has, the
sporadic buying of cataloger L.L.Bean’s customers means that it makes no sense to talk
about the number of current L.L.Bean customers. L.L.Bean will know the number of
orders it receives daily, it will know the number of catalogs it mails monthly, but it can-
not be expected to know the number of current customers it has because it is difficult to
define a “current” customer.
Instead, firms in non-contractual situations count how many customers have bought
within a certain period of time. This is the concept of recency —the length of time since
the last purchase. Customers of recency one year or less are customers who bought
within the last year. Firms in non-contractual situations with identifiable customers will
count customers of various recencies.
Recency: The length of time since a customer’s last purchase.
For example, eBay reported 60.5 million active users in the first quarter of 2005. Active
users were defined as the number of users of the eBay platform who bid, bought, or
listed an item within the previous 12-month period. They go on to report that 45.1 mil-
lion active users were reported in the same period a year ago.
Notice that eBay counts “active users” rather than “customers” and uses the concept
of recency to track its number of active users across time. The number of active
(12-month) users increased from 45.1 million to 60.5 million in one year. This tells
the firm that the number of active customers increased due in part to customer acqui-
sition. A measure of how well the firm maintained existing customer relationships is
the percentage of the 45.1 million active customers one year ago who were active in the
previous 12 months. That ratio measure is similar to retention in that it reflects the
percentage of active customers who remained active in the subsequent period.
Retention: Applies to contractual situations in which customers are either retained
or not. Customers either renew their magazine subscriptions or let them run out.
Customers maintain a checking account with a bank until they close it out. Renters pay
rent until they move out. These are examples of pure customer retention situations
where customers are either retained or considered lost for good.


In these situations, firms pay close attention to retention rates.
Retention Rate: The ratio of the number of customers retained to the number
at risk.
If 40,000 subscriptions to Fortune magazine are set to expire in July and the publisher
convinces 26,000 of those customers to renew, we would say that the publisher retained
65% of its subscribers.
The complement of retention is attrition or churn. The attrition or churn rate for the
40,000 Fortune subscribers was 35%.
Notice that this definition of retention is a ratio of the number retained to the number
at risk (of not being retained). The key feature of this definition is that a customer must
be at risk of leaving in order to be counted as a customer successfully retained. This
means that new Fortune subscribers obtained during July are not part of the equation,
nor are the large number of customers whose subscriptions were set to run out in
later months.
Finally, we point out that it sometimes makes better sense to measure retention in
“customer time” rather than “calendar time.” Rather than ask what the firm’s retention
rate was in 2004, it may be more informative to ask what percentage of customers
surviving for three years were retained throughout year four.

The ratio of the total number of customers at the end of the period to the number of
customers at the beginning of the period is not a retention rate. Retention during the
period does affect this ratio, but customer acquisitions also affect the ratio.
The percentage of customers starting the period who remained customers throughout
the period is a lot closer to being a retention rate. This percentage would be a true
retention rate if all the customers starting the period were at risk of leaving during
the period.

Advice on Counting Customers3
Defining the customer properly is critical.
Marketers tend to count “customers” in ways that are easy and consequently get
the wrong answers. They tend to gloss over the fundamental and critically
important step of defining the customer. With the wrong definition, counting
doesn’t matter.


Banks look at “households” because they are “relationship” obsessed (relationship
being defined as the number of products sold to customers with a common account
address). Banks tend to emphasize the number of products sold. No matter that the
household may contain a business owner with nearly all the accounts, a spouse
who banks mostly elsewhere, and children who do not bank at all. Household in
this situation is meaningless. There are at least three “customers” here: business
owner (a great customer), spouse (almost a non-customer), and kids (definitely
non-customers).
Retailers count transactions or “tickets” (cash register receipts), which may cover
stuff sold to Mom, Dad, and the kids, along with Aunt Mary and neighbor Sue. Or,
it may reflect a purchase by a spouse who is buying for his or her partner under
specific instructions. In this circumstance, the spouse is the real customer, with
the other taking on the role of gofer.
Defining the customer is nearly always hard because it requires a clear understanding
of both business strategy and buyer behavior.
Not all “customers” are the same.
Attracting and retaining “customers” cannot be measured for management action
purposes without understanding the differences between customers. Last year, a
major software firm we will call Zapp bought a single copy of a piece of software.
Another company we will call Tancat bought 100 copies. Are these both “customers?”
Of course not. Tancat is almost certainly a customer that needs to be retained and
possibly expanded into other products. Zapp is probably just evaluating the product
in order to stay on top of new software concepts and potentially copy it. One option
is to follow up with Zapp with their one-copy purchase to see what is really going on.
Zapp could become a great “customer” if we understand what motivated their pur-
chase or if we use that purchase to gain a contact base.
Before you count anything, you have to segment your potential and current product
or service users into groups that can be strategically addressed. Some current buyers
like Zapp are actually potential buyers in terms of what you should do about them.
You must count buyers and prospects who are alike in defined ways.
Where is the “customer?”
Large customers often buy independently from each user location. Is Bank of
America the customer, or is each branch office a customer? If Citicorp were to buy
centrally, how could you count it as one customer while Bank of America counts as
hundreds of customers?
Who is the “customer?”
Defining who is the customer is even trickier. Many “customers” are not those who
place the order with your salespeople. The real customer is deep within the bowels of


the buyer organization, someone who may take a great deal of effort to even identify.
The account name may be GM, but the real customer may be Burt Cipher, an engi-
neer in some unknown facility. Or, the Ford buyer may have consolidated orders
from several individuals scattered across the country. In this case, Ford is not the cus-
tomer for anything but billing purposes. So, what do you count?
Even more common is the multi-headed customer. Buying decisions are made by
several people. Different people may be central to a decision at different times or for
different products. Big companies have sales teams dedicated to selling into such
buying groups. Although they may be counted as a single customer, the dynamics of
their buying decision is substantially more complicated than decisions made by a
single individual.
Apparel retailers who sell pre-teen clothing have at least two customers: Mom and
the pre-teen wearer. Do you count one or both as customers? Marketing might want
to treat each as a customer for deciding how to design and place ads. The store might
treat them both as a single customer or choose the pre-teen as their target.
The key takeaway is that customer definition for counting depends fundamentally on
the purpose of the count. You may have to count the same “customer” in different
ways for different purposes. There is no universal customer definition.

5.2 Customer Profit
Customer profit (CP) is the profit the firm makes from serving a customer or cus-
tomer group over a specified period of time.
Calculating customer profitability is an important step in understanding which cus-
tomer relationships are better than others. Often, the firm will find that some cus-
tomer relationships are unprofitable. The firm may be better off (more profitable)
without these customers. At the other end, the firm will identify its most profitable
customers and be in a position to take steps to ensure the continuation of these most
profitable relationships.

Purpose: To identify the profitability of individual customers.
Companies commonly look at their performance in aggregate. A common phrase
within a company is something like: “We had a good year, and the business units
delivered $400,000 in profits.” When customers are considered, it is often using an
average such as “We made a profit of $2.50 per customer.” Although these can be useful


metrics, they sometimes disguise an important fact that not all customers are equal
and, worse yet, some are unprofitable. Simply put, rather than measuring the “average
customer,” we can learn a lot by finding out what each customer contributes to our
bottom line.4
Customer Profitability: The difference between the revenues earned from and the
costs associated with the customer relationship during a specified period.
The overall profitability of the company can be improved by treating dissimilar
customers differently.
In essence, think of three different tiers of customer:

1. Top Tier customers—REWARD: Your most valuable customers are the ones you
most want to retain. They should receive more of your attention than any other
group. If you lose these guys, your profit suffers the most. Look to reward them
in ways other than simply lowering your price. These customers probably value
what you do the most and may not be price-sensitive.
2. Second Tier customers—GROW: The customers in the middle—with middle to
low profits associated with them—might be targeted for growth. Here you have
customers whom you may be able to develop into Top Tier customers. Look to
the share of customer metrics described in Section 5.3 to help figure out which
customers have the most growth potential.
3. Third Tier customers—FIRE: The company loses money on servicing these
people. If you cannot easily promote them to the higher tiers of profitability,
you should consider charging them more for the services they currently con-
sume. If you can recognize this group beforehand, it may be best not to acquire
these customers in the first place.
A database that can analyze the profitability of customers at an individual level can be a
competitive advantage. If you can figure out profitability by customer, you have a
chance to defend your best customers and maybe even poach the most profitable con-
sumers from your competitors.

Construction
In theory, this is a trouble-free calculation. Find out the cost to serve each customer and
the revenues associated with each customer for a given period. Do the subtraction to get
profit for the customer and sort the customers based on profit. Although painless in the-
ory, large companies with a multitude of customers will find this a major challenge even
with the most sophisticated of databases.


To do the analysis with large databases, it may be necessary to abandon the notion of
calculating profit for each individual customer and work with meaningful groups of
customers instead.
After you have the sorted list of customer profits (or customer-group profits), the cus-
tom is to plot cumulative percentage of total profits versus cumulative percentage of
total customers. Given that the customers are sorted from highest to lowest profit, the
resulting graph usually looks something like the head of a whale.
Profitability will increase sharply and tail off from the very beginning. (Remember, our
customers have been sorted from most to least profitable.) Whenever there are some
negative profit customers, the graph reaches a peak—above 100%—as profit per
customer moves from positive to negative. As we continue through the negative-profit
customers, cumulative profits decrease at an ever-increasing rate. The graph always ends
at 100% of the customers accounting for 100% of the total profit.
Robert Kaplan (co-developer of Activity-Based Costing and the Balanced Scorecard)
likes to refer to these curves as “whale curves.”5 In Kaplan’s experience, the whale
curve usually reveals that the most profitable 20% of customers can sometimes gener-
ate between 150% and 300% of total profits so that the resulting curve resembles
a sperm whale rising above the water’s surface. See Figure 5.2 for an example of a
whale curve.

EXAMPLE: A catalog retailer has grouped customers in 10 deciles based on profitabil-
ity (see Table 5.1 and Figure 5.1). (A decile is a tenth of the population, so 0-10% is the
most profitable 10% of customers.)

Table 5.1 Customer Profitability Ranked by Profitability
90–100%
10–20%

20–30%

30–40%

40–50%

50–60%

60–70%

70–80%

80–90%

Customers
0–10%

Decile by
Profitability

Band ($m) $100 $50 $25 $10 $5 $3 $2 $0 ($8) ($20)
Profitability
% of Total 60% 30% 15% 6% 3% 2% 1% 0% 5% 12%
Profits

Here we have a clear illustration that if they were no longer to serve the least profitable
20% of customers, they would be $28 million better off.


Customer Profitability

70%

60%

50%
% of Company Profits

40%

30%

20%

10%

0%
0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90% 90-
100%
–10%

–20%

Decile

Figure 5.1 Customer Profitability by Decile

Table 5.2 Cumulative Profitability Peaks Before All Customers Are Served 90–100%
10–20%

20–30%

30–40%

40–50%

50–60%

60–70%

70–80%

80–90%

Customers
0–10%

Decile by
Profitability

Cumulative $100 $150 $175 $185 $190 $193 $195 $195 $187 $167
Profits
Cumulative 59.9 89.8 104.8 110.8 113.8 115.6 116.8 116.8 112.0 100.0
Profits %

Table 5.2 presents this same customer information in cumulative form. Cumulative
profits plotted across deciles begins to look like a whale with a steeply rising ridge reach-
ing a peak of total profitability above 100% and tapering off thereafter (see Figure 5.2).


Cumulative Profits

140%

120%

100%

80%
Cum Profits

60%

40%

20%

0%
%

0%

0%

0%

0%

0%

0%

0%
0%

%
10

00
–2

–3

–4

–5

–6

–7

–9
–8
0–

–1
10

20

30

40

50

60

80
70

90
Decile

Figure 5.2 The Whale Curve

Measuring customer profitability requires detailed information. Assigning revenues to
customers is often the easy part; assigning your costs to customers is much harder. The
cost of goods sold obviously gets assigned to the customers based on the goods each
customer purchased. Assigning the more indirect costs may require the use of some
form of activity-based costing (ABC) system. Finally, there may be some categories of
costs that will be impossible to assign to the customer. If so, it is probably best to keep
these costs as company costs and be content with the customer profit numbers adding
up to something less than the total company profit.
When considering the profits from customers, it must be remembered that most things
change over time. Customers who were profitable last year may not be profitable


this year. Because the whale curve reflects past performance, we must be careful when
using it to make decisions that shape the future. For example, we may very well want to
continue a relationship that was unprofitable in the past if we know things will change
for the better in the future. For example, banks typically offer discount packages to stu-
dents to gain their business. This may well show low or negative customer profits in the
short term. The “plan” is that future profits will compensate for current losses. Customer
lifetime value (addressed in Section 5.3) is a forward-looking metric that attempts to
account for the anticipated future profitability of each customer relationship.
When capturing customer information to decide which customers to serve, it is impor-
tant to consider the legal environment in which the company operates. This can change
considerably across countries, where there may be anti-discrimination laws and special
situations in some industries. For instance, public utilities are sometimes obligated to
serve all customers.
It is also worth remembering that intrusive capturing of customer-specific data can
damage customer relationships. Some individuals will be put off by excess data gather-
ing. For a food company, it may help to know which of your customers are on a diet. But
the food company’s management should think twice before adding this question to their
next customer survey.
Sometimes there are sound financial reasons for continuing to serve unprofitable cus-
tomers. For example, some companies rely on network effects. Take the case of the
United States Postal Service—part of its strength is the ability to deliver to the whole
country. It may superficially seem profitable to stop deliveries to remote areas. But when
that happens, the service becomes less valuable for all customers. In short, sometimes
unprofitable customer relationships are necessary for the firm to maintain their
profitable ones.
Similarly, companies with high fixed costs that have been assigned to customers during
the construction of customer profit must ask whether those costs will go away if they
terminate unprofitable customer relationships. If the costs do not go away, ending
unprofitable relationships may only serve to make the surviving relationships look even
less profitable (after the reallocation of costs) and result in the lowering of company
profits. In short, make certain that the negative profit goes away if the relationship is ter-
minated. Certainly the revenue and cost of goods sold will go away, but if some of the
other costs do not, the firm could be better off maintaining a negative profit relation-
ship as it contributes to covering fixed cost (refer to Sections 3.4 and 3.6).
Abandoning customers is a very sensitive practice, and a business should always con-
sider the public relations consequences of such actions. Similarly, when you get rid of a
customer, you cannot expect to attract them back very easily should they migrate into
your profitable segment.


Finally, because the whale curve examines cumulative percentage of total profits, the
numbers are very sensitive to the dollar amount of total profit. When the total dollar
profit is a small number, it is fairly easy for the most profitable customers to represent a
huge percentage of that small number. So when you hear that 20% of the firm’s cus-
tomers represent 350% of the firm’s profit, one of the first things you should consider is
the total dollar value of profits. If that total is small, 350% of it can also be a fairly small
number of dollars. To cement this idea, ask yourself what the whale curve would look
like for a firm with $0 profit.

5.3 Customer Lifetime Value
Customer lifetime value is the dollar value of a customer relationship based on the
present value of the projected future cash flows from the customer relationship.
When margins and retention rates are constant, the following formula can be used to
calculate the lifetime value of a customer relationship:

Customer Lifetime Value ($) Margin ($) * Retention Rate (%)
1 Discount Rate (%) Retention Rate (%)

Customer lifetime value (CLV) is an important concept in that it encourages firms to
shift their focus from quarterly profits to the long-term health of their customer rela-
tionships. Customer lifetime value is an important number because it represents an
upper limit on spending to acquire new customers.

Purpose: To assess the value of each customer.
As Don Peppers and Martha Rogers are fond of saying, “some customers are more equal
than others.”6 We saw a vivid illustration of this in the last section, which examined the
profitability of individual customer relationships. As we noted, customer profit (CP) is
the difference between the revenues and the costs associated with the customer rela-
tionship during a specified period. The central difference between CP and customer
lifetime value (CLV) is that CP measures the past and CLV looks forward. As such, CLV
can be more useful in shaping managers’ decisions but is much more difficult to quan-
tify. Quantifying CP is a matter of carefully reporting and summarizing the results of
past activity, whereas quantifying CLV involves forecasting future activity.
Customer Lifetime Value (CLV): The present value of the future cash flows
attributed to the customer relationship.


The concept of present value will be talked about in more detail in Section 10.4. For
now, you can think of present value as the discounted sum of future cash flows. We dis-
count (multiply by a carefully selected number less than one) future cash flows before
we add them together to account for the fact that there is a time value of money. The
time value of money is another way of saying that everyone would prefer to get paid
sooner rather than later and everyone would prefer to pay later rather than sooner. This
is true for individuals (the sooner I get paid, the sooner I can pay down my credit card
balance and avoid interest charges) as well as for firms. The exact discount factors used
depend on the discount rate chosen (10% per year as an example) and the number of
periods until we receive each cash flow (dollars received 10 years from now must be dis-
counted more than dollars received five years in the future).
The concept of CLV is nothing more than the concept of present value applied to cash
flows attributed to the customer relationship. Because the present value of any stream
of future cash flows is designed to measure the single lump sum value today of the
future stream of cash flows, CLV will represent the single lump sum value today of the
customer relationship. Even more simply, CLV is the dollar value of the customer
relationship to the firm. It is an upper limit on what the firm would be willing to pay
to acquire the customer relationship as well as an upper limit on the amount the firm
would be willing to pay to avoid losing the customer relationship. If we view a
customer relationship as an asset of the firm, CLV would present the dollar value of
that asset.

COHORT AND INCUBATE
One way to project the value of future customer cash flows is to make the heroic
assumption that the customers acquired several periods ago are no better or worse (in
terms of their CLV) than the ones we currently acquire. We then go back and collect
data on a cohort of customers all acquired at about the same time and carefully recon-
struct their cash flows over some finite number of periods. The next step is to discount
the cash flows for each customer back to the time of acquisition to calculate that cus-
tomer’s sample CLV and then average all of the sample CLVs together to produce an
estimate of the CLV of each newly acquired customer. We refer to this method as the
“cohort and incubate” approach. Equivalently, one can calculate the present value of the
total cash flows from the cohort and divide by the number of customers to get the aver-
age CLV for the cohort. If the value of customer relationships is stable across time, the
average CLV of the cohort sample is an appropriate estimator of the CLV of newly
acquired customers.
As an example of this cohort and incubate approach, Berger, Weinberg, and Hanna
(2003) followed all the customers acquired by a cruise-ship line in 1993. The 6,094
customers in the cohort of 1993 were tracked (incubated) for five years. The total net


present value of the cash flows from these customers was $27,916,614. These flows
included revenues from the cruises taken (the 6,094 customers took 8,660 cruises
over the five-year horizon), variable cost of the cruises, and promotional costs. The total
five-year net present value of the cohort expressed on a per-customer basis came out
to be $27,916,614/6,094 or $4,581 per customer. This is the average five-year CLV for
the cohort.

“Prior to this analysis, [cruise-line] management would never spend more than
$3,314 to acquire a passenger . . . Now, aware of CLV (both the concept and the
actual numerical results), an advertisement that [resulted in a cost per acquisition of
$3 to $4 thousand] was welcomed—especially because the CLV numbers are
conservative (again, as noted, the CLV does not include any residual business after
five years.)”7
The cohort and incubate approach works well when customer relationships are
stationary—changing slowly over time. When the value of relationships changes
slowly, we can use the value of incubated past relationships as predictive of the value of
new relationships.
In situations where the value of customer relationships changes more rapidly, firms
often use a simple model to forecast the value of those relationships. By a model, we
mean some assumptions about how the customer relationship will unfold. If the model
is simple enough, it may even be possible to find an equation for the present value
of our model of future cash flows. This makes the calculation of CLV even easier
because it now requires only the substitution of numbers for our situation into the
equation for CLV.
Next, we will explain what is perhaps the simplest model for future customer cash flows
and the equation for the present value of those cash flows. Although it’s not the only
model of future customer cash flows, this one gets used the most.

Construction
The model for customer cash flows treats the firm’s customer relationships as something
of a leaky bucket. Each period, a fraction (1 less the retention rate) of the firm’s cus-
tomers leave and are lost for good.
The CLV model has only three parameters: 1) constant margin (contribution after
deducting variable costs including retention spending) per period, 2) constant retention
probability per period, and 3) discount rate. Furthermore, the model assumes that in
the event that the customer is not retained, they are lost for good. Finally, the model
assumes that the first margin will be received (with probability equal to the retention
rate) at the end of the first period.


The one other assumption of the model is that the firm uses an infinite horizon when it
calculates the present value of future cash flows. Although no firm actually has an infi-
nite horizon, the consequences of assuming one are discussed in the following.
Customer Lifetime Value: The CLV formula8 multiplies the per-period cash margin
(hereafter we will just use the term “margin”) by a factor that represents the present
value of the expected length of the customer relationship:
Retention Rate (%)
Customer Lifetime Value ($) Margin ($) *
Under the assumptions of the model, CLV is a multiple of the margin. The multiplica-
tive factor represents the present value of the expected length (number of periods) of
the customer relationship. When retention equals 0, the customer will never be retained,
and the multiplicative factor is zero. When retention equals 1, the customer is always
retained, and the firm receives the margin in perpetuity. The present value of the margin
in perpetuity turns out to be Margin/Discount Rate. For retention values in between,
the CLV formula tells us the appropriate multiplier.

EXAMPLE: An Internet Service Provider (ISP) charges $19.95 per month. Variable
costs are about $1.50 per account per month. With marketing spending of $6 per year,
their attrition is only 0.5% per month. At a monthly discount rate of 1%, what is the CLV
of a customer?
Contribution Margin ($19.95 $1.50 $6 12) $17.95
Retention Rate 0.995
Discount Rate 0.01

Customer Lifetime Value (CLV) Margin * Retention Rate (%)

CLV $17.95 * [0.995/(1 0.01 0.995)]
CLV [$17.95] * [66.33]
CLV $1,191

The retention rate (and by extension the attrition rate) is a driver of customer value.
Very small changes can make a major difference to the lifetime value calculated.
Accuracy in this parameter is vital to meaningful results.


The retention rate is assumed to be constant across the life of the customer relation-
ship. For products and services that go through a trial, conversion, and loyalty
progression, retention rates will increase over the lifetime of the relationship. In
those situations, the model explained here might be too simple. If the firm wants to
estimate a sequence of retention rates, a spreadsheet model might be more useful in
calculating CLV.
The discount rate is also a sensitive driver of the lifetime value calculation—as with
retention, seemingly small changes can make major differences to customer lifetime
value. The discount rate should be chosen with care.
The contribution is assumed to be constant across time. If margin is expected to increase
over the lifetime of the customer relationship, the simple model will not apply.
Take care not to use this CLV formula for relationships in which customer inactivity
does not signal the end of the relationship. In catalog sales, for example, a small per-
centage of the firm’s customers purchase from any given catalog. Don’t confuse the per-
centage of customers active in a given period (relevant for the cataloger) with the
retention rates in this model. If customers often return to do business with the firm after
a period of inactivity, this CLV formula does not apply.
Customer Lifetime Value (CLV) with Initial Margin: One final source of confusion
concerns the timing assumptions inherent in the model. The first cash flow accounted
for in the model is the margin received at the end of one period with probability equal
to the retention rate. Other models also include an initial margin received at the begin-
ning of the period. If a certain receipt of an initial margin is included, the new CLV will
equal the old CLV plus the initial margin. Furthermore, if the initial margin is equal to
all subsequent margins, there are at least two ways to write formulas for the CLV that
include the initial margin:
Retention Rate (%)
CLV with Initial Margin ($) Margin ($) *
Margin ($) 1 Discount Rate (%) Retention Rate (%)

or

Margin ($) * 1 Discount Rate (%)

The second formula looks just like the original formula with 1 Discount Rate taking
the place of the retention rate in the numerator of the multiplicative factor. Just remem-
ber that the new CLV formula and the original CLV formula apply to the same situa-
tions and differ only in the treatment of an initial margin. This new CLV formula
includes it, whereas the original CLV formula does not.


THE INFINITE HORIZON ASSUMPTION
In some industries and companies it is typical to calculate four- or five-year customer
values instead of using the infinite time horizon inherent in the previous formulas. Of
course, over shorter periods customer retention rates are less likely to be affected by
major shifts in technology or competitive strategies and are more likely to be captured
by historical retention rates. For managers, the question is “Does it make a difference
whether I use the infinite time horizon or (for example) the five-year customer value?”
The answer to this question is yes, sometimes, it can make a difference because the
value over five years can be less than 70% of the value over an infinite horizon (see
Table 5.3).
Table 5.3 calculates the percentages of (infinite horizon) CLV accruing in the first five
years. If retention rates are higher than 80% and discount rates are lower than 20%,
differences in the two approaches will be substantial. Depending on the strategic risks
that companies perceive, the additional complexities of using a finite horizon can be
informative.

Table 5.3 Finite-Horizon CLV As a Percentage of Infinite-Horizon CLV

Percent of CLV Accruing in First Five Years

Discount
Rates Retention Rates

40% 50% 60% 70% 80% 90%

2% 99% 97% 93% 85% 70% 47%

4% 99% 97% 94% 86% 73% 51%
6% 99% 98% 94% 87% 76% 56%

8% 99% 98% 95% 89% 78% 60%

10% 99% 98% 95% 90% 80% 63%

12% 99% 98% 96% 90% 81% 66%

14% 99% 98% 96% 91% 83% 69%

16% 100% 99% 96% 92% 84% 72%

18% 100% 99% 97% 93% 86% 74%

20% 100% 99% 97% 93% 87% 76%


5.4 Prospect Lifetime Value Versus Customer Value
Prospect lifetime value is the expected value of a prospect. It is the value expected
from the prospect minus the cost of prospecting. The value expected from the
prospect is the expected fraction of prospects who will make a purchase times the
sum of the average margin the firm makes on the initial purchase and the CLV of
the newly acquired customer.
Only if prospect lifetime value is positive should the firm proceed with the planned
acquisition spending.

Purpose: To account for the lifetime value of a newly acquired customer
(CLV) when making prospecting decisions.
One of the major uses of CLV is to inform prospecting decisions. A prospect is someone
whom the firm will spend money on in an attempt to acquire her or him as a customer.
The acquisition spending must be compared not just to the contribution from the
immediate sales it generates but also to the future cash flows expected from the newly
acquired customer relationship (the CLV). Only with a full accounting of the value of
the newly acquired customer relationship will the firm be able to make an informed,
economic prospecting decision.

Construction
The expected prospect lifetime value (PLV) is the value expected from each prospect
minus the cost of prospecting. The value expected from each prospect is the acquisition
rate (the expected fraction of prospects who will make a purchase and become cus-
tomers) times the sum of the initial margin the firm makes on the initial purchases and
the CLV. The cost is the amount of acquisition spending per prospect. The formula for
expected PLV is as follows:
Prospect Lifetime Value ($) Acquisition Rate (%) * [Initial Margin ($) CLV ($)]
Acquisition Spending ($)
If PLV is positive, the acquisition spending is a wise investment. If PLV is negative, the
acquisition spending should not be made.
The PLV number will usually be very small. Although CLV is sometimes in the hundreds
of dollars, PLV can come out to be only a few pennies. Just remember that PLV applies
to prospects, not customers. A large number of small but positive-value prospects can
add to a considerable amount of value for a firm.


EXAMPLE: A service company plans to spend $60,000 on an advertisement reaching
75,000 readers. If the service company expects the advertisement to convince 1.2% of the
readers to take advantage of a special introductory offer (priced so low that the firm
makes only $10 margin on this initial purchase) and the CLV of the acquired customers
is $100, is the advertisement economically attractive?
Here Acquisition Spending is $0.80 per prospect, the expected acquisition rate is 0.012,
and the initial margin is $10. The expected PLV of each of the 75,000 prospects is
PLV 0.012 * ($10 $100) $0.80
$0.52
The expected PLV is $0.52. The total expected value of the prospecting effort will be
75,000 * $0.52 $39,000. The proposed acquisition spending is economically attractive.
If we are uncertain about the 0.012 acquisition rate, we might ask what the acquisition
rate from the prospecting campaign must be in order for it to be economically successful.
We can get that number using Excel’s goal seek function to find the acquisition rate that
sets PLV to zero. Or we can use a little algebra and substitute $0 in for PLV and solve for
the break-even acquisition rate:
Break-Even Acquisition Rate
Initial Margin ($) CLV ($)
$0.80
0.007273
$10 $100

The acquisition rate must exceed 0.7273% in order for the campaign to be successful.

In addition to the CLV of the newly acquired customers, the firm needs to know
the planned amount of acquisition spending (expressed on a per-prospect basis), the
expected success rate (the fraction of prospects expected to become customers), and the
average margin the firm will receive from the initial purchases of the newly acquired
customers. The initial margin number is needed because CLV as defined in the previous
section accounts for only the future cash flows from the relationship. The initial cash
flow is not included in CLV and must be accounted for separately. Note also that the ini-
tial margin must account for any first-period retention spending.
Perhaps the biggest challenge in calculating PLV is estimating CLV. The other terms
(acquisition spending, acquisition rate, and initial margin) all refer to flows or outcomes
in the near future, whereas CLV requires longer-term projections.


Another caution worth mentioning is that the decision to spend money on customer
acquisition whenever PLV is positive rests on an assumption that the customers
acquired would not have been acquired had the firm not spent the money. In other
words, our approach gives the acquisition spending “full credit” for the subsequent cus-
tomers acquired. If the firm has several simultaneous acquisition efforts, dropping one
of them might lead to increased acquisition rates for the others. Situations such as these
(where one solicitation cannibalizes another) require a more complicated analysis.
The firm must be careful to search for the most economical way to acquire new cus-
tomers. If there are alternative prospecting approaches, the firm must be careful not to
simply go with the first one that gives a positive projected PLV. Given a limited number
of prospects, the approach that gives the highest expected PLV should be used.
Finally, we want to warn you that there are other ways to do the calculations necessary
to judge the economic viability of a given prospecting effort. Although these other
approaches are equivalent to the one presented here, they differ with respect to what
gets included in “CLV.” Some will include the initial margin as part of “CLV.” Others will
include both the initial margin and the expected acquisition cost per acquired customer
as part of “CLV.” We illustrate these two approaches using the service company example.

EXAMPLE: A service company plans to spend $60,000 on an advertisement reaching
75,000 readers. If the service company expects the advertisement to convince 1.2% of the
readers to take advantage of a special introductory offer (priced so low that the firm
makes only $10 margin on this initial purchase) and the CLV of the acquired customers
is $100, is the advertisement economically attractive?
If we include the initial margin in “CLV” we get
“CLV” [with Initial Margin ($)] Initial Margin ($) CLV ($)
$10 $110 $110
The expected PLV is now
PLV ($) Acquisition Rate (%) * “CLV” [with Initial Margin ($)] Acquisition Cost ($)
0.012 * $110 $0.85 $0.52
This is the same number as before calculated using a slightly different “CLV”—one that
includes the initial margin.
We illustrate one final way to do the calculations necessary to judge the economics of a
prospecting campaign. This last way does things on a per-acquired-customer basis using
a “CLV” that includes both initial margin and an allocated acquisition spending. The
thinking goes as follows: The expected value of a new customer is $10 now plus $100
from future sales, or $110 in total. The expected cost to acquire a customer is the total
cost of the campaign divided by the expected number of new customers. This average


acquisition cost is calculated as $60,000 /(0.012 * 75,000) $66.67. The expected value
of a new customer net of the expected acquisition cost per customer is $110 $66.67
$43.33. Because this new “net” CLV is positive, the campaign is economically attractive.
Some will even label this $43.33 number as the “CLV” of a new customer.
Notice that $43.33 times the 900 expected new customers equals $39,000, the same total
net value from the campaign calculated in the original example as the $0.52 PLV times
the 75,000 prospects. The two ways to do the calculations are equivalent.

5.5 Acquisition Versus Retention Cost
The firm’s average acquisition cost is the ratio of acquisition spending to the number
of customers acquired. The average retention cost is the ratio of retention spending
directed toward a group of customers to the number of those customers successfully
retained.
Average Acquisition Cost ($)
Number of Customers Acquired (#)

Retention Spending ($)
Average Retention Cost ($)
Number of Customers Retained (#)

These two metrics help the firm monitor the effectiveness of two important cate-
gories of marketing spending.

Purpose: To determine the firm’s cost of acquisition and retention.
Before the firm can optimize its mix of acquisition and retention spending, it must first
assess the status quo. At the current spending levels, how much does it cost the firm (on
average) to acquire new customers, and how much is it spending (on average) to retain
its existing customers? Does it cost five times as much to acquire a new customer as it
does to retain an existing one?

Construction
Average Acquisition Cost: This represents the average cost to acquire a customer
and is the total acquisition spending divided by the number of new customers
acquired.


Average Acquisition Cost ($)
Number of Customers Acquired (#)

Average Retention Cost: This represents the average “cost” to retain an existing
customer and is the total retention spending divided by the number of customers
retained.
Retention Spending ($)
Average Retention Cost ($)
Number of Customers Retained (#)

EXAMPLE: During the past year, a regional pest control service spent $1.4 million and
acquired 64,800 new customers. Of the 154,890 customer relationships in existence at the
start of the year, only 87,957 remained at the end of the year, despite about $500,000
spent during the year in attempts to retain the 154,890 customers. The calculation of
average acquisition cost is relatively straightforward. A total of $1.4 million resulted in
64,800 new customers. The average acquisition cost is $1,400/64.8 $21.60 per cus-
tomer. The calculation of average retention cost is also straightforward. A total of
$500,000 resulted in 87,957 retained customers. The average yearly retention cost is
$500,000 / 87,957 $5.68. Thus, for the pest control firm, it cost about four times as
much to acquire a new customer as it did to retain an existing one.

For any specific period, the firm needs to know the total amount it spent on customer
acquisition and the number of new customers that resulted from that spending. With
respect to customer retention, the firm needs to measure the total amount spent during
the period attempting to retain the customers in existence at the start of the period and
the number of the existing customers successfully retained at the end of the period.
Notice that retention spending directed at customers acquired within the period is not
included in this figure. Similarly, the number retained refers only to those retained from
the pool of customers in existence at the start of the period. Thus, the average retention
cost calculated will be associated with the length of the period in question. If the period
is a year, the average retention cost will be a cost per year per customer retained.
The calculation and interpretation of average acquisition cost is much easier than the
calculation and interpretation of average retention cost. This is so because it is often
possible to isolate acquisition spending and count the number of new customers
that resulted from that spending. A simple division results in the average cost to acquire


a customer. The reasonable assumption underlying this calculation is that the new cus-
tomers would not have been acquired had it not been for the acquisition spending.
Things are not nearly so clear when it comes to average retention cost. One source of
difficulty is that retention rates (and costs) depend on the period of time under consid-
eration. Yearly retention is different from monthly retention. The cost to retain a
customer for a month will be less than the cost to retain a customer for a year. Thus,
the definition of average retention cost requires a specification of the time period
associated with the retention.
A second source of difficulty stems from the fact that some customers will be retained
even if the firm spends nothing on retention. For this reason it can be a little misleading
to call the ratio of retention spending to the number of retained customers the average
retention cost. One must not jump to the conclusion that retention goes away if
the retention spending goes away. Nor should one assume that if the firm increases the
retention budget by the average retention cost that it will retain one more customer. The
average retention cost number is not very useful to help make retention budgeting
decisions.
One final caution involves the firm’s capability to separate spending into acquisition and
retention classifications. Clearly there can be spending that works to improve both the
acquisition and retention efforts of the firm. General brand advertisements, for exam-
ple, serve to lower the cost of both acquisition and retention. Rather than attempt to
allocate all spending as either acquisition or retention, we suggest that it is perfectly
acceptable to maintain a separate category that is neither acquisition nor retention.

Berger, Weinberg, and Hanna. (2003). “Customer Lifetime Value Determination and Strategic
Implications for a Cruise-Ship Line,” Database Marketing and Customer Strategy Management,
11(1).
Blattberg, R.C., and S.J. Hoch. (1990). “Database Models and Managerial Intuition: 50%
Model 50% Manager,” Management Science, 36(8), 887–899.
Gupta, S., and Donald R. Lehmann. (2003). “Customers As Assets,” Journal of Interactive
Marketing, 17(1).
Kaplan, R.S., and V.G. Narayanan. (2001). “Measuring and Managing Customer Profitability,”
Journal of Cost Management, September/October: 5–15.
Little, J.D.C. (1970). “Models and Managers: The Concept of a Decision Calculus,” Management
Science, 16(8), B-466; B-485.
McGovern, G.J., D. Court, J.A. Quelch, and B. Crawford. (2004). “Bringing Customers into the
Boardroom,” Harvard Business Review, 82(11), 70–80.


Much, J.G., Lee S. Sproull, and Michal Tamuz. (1989). “Learning from Samples of One or Fewer,”
Organization Science: A Journal of the Institute of Management Sciences, 2(1), 1–12.
Peppers, D., and M. Rogers. (1997). Enterprise One-to-One: Tools for Competing in the Interactive
Age (1st ed.), New York: Currency Doubleday.
Pfeifer, P.E., M.E. Haskins, and R.M. Conroy. (2005). “Customer Lifetime Value, Customer
Profitability, and the Treatment of Acquisition Spending,” Journal of Managerial Issues, 17(1),
11–25.


6
SALES FORCE AND CHANNEL
MANAGEMENT

Introduction

Sales Force Coverage Facings and Share of Shelf
Sales Force Goals Out-of-Stock and Service Levels
Sales Force Results Inventory Turns
Sales Force Compensation Markdowns
Pipeline Analysis Gross Margin Return on Inventory
Investment (GMROII)
Numeric Distribution, ACV
Distribution, and PCV Distribution Direct Product Profitability (DPP)

This chapter deals with push marketing. It describes how marketers measure the ade-
quacy and effectiveness of the systems that provide customers with reasons and oppor-
tunities to buy their products.
The first sections discuss sales force metrics. Here, we list and define the most common
measures for determining whether sales force effort and geographic coverage are
adequate. We discuss pipeline analysis, which is useful in making sales forecasts and in
allocating sales force effort to different stages of the selling process. Pipeline metrics are
used to examine a sequence of selling activities, from lead generation, through follow-
up, to conversion and sales. Although the most important of these represents the

181

percentage of initial leads who ultimately buy, other measures of activity, productivity,
efficiency, and cost can be useful at each stage of the selling process.
In further sections of this chapter, we discuss measures of product distribution
and availability. For manufacturers who approach their market through resellers,
three key metrics provide an indication of “listings”—the percentage of potential
outlets that stock their products. These include numeric distribution, which is
unweighted; ACV, the industry standard; and PCV, a category-specific measure of
product availability.
Marketing logistics tracking metrics are used to measure the operational effectiveness of
the systems that service retailers and distributors. Inventory turns, out-of-stocks, and
service levels are key factors in this area.
At the retail level, gross margin return on inventory investment (GMROII) and direct
product profitability (DPP) offer SKU-specific metrics of product performance, com-
bining movement rates, gross margins, costs of inventory, and other factors.

6.1 Workload Hours required to Prospect numbers To assess the
service clients and may be debatable. number of sales-
prospects. Time spent trying people required
to convert to service a terri-
prospects can tory, and to
vary by territory, ensure balanced
salesperson, workloads.
and potential
client.
6.1 Sales Potential This comprises Doesn’t assess the To determine
Forecast the number of likelihood of con- sales targets. Can
prospects and verting “poten- also help identify
their buying tial” accounts. territories worthy
power. Definitions of of an allocation
buying power are of limited sales
more an art than resources.
a science.


6.2 Sales Goal Individual sales Setting individual To set targets for
projections may targets on the individual sales-
be based on a basis of prior year people and for
salesperson’s sales can discour- territories.
share of fore- age optimal
casted sales, on performance, as
prior year sales strong perform-
and a share of ance in one year
increased district leads to more
projections, or on aggressive targets
a management- in the next.
designed weight-
ing system.
6.3 Sales Force Effectiveness Depends on fac- To assess the
Effectiveness metrics analyze tors that also performance of
sales in the con- affect sales poten- a salesperson
text of various tial and workload. or team.
criteria, including
calls, contacts,
potential
accounts, active
accounts, buying
power of territory,
and expenses.
6.4 Compensation Total payments Perceived rela- To motivate
made to a sales- tionship between maximum sales
person, typically incentive reward effort. To enable
consisting of and controllable salespeople and
base salary, activities may management to
bonus, and/or vary widely track progress
commission. among industries toward goals.
and firms.

6.4 Break-Even Sales revenue, Margins may vary To determine the
Number of multiplied by across products, appropriate
Employees margin net of time, and sales- personnel level
commission, people. Sales are for a projected
divided by cost not independent sales volume.
per staff member. of the number of
salespeople.
Continues

Chapter 6 Sales Force and Channel Management 183

6.5 Sales Funnel, Portrayal of the Funnel dimen- To monitor sales
Sales Pipeline number of clients sions depend on effort and project
and potential type of business future sales.
clients at various and definition of
stages of the sales potential clients.
cycle.
6.6 Numeric Percentage of Outlets’ size or To assess the
Distribution outlets in a sales levels are not degree to which a
defined universe reflected in this brand or product
that stock a par- measure. has penetrated its
ticular brand or Boundaries by potential
product. which distribu- channels.
tion universe is
defined may be
arbitrary.
6.6 All Commodity Numeric distribu- Reflects sales of To assess the
Volume (ACV) tion, weighted by “all commodi- degree to which a
penetrated out- ties,” but may not brand or product
lets’ share of sales reflect sales of the has access to retail
of all product relevant product traffic.
categories. or category.
6.6 Product Category Numeric distribu- Strong indicator To assess the
Volume (PCV) tion, weighted of share potential, degree to which a
by penetrated but may miss brand or product
outlets’ share of opportunities to has access to
sales of the expand category. established outlets
relevant product for its category.
category.
6.6 Total Distribution Usually based on Strong indicator To assess the
ACV or PCV. of the distribu- extent to which a
Sums the relevant tion of a product product line is
measures for line, as opposed available.
each SKU in a to an individual
brand or product SKU.
line.



6.6 Category The ratio of a Same as for ACV To assess whether
Performance PCV to ACV and PCV. a brand’s distri-
Ratio distribution. bution or a par-
ticular retailer is
performing above
or below average
for the category.

6.7 Out-of-Stock Percentage of Out-of-stocks can To monitor the
outlets that “list” be measured in ability of logistics
or normally stock Numeric, ACV, or systems to match
a product or PCV terms. supply with
brand, but have demand.
none available
for sale.

6.7 Inventories Total amount of May be held at To calculate
product or brand different levels ability to meet
available for sale and valued in demand and
in a channel. ways that may or determine chan-
may not reflect nel investments.
promotional
allowances and
discounts.
6.8 Markdowns Percentage dis- For many prod- To determine
count from the ucts, a certain whether channel
regular selling percentage of sales are being
price. markdowns are made at planned
expected. Too margins.
few markdowns
may reflect
“under-ordering.”
If markdowns are
too high, the
opposite may
be true.

Continues



6.8 Direct Product The adjusted Cost allocation is To identify
Profitability gross margin of often imprecise. profitable SKUs
(DPP) products, less Some products and realistically
direct product may be intended calculate their
costs. not to generate earnings.
profit but to drive
traffic.
6.8 Gross Margin Margin divided Allowances and To quantify
Return on by the average rebates must be return on
Inventory dollar value of considered in working capital
Investment inventory held margin calcula- invested in
(GMROII) during a specific tions. For “loss inventory.
period of time. leaders” this
measure may be
consistently
negative and still
not present a
problem. For
most products,
negative trends
in GMROII are
signs of future
problems.

6.1 Sales Force Coverage: Territories
Sales force territories are the customer groups or geographic districts for which
individual salespeople or sales teams hold responsibility. Territories can be defined
on the basis of geography, sales potential, history, or a combination of factors.
Companies strive to balance their territories because this can reduce costs and
increase sales.
Workload (#) [Current Accounts (#) * Average Time to Service an Active Account (#)]
[Prospects (#) * Time Spent Trying to Convert a Prospect into
an Active Account (#)]
Sales Potential ($) Number of Possible Accounts (#) * Buying Power ($)


Purpose: To create balanced sales territories.
There are a number of ways to analyze territories.1 Most commonly, territories are com-
pared on the basis of their potential or size. This is an important exercise. If territories
differ sharply or slip out of balance, sales personnel may be given too much or too little
work. This can lead to under- or over-servicing of customers.
When sales personnel are stretched too thin, the result can be an under-servicing of
customers. This can cost a firm business because over-taxed salespeople engage in sub-
optimal levels of activity in a number of areas. They seek out too few leads, identify too
few prospects, and spend too little time with current customers. Those customers, in
turn, may take their business to alternate providers.
Over-servicing, by contrast, may raise costs and prices and therefore indirectly reduce
sales. Over-servicing in some territories may also lead to under-servicing in others.
Unbalanced territories also raise the problem of unfair distribution of sales potential
among members of a sales force. This may result in distorted compensation and cause
talented salespeople to leave a company, seeking superior balance and compensation.
Achieving an appropriate balance among territories is an important factor in maintain-
ing satisfaction among customers, salespeople, and the company as a whole.

Construction
In defining or redefining territories, companies strive to

■ Balance workloads
■ Balance sales potential
■ Develop compact territories
■ Minimize disruptions during the redesign
These goals can have different effects on different stakeholders, as represented in
Table 6.1.2
Before designing new territories, a sales force manager should evaluate the workloads
of all members of the sales team. The workload for a territory can be calculated
as follows:
Workload (#) [Current Accounts (#) * Average Time to Service an Active Account (#)]
[Prospects (#) * Time Spent Trying to Convert a Prospect into an
Active Account (#)]
The sales potential in a territory can be determined as follows:
Sales Potential ($) Number of Possible Accounts (#) * Buying Power ($)


Table 6.1 Effects of Balancing Sales Territories

Balance Develop
Balance the Sales Minimize Compact
Workload Potential Disruption Territories
Customers Responsiveness X X
Relationships X
Salespeople Earnings
opportunities X
Manageable
workload X X
Reduced
uncertainty X
Control of
overnights X
Firm Sales results X X X
Effort control X
Motivation X X X X
Travel cost
control X

Buying power is a dollar figure based on such factors as average income levels, number
of businesses in a territory, average sales of those businesses, and population demo-
graphics. Buying power indices are generally specific to individual industries.

EXAMPLE: Among the sales prospects in one of its territories, a copier manufacturer
has identified six small businesses, eight medium-sized firms, and two large companies.
Enterprises of these sizes have historically made annual copier purchases that average
$500, $700, and $1,000, respectively. The sales potential for the territory is thus:
Sales Potential (6 * $500) (8 * $700) (2 * $1,000) $10,600

In addition to workload and sales potential, a third key metric is needed to compare
territories. This is size or, more specifically, travel time. In this context, travel time is
more useful than size because it more accurately represents the factor that size implies—
that is, the amount of time needed to reach customers and potential customers.
As a manager’s goal is to balance workload and potential among sales personnel, it can
be beneficial to calculate combined metrics—such as sales potential or travel time—in
order to make comparisons between territories.


Sales potential can be represented in a number of ways. Of these, the most basic is
population—the number of potential accounts in a territory. In the copier case cited
earlier, this might be the number of offices in a territory.
Estimating the size of a territory might involve simply calculating the geographic area that
it covers. It is likely, however, that average travel time will also be important. Depending
on the quality of roads, density of traffic, or distance between businesses, one may find
that territories of equal area entail very different travel time requirements. In evaluating
such distinctions, sales force records of the time needed to travel from call to call can be
useful. Specialized computer software programs are available for these purposes.
Redefining territories is a famously difficult process. To perform it well, in addition to
the metrics cited earlier, disruption of customer relationships and feelings of ownership
among sales personnel must also be considered.

6.2 Sales Force Objectives: Setting Goals
Sales goals are generally needed to motivate salespeople. These can have negative effects,
however, if set too high or low. Means of establishing sales goals include the following:
Sales Goal ($) Salesperson’s Share of Prior-Year Sales in District (%)
* Forecasted Sales for District ($)
Sales Goal ($) Salesperson’s Prior-Year Sales ($) [Forecasted Sales Increase for
District ($) * Territory’s Share of Sales Potential in District (%)]
Weighted Share of Sales Allotment (%) {Salesperson’s Share of Prior-Year Sales in
District (%) * Assigned Weighting (%)}
{Territory’s Share of Sales Potential in District
(%) * [1 Assigned Weighting (%)]}
Sales Goal ($) Weighted Share of Sales Allotment (%)
* Forecasted Sales for District ($)
Many of these approaches involve a combination of historical results and a weighting
of sales potential among the territories. This ensures that overall goals will be
attained if all salespeople meet their individual goals.

Purpose: To motivate sales personnel and establish benchmarks
for evaluating and rewarding their performance.
In setting sales goals, managers strive to motivate their personnel to stretch themselves
and generate the most sales possible. But they don’t want to set the bar too high. The
correct goal levels will motivate all salespeople and reward most of them.


When planning sales goals, certain guidelines are important. Under the SMART strategy
recommended by Jack D. Wilner, author of Seven Secrets to Successful Sales
Management,3 goals should be Specific, Measurable, Attainable, Realistic, and Time-
bound. Goals should be specific to a department, a territory, and even a salesperson.
They should be clear and applicable to each individual so that salespeople do not have
to derive part of their goal. Measurable goals, expressed in concrete numbers such as
“dollar sales” or “percentage increase,” enable salespeople to set precise targets and
track their progress. Vague goals, such as “more” or “increased” sales, are not effective
because they make it difficult to measure progress. Attainable goals are in the realm of
possibility. They can be visualized and understood by both the manager and the sales-
person. Realistic goals are set high enough to motivate, but not so high that salespeople
give up before they even start. Finally, time-bound goals must be met within a precise
time frame. This applies pressure to reach them sooner rather than later and defines an
endpoint when results will be checked.

Construction
There are numerous ways of allotting a company’s forecast across its sales force. These
methods are designed to set goals that are fair, achievable, and in line with historic
results. Goals are stated in terms of sales totals for individual salespeople. In the follow-
ing formulas, which encapsulate these methods, a district is composed of the individual
territories of multiple salespeople.
A sales goal or allocation based on prior-year sales can be calculated as follows:4
Sales Goal ($) Salesperson’s Share of Prior-Year Sales in District (%) * Forecasted Sales
for District ($)
A sales goal based on prior-year sales and the sales potential of a territory can be calcu-
lated as follows:
Sales Goal ($) Salesperson’s Prior-Year Sales ($) [Forecasted Sales Increase for District($)
* Territory’s Share of Sales Potential in District (%)]
Sales goals can also be set by a combined method, in which management assigns weight-
ings to both the prior-year sales of each salesperson and the sales potential of each ter-
ritory. These weightings are then used to calculate each salesperson’s percentage share of
the relevant sales forecast, and percentage shares are used to calculate sales goals in
dollar terms.
Weighted Share of Sales Allotment (%) {Salesperson’s Share of Prior-Year Sales in District (%)
* Assigned Weighting (%)} {Territory’s Share
of Sales Potential in District (%) * [1 Assigned
Weighting (%)]}
Sales Goal ($) Weighted Share of Sales Allotment (%) * Forecasted Sales for District ($)


EXAMPLE: A salesperson achieved prior-year sales of $1,620, which represented 18% of
the sales in her district. This salesperson was responsible for a territory that held 12% of the
sales potential in the district. If the salesperson’s employer mandates a district sales goal of
$10,000 for the coming year—representing an overall increase of $1,000 over prior-year
results—then the salesperson’s individual sales goal can be calculated in several ways that
involve different emphasis on historical sales versus sales potential. Here are four examples:
1. Sales Goal Based on Prior-year Sales 18% * $10,000 $1,800
2. Sales Goals Based on Sales Potential 12% * $10,000 $1,200
3. Sales Goal Based on Prior-year Sales Sales Potential * Increase $1,620
(12% * $1,000) $1,740
4. Weighted Share of Sales Allotment, in Which Prior-year Sales and Sales Potential
Are Weighted (for Example) by a Factor of 50% Each (18% * 50%) (12% *
50%) 15%. Then…
Sales Goal Based on Weighted Share of Sales Allotment 15% * $10,000 $1,500

Sales goals are generally established by using combinations of bottom-up and top-down
procedures. Frequently, top management sets objectives at a corporate level, while the sales
manager allocates shares of that overall goal among the various members of the sales force.
Top management generally uses multiple metrics to forecast sales, including prior-year
sales of the product in question, total prior-year sales in the relevant market, prior-year
sales by competitors, and the company’s current market share. After the corporate sales
forecast is derived, a sales force manager verifies that these targets are reasonable, push-
ing back where necessary. The manager then allots the projected sales among the sales
force in a district, based at least in part on measures of individual performance from the
prior year. Of greatest importance in this calculation are each salesperson’s historic
percentage of sales and the sales potential of his or her territory.
It is important to re-evaluate sales goals during the year to ensure that actual performance
is running reasonably close to projections. If, at this checkpoint, it appears that more than
90% or less than 50% of the sales force is on track to achieve their goals, then it may be
advisable to alter the goals. This will prevent salespeople from easing off too early because
their goals are in sight, or giving up because their goals are unattainable. In setting goals,
one possible rule of thumb would be to plan for a success rate of 75%. That would ensure
that enough salespeople reach their goal and that the goal is sufficiently challenging.
If “rebudgeting” becomes necessary, it is important to ensure that this is properly
recorded. Unless care is taken, revised sales goals can slip out of alignment with finan-
cial budgets and the expectations of senior management.


6.3 Sales Force Effectiveness: Measuring Effort,
Potential, and Results
By analyzing sales force performance, managers can make changes to optimize sales
going forward. Toward that end, there are many ways to gauge the performance of indi-
vidual salespeople and of the sales force as a whole, in addition to total annual sales.
Sales Force Effectiveness Ratios
Sales ($)
Contacts with Clients (Calls) (#)
Sales ($)
Potential Accounts (#)
Sales ($)
Active Accounts (#)
Sales ($)
Buying Power ($)
Expenses ($)
(Also Known As Cost of Sales)
Sales ($)

Each can also be calculated on a dollar contribution basis.

Purpose: To measure the performance of a sales force
and of individual salespeople.
When analyzing the performance of a salesperson, a number of metrics can be com-
pared. These can reveal more about the salesperson than can be gauged by his or her
total sales.

Construction
An authoritative source lists the following ratios as useful in assessing the relative effec-
tiveness of sales personnel:5
Sales ($)
Contacts with Clients (Calls) (#)
Sales ($)
Potential Accounts (#)


Sales ($)
Active Accounts (#)
Sales ($)
Buying Power ($)

These formulas can be useful for comparing salespeople from different territories and
for examining trends over time. They can reveal distinctions that can be obscured by
total sales results, particularly in districts where territories vary in size, in number of
potential accounts, or in buying power.
These ratios provide insight into the factors behind sales performance. If an individual’s
sales per call ratio is low, for example, that may indicate that the salesperson in question
needs training in moving customers toward larger purchases. Or it may indicate a lack
of closing skills. If the sales per potential account or sales per buying power metric is
low, the salesperson may not be doing enough to seek out new accounts. These metrics
reveal much about prospecting and lead generation because they’re based on each sales-
person’s entire territory, including potential as well as current customers. The sales per
active account metric provides a useful indicator of a salesperson’s effectiveness in max-
imizing the value of existing customers.
Although it is important to make the most of every call, a salesperson will not reach his
or her goal in just one call. A certain amount of effort is required to complete sales. This
can be represented graphically (see Figure 6.1).6
Although one can increase sales by expending more time and attention on a customer,
at a certain point, a salesperson encounters diminishing returns in placing more calls to

Sales ($)/Potential Account (#)

Calls (#)/Potential Account (#)

Figure 6.1 Sales Resulting from Calls to Customers


the same customers. Eventually, the incremental business generated by each call will be
worth less than the cost of making the call.
In addition to the formulas described earlier, one other important measure of effective-
ness is the ratio of expenses to sales. This cost metric is commonly expressed as a per-
centage of sales and is calculated as follows:
Expenses ($)
Sales ($)
If this ratio is substantially higher for one salesperson than for others, it may indicate
that the individual in question has poor control of his or her expenses. Examples of
poor expense control could include making unnecessary trips to a client, overproducing
product pamphlets, or hosting too many dinners. Alternatively, expenses may represent
a high percentage of sales if an individual possesses poor closing skills. If a salesperson’s
expenses are comparable to those of his peers, but his sales are lower, then he may be
failing to deliver sales after spending significant money on a potential customer.
A more challenging set of sales force performance metrics involves customer service.
Customer service is difficult to measure because there are no concrete numbers repre-
senting it, other than repeat rates or customer complaints. Each of those is telling, but
how can a sales manager evaluate the service provided to customers who are not repeat-
ing, leaving, or complaining? One possibility is to develop a survey, including an
itemized scale to help customers quantify their opinions. After enough of these surveys
are completed, managers will be able to calculate average scores for different service
metrics. By comparing these with sales figures, managers can correlate sales with
customer service and grade salespeople on their performance.

EXAMPLE: To translate customers’ opinions into a metric, a company might pose
survey questions such as the following:
Please circle the level of service your business received from our sales staff after shipment
of the products you ordered:
1 2 3 4 5 6 7 8 9 10
Extremely Poor Satisfactory Extremely Good

Calculating the effectiveness of a salesperson is not difficult, but it does require keeping
track of a few important numbers. Fortunately, these are commonly recorded in the
sales industry.


The most important statistics are the amount of each sale (in dollars) and the contribu-
tion generated by that sale. It may also be important to keep track of which items are
sold if a salesperson has been instructed to emphasize a certain product line. Additional
useful information would include measures of the number of calls made (including
both face-to-face and phone meetings), total accounts active, and total accounts in the
territory. Of these, the latter two are needed to calculate the buying power of a territory.
The largest problem in performance review is a tendency to rely on only one or two
metrics. This can be dangerous because an individual’s performance on any one meas-
ure may be anomalous. A salesperson who generates $30,000 per call may be more valu-
able than one who generates $50,000 per call, for example, if he generates greater sales
per potential account. A salesperson in a small territory may generate low total contri-
bution but high dollar sales per buying power. If this is true, it may be advisable to
increase the size of that person’s territory. Another salesperson may show a dramatic
increase in dollar sales per active account. If he achieves this simply by eliminating
weaker accounts without generating incremental sales, it would not be grounds for
reward. In reviewing sales personnel, managers are advised to evaluate as many per-
formance metrics as possible.
Although the customer service survey described earlier is grounded upon a straightfor-
ward concept, managers can find it difficult to gather enough data—or sufficiently repre-
sentative data—to make it useful. This could be because customers hesitate to fill out the
surveys, or because they do so only when they encounter a problem. A small sample size or
a prevalence of negative responses might distort the results. Even so, some effort to meas-
ure customer satisfaction is needed to ensure that salespeople don’t emphasize the wrong
issues—or neglect issues that have a substantial impact on customers’ lifetime value.

6.4 Sales Force Compensation: Salary/Reward Mix
“The incentive plan needs to align the salesperson’s activities with the firm’s objec-
tives.”7 Toward that end, an effective plan may be based on the past (growth), the
present (comparison with others), or the future (percentage of goal achieved). Key
formulas in this area include the following:
Compensation ($) Salary ($) Bonus 1 ($) Bonus 2 ($)
Compensation ($) Salary ($) [Sales ($) * Commission (%)]
(Sales ($) * [Margin (%) Commission (%)])
Break-Even Number of Employees (#)
[Salary ($) Expenses ($) Bonus ($)]


Purpose: To determine the mix of salary, bonus, and commission that
will maximize sales generated by the sales force.
When designing a compensation plan for a sales force, managers face four key consider-
ations: level of pay, mix between salary and incentive, measures of performance, and
performance-payout relationships. The level of pay, or compensation, is the amount
that a company plans to pay a salesperson over the course of a year. This can be viewed
as a range because its total will vary with bonuses or commissions.
The mix between salary and incentive represents a key allocation within total compen-
sation. Salary is a guaranteed sum of money. Incentives can take multiple forms,
including bonuses or commissions. In the case of a bonus, a salesperson will receive a
lump sum for reaching certain sales targets. With a commission, the incentive is incre-
mental and is earned on each sale. In order to generate incentives, it is important to
measure accurately the role a salesperson plays in each sale. The higher the level of
causality that can be attributed to a salesperson, the easier it is to use an incentive
system.
Various metrics can be used to measure a salesperson’s performance. With these, man-
agers can evaluate a salesperson’s performance in the context of past, present, or future
comparators, as follows:
■ The past: Measure the salesperson’s percentage growth in sales over prior-year
results.
■ The present: Rank salespeople on the basis of current results.
■ The future: Measure the percentage of individual sales goals achieved by each
salesperson.
Sales managers can also select the organizational level on which to focus an incentive
plan. The disbursement of incentive rewards can be linked to results at the company,
division, or product-line level. In measuring performance and designing compensation
plans along all these dimensions, managers seek to align salespeople’s incentives with
the goals of their firm.
Lastly, a time period should be defined for measuring the performance of
each salesperson.

Construction
Managers enjoy considerable freedom in designing compensation systems. The key is to
start with a forecast for sales and a range within which each salesperson’s compensation
should reside. After these elements are determined, there are many ways to motivate
a salesperson.


In a multi-bonus system, the following formula can represent the compensation struc-
ture for a salesperson:
Compensation ($) Salary ($) Bonus 1 ($) Bonus 2 ($)
In this system, bonus 1 might be attained at a level approximately halfway to the individ-
ual’s sales goal for the year. The second bonus might be awarded when that goal is met.
In a commission system, the following formula would represent compensation for
a salesperson:
Compensation ($) Salary ($) [Sales ($) * Commission (%)]
Theoretically, in a 100% commission structure, salary might be set as low as $0. Many
jurisdictions, however, place limits on such arrangements. Managers must ensure that
their chosen compensation structures comply with employment law.
Managers can also combine bonus and commission structures by awarding bonuses on
top of commissions at certain sales levels, or by increasing the commission rate at cer-
tain sales levels.

EXAMPLE: Tina earns a commission of 2% on sales up to $1,000,000, and a 3% com-
mission on sales beyond that point. Her salary is $20,000 per year. If she makes
$1,200,000 in sales, her compensation can be calculated as follows:

Compensation $20,000 (.02) * ($1,000,000) (.03) * ($200,000)
$46,000

After a sales compensation plan has been established, management may want to re-
evaluate the size of its sales force. Based on forecasts for the coming year, a firm may
have room to hire more salespeople, or it may need to reduce the size of the sales force.
On the basis of a given value for projected sales, managers can determine the break-even
number of employees for a firm as follows:
Sales ($) * [Margin (%) Commission (%)]
Break-Even Number of Employees (#)
[Salary ($) Expenses ($) Bonus ($)]

Measurements commonly used in incentive plans include total sales, total contribution,
market share, customer retention, and customer complaints. Because such a plan
rewards a salesperson for reaching certain goals, these targets must be defined at the


beginning of the year (or other time period). Continual tracking of these metrics will
help both the salesperson and the company to plan for year-end compensation.
Timing is an important issue in incentive plans. A firm must collect data in a timely
fashion so that both managers and salespeople know where they stand in relation to
established goals. The time frame covered by a plan also represents an important con-
sideration. If a company tries to generate incentives through weekly rewards, its com-
pensation program can become too expensive and time-consuming to maintain. By
contrast, if the program covers too long a period, it may slip out of alignment with com-
pany forecasts and goals. This could result in a sales force being paid too much or too
little. To guard against these pitfalls, managers can develop a program that mixes both
short- and long-term incentives. They can link some rewards to a simple, short-term
metric, such as calls per week, and others to a more complex, long-term target, such as
market share achieved in a year.
A further complication that can arise in incentive programs is the assignment of causal-
ity to individual salespeople. This can become a problem in a number of instances,
including team collaborations in landing sales. In such a scenario, it can be difficult to
determine which team members deserve which rewards. Consequently, managers may
find it best to reward all members of the team with equal bonuses for meeting a goal.
A last concern: When an incentive program is implemented, it may reward the “wrong”
salespeople. To avoid this, before activating any newly proposed program, sales man-
agers are advised to apply that program to the prior year’s results as a test. A “good” plan
will usually reward the salespeople whom the manager knows to be the best.

6.5 Sales Force Tracking: Pipeline Analysis
Pipeline analysis is used to track the progress of sales efforts in relation to all current
and potential customers in order to forecast short-term sales and to evaluate sales
force workload.

Purpose: To forecast upcoming sales and evaluate workload distribution.
A convenient way to forecast sales in the short term and to keep an eye on sales force
activity is to create a sales pipeline or sales funnel. Although this concept can be
represented graphically, the data behind it are stored electronically in a database or
spreadsheet.
The concept of the sales funnel originates in a well-known dynamic: If a sales force
approaches a large number of potential customers, only a subset of these will actually


make purchases. As salespeople proceed through multiple stages of customer interac-
tion, a number of prospects are winnowed out. At the conclusion of each stage, fewer
potential customers remain. By keeping track of the number of potential customers at
each stage of the process, a sales force manager can balance the workload within a team
and make accurate forecasts of sales.
This analysis is similar to the hierarchy of effects discussed in Section 2.7. Whereas the
hierarchy of effects focuses on the impact of advertising or mass media, the sales funnel is
used to track individual customers (often by name) and sales force efforts. (Note: In some
industries, such as consumer packaged goods, the term “pipeline sales” can refer to sales
into a distribution channel. Please do not confuse pipeline sales with a sales pipeline.)

Construction
In order to conceptualize a sales funnel or pipeline, it is helpful to draw a diagram show-
ing the stages of the selling process (see Figure 6.2). At any point in the year, it is likely
that all stages of the pipeline will include some number of customers. As Figure 6.2
illustrates, although there may be a large number of potential customers, those who
actually make purchases represent only a percentage of these original leads.

Cold Leads Interest

Creation
Warm Leads

Prospects
Pre-purchase

1st Meeting

2nd Meeting Purchase

3rd Meeting

Post-purchase
Delivery

Support

Figure 6.2 Sales Force Funnel


Interest Creation: This entails building awareness of a product through such activities
as trade shows, direct mail, and advertising. In the course of interest creation, salespeo-
ple can also generate leads. That is, they can identify targets to add to their pool of
potential customers. Two main classifications of leads include cold leads and warm
leads.

Cold Lead: A lead that has not specifically expressed interest. These can be identi-
fied through mailing lists, phone books, business listings, and so on.
Warm Lead: A lead that is expected to be responsive. These potential customers
may have registered through a Web site or requested product information, for
example.
Pre-Purchase: This stage involves identifying prospects from among cold and warm
leads. Salespeople make this distinction through initial meetings with leads, in which
they explain product features and benefits, and cooperate in problem solving with the
customer. The desired result of such an early-stage meeting is not a sale but rather the
identification of a prospect and the scheduling of another meeting.
Prospect: A potential customer who has been identified as a likely buyer, possess-
ing the ability and willingness to buy.8
Purchase: After prospects are identified and agree to additional calls, salespeople engage
in second and third meetings with them. It is in these sessions that traditional “selling”
takes place. Salespeople will engage in persuading, negotiating, and/or bidding. If a
purchase is agreed upon, a salesperson can close the deal through a written proposal,
contract, or order.
Post-Purchase: After a customer has made a purchase, there is still considerable work
to be done. This includes delivery of the product or service, installation (if necessary),
collection of payments, and possibly training. There is then an ongoing commitment to
customer service.
After salespeople visualize the different stages represented in a sales funnel, they can
track their customers and accounts more accurately. They can do this electronically by
using a database or spreadsheet. If a sales pipeline file is maintained on a shared drive,
any member of a sales force will be able to update the relevant data on a regular basis.
This will also enable a sales manager to view the progress of the team at any point in
time. Table 6.2 is an example of a spreadsheet form of a sales funnel.
A manager can use the information stored in such a funnel to prepare for sales in the
near future. This is a form of pipeline analysis. When a firm faces inventory issues, or
when sales goals are being missed, this represents vital information. By applying histor-
ical averages, a sales or marketing manager can improve sales forecasts by using the data
in a sales funnel. This can be done manually or with specialized software. The underly-


Table 6.2 Spreadsheet Sales Funnel

Interest Creation Pre-purchase Purchase Post-purchase
Cold Warm 1st/2nd 2nd/3rd
Salesperson Leads Leads Prospects Meeting Meeting Delivery Support
Sandy 56 30 19 5 8 7 25
Bob 79 51 33 16 4 14 35

ing assumption behind a sales funnel is that failure at any stage eliminates a prospect
from the funnel. The following example illustrates how this bottom-up forecasting
could be applied.

EXAMPLE: Using the sales funnel from earlier, Sandy and Bob’s manager wants to
forecast the number of sales that will require fulfillment in the next five months. Toward
that end, she applies certain historical averages:
■ 2% of cold calls are converted to sales within five months.
■ 14% of warm calls are converted to sales within four months.
■ 25% of prospects are converted to sales within three months.
■ 36% of customers who agree to a pre-purchase meeting are converted to sales
within two months.
■ 53% of customers who agree to a purchase meeting are converted to sales
within one month.
On this basis:
Upcoming Sales [(56 79) * 2%] [(30 51) * 14%] [(19 33) * 25%]
[(5 16) * 36%)] [(8 4) * 53%] 41
Note: This example applies to only one product. Often, a firm will need multiple sales
funnels for different products or product lines. Additionally, a sale may comprise a single
item or thousands of items. In the latter case, it would be appropriate to use a metric for
“average sale size/customer” in forecasting.

In order to populate a sales funnel correctly, salespeople must maintain records of all
their current and potential customers, and the status of each within the purchase
process. Each salesperson must also share this information, which can then be aggregated


in a comprehensive database of sales force activities. By applying assumptions to
these—including assumptions drawn from historical sales results—a firm can project
future sales. For example, if 25% of warm leads are generally converted to sales within
two months, and 200 warm leads currently appear in a sales funnel, management can
estimate that 50 of these will be converted to sales within two months.
At times, the use of a sales funnel leads to the pitfall of over-prospecting. If the incre-
mental contribution generated by a customer is less than the cost of acquiring that cus-
tomer, then prospecting for that customer yields a negative result. Salespeople are
advised to use customer lifetime value metrics as a guide in deciding the appropriate
scale and direction of their prospecting. Increasing pre-purchase sales funnel metrics
will not be worthwhile unless that increment leads to improved figures further down the
pipeline as well.
Difficulties in the sales cycle can also arise when a salesperson judges that a potential
customer may be a prospect because he or she has the willingness and ability to buy.
To solidify this judgment, the salesperson must also confirm that the customer possesses
the authority to buy. When prospecting, salespeople should take the time needed to ver-
ify that their contacts can make purchase decisions without approval from another
source.

6.6 Numeric, ACV and PCV Distribution,
Facings/Share of Shelf
Distribution metrics quantify the availability of products sold through resellers, usu-
ally as a percentage of all potential outlets. Often, outlets are weighted by their share
of category sales or “all commodity” sales.
Number of Outlets Carrying Brand (#)
Numeric Distribution (%)
Total Number of Outlets (#)
Total Sales of Outlets Carrying Brand ($)
All Commodity Volume (ACV) Distribution (%)
Total Sales of All Outlets ($)
Total Category Sales of Outlets
Carrying Brand ($)
Product Category Volume (PCV) Distribution9 (%)
Total Category Sales of All
Outlets ($)
PCV (%)
Category Performance Ratio (%)
ACV (%)


For marketers who sell through resellers, distribution metrics reveal a brand’s per-
centage of market access. Balancing a firm’s efforts in “push” (building and maintain-
ing reseller and distributor support) and “pull” (generating customer demand) is an
ongoing strategic concern for marketers.

Purpose: To measure a firm’s ability to convey a product to its customers.
In broad terms, marketing can be divided into two key challenges:
■ The first—and most widely appreciated—is to ensure that consumers or end
users want a firm’s product. This is generally termed pull marketing.
■ The second challenge is less broadly recognized, but often just as important.
Push marketing ensures that customers are given opportunities to buy.
Marketers have developed numerous metrics by which to judge the effectiveness of the
distribution system that helps create opportunities to buy. The most fundamental of
these are measures of product availability.
Availability metrics are used to quantify the number of outlets reached by a product, the
fraction of the relevant market served by those outlets, and the percentage of total sales
volume in all categories held by the outlets that carry the product.

Construction
There are three popular measures of distribution coverage:

1. Numeric distribution
2. All commodity volume (ACV)
3. Product category volume (PCV), also known as weighted distribution

NUMERIC DISTRIBUTION
This measure is based on the number of outlets that carry a product (that is, outlets that
list at least one of the product’s stock-keeping units, or SKUs). It is defined as the per-
centage of stores that stock a given brand or SKU, within the universe of stores in the
relevant market.
The main use of numeric distribution is to understand how many physical locations
stock a product or brand. This has implications for delivery systems and for the cost of
servicing these outlets.


Numeric Distribution: To calculate numeric distribution, marketers divide the number
of stores that stock at least one SKU of a product or brand by the number of outlets in
the relevant market.

Number of Outlets Carrying Product (#)
Numeric Distribution (%)
Total Number of Outlets in the Market (#)

For further information about stock-keeping units (SKUs), refer to Section 3.3.

EXAMPLE: Alice sells photo albums to gift shops. There are 60 such stores in her area.
In order to generate adequate distribution coverage, Alice believes she must reach at least
60% of these. In initiating her relationship with each store, however, Alice must provide
the store with $4,000 worth of inventory to build a presence. To attain her distribution
goal, how much will Alice need to invest in inventory?
To reach her numeric distribution target of 60%, Alice must build a presence in 36 stores
(that is, 0.60 * 60).
She will therefore have to spend at least $144,000 on inventory (36 stores * $4,000 per
store).

ALL COMMODITY VOLUME
All commodity volume (ACV) is a weighted measure of product availability, or dis-
tribution, based on total store sales. ACV can be expressed as a dollar value or
percentage.
All Commodity Volume (ACV): The percentage of sales in all categories that are
generated by the stores that stock a given brand (again, at least one SKU of that
brand).
All Commodity Volume (ACV Distribution) (%)
Total Sales of Stores Carrying Brand ($)
Total Sales of All Stores ($)
All Commodity Volume (ACV Distribution) ($) Total Sales of Stores Carrying Brand ($)

EXAMPLE: The marketers at Madre’s Tortillas want to know the all commodity vol-
ume of their distribution network (Table 6.3).


Table 6.3 Madre’s Tortillas’ Distribution

Madre’s Tortillas Padre’s Tortillas
Outlet All Sales Tortilla Sales SKUs Stocked SKUs Stocked
Store 1 $100,000 $1,000 12 ct, 24 ct 12 ct, 24 ct

Store 2 $75,000 $500 12 ct 24 ct

Store 3 $50,000 $300 12 ct, 24 ct none

Store 4 $40,000 $400 none 12 ct, 24 ct

Madre’s Tortillas are carried by Stores 1-3, but not by Store 4. The ACV of its distribution
network is therefore the total sales of Stores 1, 2, and 3, divided by the total sales of all
stores. This represents a measure of the sales of all commodities in these stores, not just
tortilla sales.

Sales Stores 1 3
Madre’s Tortillas ACV (%) =
All Store Sales
($100k $75k $50k)
=
($100k $75k $50k $40k)
$225k
= 84.9%
$265k

The principal benefit of the ACV metric, by comparison with numeric distribution, is
that it provides a superior measure of customer traffic in the stores that stock a brand.
In essence, ACV adjusts numeric distribution for the fact that not all retailers generate
the same level of sales. For example, in a market composed of two small stores, one
superstore, and one kiosk, numeric distribution would weight each outlet equally,
whereas ACV would place greater emphasis on the value of gaining distribution in
the superstore. In calculating ACV when detailed sales data are not available,
marketers sometimes use the square footage of stores as an approximation of their total
sales volume.
The weakness of ACV is that it does not provide direct information about how
well each store merchandises and competes in the relevant product category. A store
can do a great deal of general business but sell very little of the product category under
consideration.


PRODUCT CATEGORY VOLUME
Product category volume (PCV)10 is a refinement of ACV. It examines the share of the
relevant product category sold by the stores in which a given product has gained distri-
bution. It helps marketers understand whether a given product is gaining distribution in
outlets where customers look for its category, as opposed to simply high-traffic stores
where that product may get lost in the aisles.
Continuing our example of the two small retailers, the kiosk, and the superstore,
although ACV may lead the marketer of a chocolate bar to seek distribution in the high-
traffic superstore, PCV might reveal that the kiosk, surprisingly, generates the greatest
volume in snack sales. In building distribution, the marketer would then be advised to
target the kiosk as her highest priority.
Product Category Volume (PCV): The percentage share, or dollar value, of cate-
gory sales made by stores that stock at least one SKU of the brand in question, in
comparison with all stores in their universe.
Total Category Sales by Stores
Carrying Brand ($)
Product Category Volume (PCV Distribution) (%)
Total Category Sales of All Stores ($)

Product Category Volume (PCV Distribution) ($) Total Category Sales of Stores
Carrying Brand ($)
When detailed sales data are available, PCV can provide a strong indication of the
market share within a category to which a given brand has access. If sales data are not
available, marketers can calculate an approximate PCV by using square footage devoted
to the relevant category as an indication of the importance of that category to a partic-
ular outlet or store type.

EXAMPLE: The marketers at Madre’s Tortillas want to know how effectively their
product is reaching the outlets where customers shop for tortillas. Using data from the
previous example:
Stores 1, 2, and 3 stock Madre’s Tortillas. Store 4 does not. The product category volume
of Madre’s Tortillas’ distribution network can be calculated by dividing total tortilla sales
in Stores 1-3 by tortilla sales throughout the market.
(Tortilla Sales of Stores Carrying Madre’s)
PCV (%)
(Tortilla Sales of All Stores)
($1,000 $500 $300)
$81.8%
($1,000 $500 $300 $400)


Total Distribution: The sum of ACV or PCV distribution for all of a brand’s
stock-keeping units, calculated individually. By contrast with simple ACV or PCV,
which are based on the all commodity or product-category sales of all stores that
carry at least one SKU of a brand, total distribution also reflects the number of
SKUs of the brand that is carried by those stores.
Category Performance Ratio: The relative performance of a retailer in a given
product category, compared with its performance in all product categories.
By comparing PCV with ACV, the category performance ratio provides insight into
whether a brand’s distribution network is more or less effective in selling the category of
which that brand is a part, compared with its average effectiveness in selling all cate-
gories in which members of that network compete.
PCV (%)
Category Performance Ratio (%)
ACV (%)

If a distribution network’s category performance ratio is greater than 1, then the outlets
comprising that network perform comparatively better in selling the category in ques-
tion than in selling other categories, relative to the market as a whole.

EXAMPLE: As noted earlier, the PCV of Madre’s Tortillas’ distribution network is
81.8%. Its ACV is 84.9%. Thus, its category performance ratio is 0.96.
Madre’s has succeeded in gaining distribution in the largest stores in its market. Tortilla
sales in those stores, however, run slightly below the average of all commodity sales in
those stores, relative to the market as a whole. That is, outlets carrying Madre’s show a
slightly weaker focus on tortillas than the overall universe of stores in this market.

In many markets, there are data suppliers such as A.C. Nielsen, which specialize in col-
lecting information about distribution. In other markets, firms must generate their own
data. Sales force reports and shipment invoices provide a place to start.
For certain merchandise—especially low-volume, high-value items—it is relatively
simple to count the limited number of outlets that carry a given product. For higher-
volume, lower-cost goods, merely determining the number of outlets that stock an
item can be a challenge and may require assumptions. Take, for instance, the num-
ber of outlets selling a specific soft drink. To arrive at an accurate number, one
would have to include vending machines and street vendors as well as traditional
grocery stores.


Total outlet sales are often approximated by quantifying selling space (measured in
square feet or square meters) and applying this measure to industry averages for sales
per area of selling space.
In the absence of specific category sales data, it is often useful to weight ACV to arrive at
an approximation of PCV. Marketers may know, for example, that pharmacies, relative
to their overall sales, sell proportionally more of a given product than do superstores. In
this event, they might increase the weighting of pharmacies relative to superstores in
evaluating relevant distribution coverage.

Facing: A facing is a frontal view of a single package of a product on a fully stocked
shelf.
Share of Shelf: A metric that compares the facings of a given brand to the total facing
positions available, in order to quantify the display prominence of that brand.
Facings for Brand (#)
Share of Shelf (%)
Total Facings (#)

Store Versus Brand Measures: Marketers often refer to a grocery chain’s ACV. This can
be either a dollar number (the chain’s total sales of all categories in the relevant geo-
graphic market) or a percentage number (its share of dollar sales among the universe of
stores). A brand’s ACV is simply the sum of the ACVs of the chains and stores that stock
that brand. Thus, if a brand is stocked by two chains in a market, and these chains have
40% and 30% ACV respectively, then the ACV of that brand’s distribution network is
30% 40%, or 70%.
Marketers can also refer to a chain’s market share in a specific category. This is equiva-
lent to the chain’s PCV (%). A brand’s PCV, by contrast, represents the sum of the PCVs
of the chains that stock that brand.
Inventory: This is the level of physical stock held. It will typically be measured at differ-
ent points in a pipeline. A retailer may have inventory on order from suppliers, at ware-
houses, in transit to stores, in the stores’ backrooms, and on the store shelves.
Breadth of Distribution: This figure can be measured by the number of SKUs held.
Typically, a company will hold a wide range of SKUs—a high breadth of distribution—
for the products that it is most interested in selling.
Features in Store: The percentage of stores offering a promotion in a given time period.
This can be weighted by product or by all commodity volume (ACV).


ACV on Display: Distinctions can be made in all commodity volume metrics to take
account of where products are on display. This will reduce the measured distribution of
products if they are not in a position to be sold.
AVC on Promotion: Marketers may want to measure the ACV of outlets where a given
product is on promotion. This is a useful shorthand way of determining the product’s
reliance on promotion.

6.7 Supply Chain Metrics
Marketing logistics tracking includes the following metrics:
Outlets Where Brand or Product Is Listed But Unavailable (#)
Out-of-Stocks (%)
Total Outlets Where Brand or Product Is Listed (#)

Deliveries Achieved in Timeframe
Promised (#)
Service Levels; Percentage on Time Delivery (%)
All Deliveries Initiated in the
Period (#)
Product Revenues ($)
Inventory Turns (I)
Average Inventory ($)

Logistics tracking helps ensure that companies are meeting demand efficiently and
effectively.

Purpose: To monitor the effectiveness of an organization in managing
the distribution and logistics process.
Logistics are where the marketing rubber meets the road. A lot can be lost at the poten-
tial point-of-purchase if the right goods are not delivered to the appropriate outlets on
time and in amounts that correspond to consumer demand. How hard can that be?
Well, ensuring that supply meets demand becomes more difficult when:
■ The company sells more than a few stock keeping units (SKUs).
■ Multiple levels of suppliers, warehouses, and stores are involved in the distribu-
tion process.
■ Product models change frequently.
■ The channel offers customer-friendly return policies.


In this complex field, by monitoring core metrics and comparing these with historical
norms and guidelines, marketers can determine how well their distribution channel is
functioning as a supply chain for their customers.
By monitoring logistics, managers can investigate questions such as the following: Did
we lose sales because the wrong items were shipped to a store that was running a pro-
motion? Are we being forced to pay for the disposal of obsolete goods that stayed too
long in warehouses or stores?

Construction
Out-of-Stocks: This metric quantifies the number of retail outlets where an item
is expected to be available for customers, but is not. It is typically expressed as a
percentage of stores that list the relevant item.

Outlets Where Brand or Product Is Listed But Unavailable (#)
Out-of-Stocks (%)
Total Outlets Where Brand or Product Is Listed (#)

Being “listed” by a chain means that a headquarters buyer has “authorized” distribution
of a brand, SKU, or product at the store level. For various reasons, being listed does not
always ensure presence on the shelf. Local managers may not approve “distribution.”
Alternatively, a product may be distributed but sold out.
Out-of-stocks are often expressed as a percentage. Marketers must note whether an out-
of-stock percentage is based on numeric distribution, ACV, PCV, or the percentage of
distributing stores for a given chain.
The in-stock percentage is the complement of the out-of-stock percentage. A 3% out-
of-stock rate would be equivalent to a 97% in-stock rate.
PCV Net Out-of-Stocks: The PCV of a given product’s distribution network,
adjusted for out-of-stock situations.
Product Category Volume (PCV), Net Out-of-Stocks: This out-of-stocks measure is
calculated by multiplying PCV by a factor that adjusts it to recognize out-of-stock situ-
ations. The adjusting factor is simply one minus the out-of-stocks figure.
Product Category Volume, Net Out-of-Stocks (%) PCV (%) * [1 Out-of-Stock (%)]
Service Levels, Percentage On-time Delivery: There are various service measures in
marketing logistics. One particularly common measure is on-time delivery. This metric
captures the percentage of customer (or trade) orders that are delivered in accordance
with the promised schedule.
Deliveries Achieved in Timeframe
Promised (#)
Service Levels, Percentage on Time Delivery (%)
All Deliveries Initiated in the Period (#)


Inventories, like out-of-stocks and service levels, should be tracked at the SKU level.
For example, in monitoring inventory, an apparel retailer will need to know not only
the brand and design of goods carried, but also their size. Simply knowing that there are
30 pairs of suede hiking boots in a store, for example, is not sufficient—particularly if
all those boots are the same size and fail to fit most customers.
By tracking inventory, marketers can determine the percentage of goods at each stage of
the logistical process—in the warehouse, in transit to stores, or on the retail floor, for
example. The significance of this information will depend on a firm’s resource manage-
ment strategy. Some firms seek to hold the bulk of their inventory at the warehouse
level, for example, particularly if they have an effective transport system to ship goods
quickly to stores.
Inventory Turns: The number of times that inventory “turns over” in a year can be cal-
culated on the basis of the revenues associated with a product and the level of inventory
held. One need only divide the revenues associated with the product in question by the
average level of inventory for that item. As this quotient rises, it indicates that inventory
of the item is moving more quickly through the process. Inventory turns can be calcu-
lated for companies, brands, or SKUs and at any level in the distribution chain, but they
are frequently most relevant for individual trade customers. Important note: In calculat-
ing inventory turns, dollar figures for both sales and inventory must be stated either
on a cost or wholesale basis, or on a retail or resale basis, but the two bases must not
be mixed.

Annual Product Revenues ($)
Inventory Turns (I)
Average Inventory ($)

Inventory Days: This metric also sheds light on the speed with which inventory moves
through the sales process. To calculate it, marketers divide the 365 days of the year by
the number of inventory turns, yielding the average number of days of inventory car-
ried by a firm. By way of example, if a firm’s inventory of a product “turned” 36.5 times
in a year, that firm would, on average, hold 10 days’ worth of inventory of the product.
High inventory turns—and, by corollary, low inventory days—tend to increase prof-
itability through efficient use of a firm’s investment in inventory. But they can also lead
to higher out-of-stocks and lost sales.

Days in Year (365)
Inventory Days (#)
Inventory Turns (I)

Inventory days represents the number of days’ worth of sales that can be supplied by
the inventory present at a given moment. Viewed from a slightly different perspective,
this figure advises logistics managers of the time expected to elapse before they suffer a
stock-out. To calculate this figure, managers divide product revenue for the year by the


value of the inventory days, generating expected annual turns for that inventory level.
This can be easily converted into days by using the previous equation.

EXAMPLE: An apparel retailer holds $600,000 worth of socks in inventory January 1,
and $800,000 the following December 31. Revenues generated by sock sales totaled $3.5
million during the year.
To estimate average sock inventory during the year, managers might take the average of
the beginning and ending numbers: ($600,000 $800,000)/2 $700,000 average inven-
tory. On this basis, managers might calculate inventory turns as follows:
Product Revenues
Inventory Turns
Average Inventory
$3,500,000
5
$700,000

If inventory turns five times per year, this figure can be converted to inventory days in
order to measure the average number of days worth of stock held during the period.

Days in Year (365)
Inventory Days
Inventory Turns
365
= 73 Days Worth of Inventory
5

Although some companies and supply chains maintain sophisticated inventory tracking
systems, others must estimate logistical metrics on the basis of less-than-perfect data.
Increasingly, manufacturers may also have difficulty purchasing research because retail-
ers that gather such information tend to restrict access or charge high fees for it. Often,
the only readily available data may be drawn from incomplete store audits or reports filed
by an overloaded sales force. Ideally, marketers would like to have reliable metrics for
the following:
■ Inventory units and monetary value of each SKU at each level of the distribu-
tion chain for each major customer.
■ Out-of-stocks for each SKU, measured at both the supplier and the store
level.


■ Percentage of customer orders that were delivered on time and in the correct
amount.
■ Inventory counts in the tracking system that don’t match the number in the
physical inventory. (This would facilitate a measure of shrinkage or theft.)
When considering the monetary value of inventory, it is important to use comparable
figures in all calculations. As an example of the inconsistency and confusion that can
arise in this area, a company might value its stock on the retail shelf at the cost to the
store, which might include an approximation of all direct costs. Or it might value that
stock for some purposes at the retail price. Such figures can be difficult to reconcile with
the cost of goods purchased at the warehouse and can also be different from accounting
figures adjusted for obsolescence.
When evaluating inventory, managers must also establish a costing system for items that
can’t be tracked on an individual basis. Such systems include the following:
■ First In, First Out (FIFO): The first unit of inventory received is the first
expensed upon sale.
■ Last In, First Out (LIFO): The last unit of inventory received is the first
expensed upon sale.
The choice of FIFO or LIFO can have a significant financial impact in inflationary
times. At such times, FIFO will hold down the cost of goods sold by reporting this fig-
ure at the earliest available prices. Simultaneously, it will value inventory at its highest
possible level—that is, at the most recent prices. The financial impact of LIFO will be
the reverse.
In some industries, inventory management is a core skill. Examples include the apparel
industry, in which retailers must ensure that they are not left with prior seasons’ fash-
ions, and the technology industry, in which rapid developments make products hard to
sell after only a few months.
In logistical management, firms must beware of creating reward structures that lead to
sub-optimal outcomes. An inventory manager rewarded solely for minimizing out-of-
stocks, for example, would have a clear incentive to overbuy—regardless of inventory
holding costs. In this field, managers must ensure that incentive systems are sophisti-
cated enough not to reward undesirable behavior.
Firms must also be realistic about what will be achieved in inventory management. In
most organizations, the only way to be completely in stock on every product all the time
is to ramp up inventories. This will involve huge warehousing costs. It will tie up a great
deal of the company’s capital in buying stocks. And it will result in painful obsolescence
charges to unload over-purchased items. Good logistics and inventory management
entails finding the right trade-off between two conflicting objectives: minimizing both
inventory holding costs and sales lost due to out-of-stocks.


Rain Checks, or Make-Goods on Promotions: These measures evaluate the effect on a
store of promotional items being unavailable. In a typical example, a store might track
the incidents in which it offers customers a substitute item because it has run out of
stock on a promoted item. Rain checks or make-goods might be expressed as a percent-
age of goods sold, or more specifically, as a percentage of revenues coded to the promo-
tion but generated by sales of items not listed as part of the promotional event.
Misshipments: This measures the number of shipments that failed arrive on time or in
the proper quantities.
Deductions: This measures the value of deductions from customer invoices caused by
incorrect or incomplete shipments, damaged goods, returns, or other factors. It is often
useful to distinguish between the reasons for deductions.
Obsolescence: This is a vital metric for many retailers, especially those involved in fash-
ion and technology. It is typically expressed as the monetary value of items that are
obsolete, or as the percentage of total stock value that comprises obsolete items. If obso-
lescence is high, then a firm holds a significant amount of inventory that is likely to sell
only at a considerable discount.
Shrinkage: This is generally a euphemism for theft. It describes a phenomenon in which
the value of actual inventory runs lower than recorded inventory, due to an unexplained
reduction in the number of units held. This measure is typically calculated as a mone-
tary figure or as a percentage of total stock value.
Pipeline Sales: Sales that are required to supply retail and wholesale channels with suf-
ficient inventory to make a product available for sale (refer to Section 6.5).
Consumer Off-Take: Purchases by consumers from retailers, as opposed to purchases
by retailers or wholesalers from their suppliers. When consumer off-take runs higher
than manufacturer sales rates, inventories will be drawn down.
Diverted Merchandise or Diverted Goods: Products shipped to one customer that are
subsequently resold to another customer. For example, if a retail drug chain overbuys
vitamins at a promotional price, it may ship some of its excess inventory to a dollar store.

6.8 SKU Profitability: Markdowns, GMROII, and DPP
Profitability metrics for retail products and categories are generally similar to other
measures of profitability, such as unit and percentage margins. Certain refinements
have been developed for retailers and distributors, however. Markdowns, for
example, are calculated as a ratio of discount to original price charged. Gross margin


return on inventory investment (GMROII) is calculated as margin divided by the
cost of inventory and is expressed as a “rate” or percentage. Direct product
profitability (DPP) is a metric that adjusts gross margin for other costs, such as
storage, handling, and allowances paid by suppliers.

Reduction in Price of SKU ($)
Markdown (%)
Initial Price of SKU ($)

Gross Margin on Product Sales
in Period ($)
Gross Margin Return on Inventory Investment (%)
Average Inventory Value at
Cost ($)

Direct Product Profitability ($) Gross Margin ($) Direct Product Costs ($)
By monitoring markdowns, marketers can gain important insight into SKU
profitability. GMROII can be a vital metric in determining whether sales rates
justify inventory positions. DPP is a theoretically powerful measure of profit that has
fallen out of favor, but it may be revived in other forms (for example, activity-based
costing).

Purpose: To assess the effectiveness and profitability of individual
product and category sales.
Retailers and distributors have a great deal of choice regarding which products to stock
and which to discontinue as they make room for a steady stream of new offerings. By
measuring the profitability of individual stock keeping units (SKUs), managers develop
the insight needed to optimize such product selections. Profitability metrics are also
useful in decisions regarding pricing, display, and promotional campaigns.
Figures that affect or reflect retail profitability include markdowns, gross margin return
on inventory investment, and direct product profitability. Taking each in turn:
Markdowns are not always applied to slow-moving merchandise. Markdowns in excess
of budget, however, are almost always regarded as indicators of errors in product assort-
ment, pricing, or promotion. Markdowns are often expressed as a percentage of regular
price. As a standalone metric, a markdown is difficult to interpret.
Gross margin return on inventory investment (GMROII) applies the concept of return
on investment (ROI) to what is often the most crucial element of a retailer’s working
capital: its inventory.
Direct product profitability (DPP) shares many features with activity-based costing
(ABC). Under ABC, a wide range of costs are weighted and allocated to specific products


through cost drivers—the factors that cause the costs to be incurred. In measuring DPP,
retailers factor such line items as storage, handling, manufacturer’s allowances, war-
ranties, and financing plans into calculations of earnings on specific product sales.

Construction
Markdown: This metric quantifies shop-floor reductions in the price of a SKU. It can be
expressed on a per-unit basis or as a total for the SKU. It can also be calculated in dollar
terms or as a percentage of the item’s initial price.
Markdown ($) Initial Price of SKU ($) Actual Sales Price ($)
Markdown ($)
Markdown (%)
Initial Price of SKU ($)

Gross Margin Return on Inventory Investment (GMROII): This metric quantifies the
profitability of products in relation to the inventory investment required to make them
available. It is calculated by dividing the gross margin on product sales by the cost of the
relevant inventory.
Gross Margin on Product Sales in
Period ($)
Gross Margin Return on Inventory Investment (%)
Average Inventory Value at Cost ($)

DIRECT PRODUCT PROFITABILITY (DPP)
Direct product profitability is grounded in a simple concept, but it can be difficult to
measure in practice. The calculation of DPP consists of multiple stages. The first stage is
to determine the gross margin of the goods in question. This gross margin figure is then
modified to take account of other revenues associated with the product, such as promo-
tional rebates from suppliers or payments from financing companies that gain business
on its sale. The adjusted gross margin is then reduced by an allocation of direct product
costs, described next.
Direct Product Costs: These are the costs of bringing a product to customers. They gen-
erally include warehouse, distribution, and store costs.
Direct Product Costs ($) Warehouse Direct Costs ($) Transportation Direct Costs ($)
Store Direct Costs ($)
Direct Product Profitability (DPP): Direct product profitability represents a product’s
adjusted gross margin, less its direct product costs.
As noted earlier, the concept of DPP is quite simple. Difficulties can arise, however, in
calculating or estimating the relevant costs. Typically, an elaborate ABC system is needed


to generate direct costs for individual SKUs. DPP has fallen somewhat out of favor as a
result of these difficulties.
Other metrics have been developed, however, in an effort to obtain a more refined and
accurate estimation of the “true” profitability of individual SKUs, factoring in the vary-
ing costs of receiving, storing, and selling them. The variations between products in the
levels of these costs can be quite significant. In the grocery industry, for example, the
cost of warehousing and shelving frozen foods is far greater—per unit or per dollar of
sales—than the cost of warehousing and shelving canned goods.
Direct Product Profitability ($) Gross Margin ($) Direct Product Costs ($)

EXAMPLE: The apparel retailer cited earlier wants to probe further into the prof-
itability of its sock line. Toward that end, it assembles the following information. For this
retailer, socks generate slotting allowances—in essence, fees paid by the manufacturer to
the retailer in compensation for shelf space—in the amount of $50,000 per year.
Warehouse costs for the retailer come to $10,000,000 per year. Socks consume 0.5% of
warehouse space. Estimated store and distribution costs associated with socks total
$80,000.
With this information, the retailer calculates an adjusted gross margin for its sock line.
Adjusted Gross Margin Gross Margin Additional Margin
$350,000 $50,000
$400,000
The retailer then calculates direct product costs for its sock line.
Direct Product Costs Store and Distribution Costs Warehouse Costs
$80,000 (0.5% * $10,000,000)
$80,000 $50,000
$130,000
On this basis, the retailer calculates the direct product profitability of its sock line.
DPP Gross Margin Direct Product Costs
$400,000 $130,000
$270,000

For GMROII calculations, it is necessary to determine the value of inventory held, at
cost. Ideally, this will be an average figure for the period to be considered. The average
of inventory held at the beginning and end of the period is often used as a proxy, and is


generally—but not always—an acceptable approximation. To perform the GMROII cal-
culation, it is also necessary to calculate a gross margin figure.
One of the central considerations in evaluating direct product profitability is an organi-
zation’s ability to capture large amounts of accurate data for analysis. The DPP calcula-
tion requires an estimate of the warehousing, distribution, store direct, and other costs
attributable to a product. To assemble these data, it may be necessary to gather all dis-
tribution costs and apportion them according to the cost drivers identified.
Inventory held, and thus the cost of holding it, can change considerably over time.
Although one may usually approximate average inventory over a period by averaging
the beginning and ending levels of this line item, this will not always be the case.
Seasonal factors may perturb these figures. Also, a firm may hold substantially more—
or less—inventory during the course of a year than at its beginning and end. This could
have a major impact on any DPP calculation.
DPP also requires a measure of the ancillary revenues tied to product sales.
Direct product profitability has great conceptual strength. It tries to account for the
wide range of costs that retailers incur in conveying a product to customers, and thus to
yield a more realistic measure of the profitability of that product. The only significant
weakness in this metric is its complexity. Few retailers have been able to implement it.
Many firms continue to try to realize its underlying concept, however, through such
programs as activity-based costing.

Shopping Basket Margin: The profit margin on an entire retail transaction,
which may include a number of products. This aggregate transaction is termed the
“basket” of purchases that a consumer makes.
One key factor in a firm’s profitability is its capability to sell ancillary products in addi-
tion to its central offering. In some businesses, more profit can be generated through
accessories than through the core product. Beverage and snack sales at movie theaters are
a prime example. With this in mind, marketers must understand each product’s role
within their firm’s aggregate offering—be it a vehicle to generate customer traffic, or to
increase the size of each customer’s basket, or to maximize earnings on that item itself.

Wilner, J.D. (1998). 7 Secrets to Successful Sales Management: The Sales Manager’s Manual, Boca
Raton: St. Lucie Press.
Zoltners, A.A., P. Sinha, and G.A. Zoltners. (2001). The Complete Guide to Accelerating Sales Force
Performance, New York: Amacom.


7
PRICING STRATEGY

Introduction

Price Premium Optimal Prices, Linear and Constant
Demand
Reservation Price
“Own,” “Cross,” and “Residual” Price
Percent Good Value
Elasticity
Price Elasticity of Demand

“The cost of . . . lack of sophistication in pricing is growing day by day. Customers and
Competitors operating globally in a generally more complex marketing environment
are making mundane thinking about pricing a serious threat to the firm’s financial
well being.”1
A full-fledged evaluation of pricing strategies and tactics is well beyond the scope of this
book. However, there are certain key metrics and concepts that are fundamental to the
analysis of pricing alternatives, and this chapter addresses them.
First we describe several of the more common methods of calculating price premiums—
also called relative prices.
Next, we discuss the concepts that form the foundation of price-quantity schedules—
also known as demand functions or demand curves. These include reservation prices
and percent good value.
In the third section, we explain the definition and calculation of price elasticity, a fre-
quently used index of market response to changes in price. This relatively simple ratio

219

of percentage changes in volumes and prices is complicated in practice by variations in
measure and interpretation.
For managers, the purpose of understanding price elasticity is to improve pricing. With
this in mind, we’ve devoted a separate section to determining optimal prices for the two
main types of demand functions: linear and constant elasticity. The final portion of this
chapter addresses the question of whether elasticity has been calculated in a manner
that incorporates likely competitive reactions. It explains three types of elasticity—
“own,” “cross,” and “residual” elasticity. Although these may seem at first glance to rest
upon subtle or pedantic distinctions, they have major pragmatic implications. The
familiar concept of the prisoner’s dilemma helps explain their import.


7.1 Price The percentage by Benchmarks include Measures how a
Premium which the price of average price brand’s price com-
a brand exceeds a paid, average pares to that of its
benchmark price. price charged, competition.
average price displayed,
and price of a relevant
competitor. Prices can
be compared at any
level in the channel
and can be calculated
on a gross basis or net
of discounts and
rebates.
7.2 Reservation The maximum Reservation prices are One way to con-
Price amount an indi- difficult to observe. ceptualize a
vidual is willing to demand curve is
pay for a product. as the aggregation
of reservation
prices of potential
customers.
7.2 Percent Good The proportion of Easier to observe than A second way to
Value customers who individual reservation conceptualize a
consider a product prices. demand curve is
to be a good as the relationship
value—that is, to between percent
have a selling price good value and
below their reser- price.
vation price.



7.3 Price The responsiveness For linear demand, Measures the
Elasticity of of demand to a linear projections responsiveness of
Demand small change in based on elasticity are quantity to
price, expressed accurate, but elasticity changes in price. If
as a ratio of changes with price. priced optimally,
percentages. For constant elasticity the margin is the
demand, linear pro- negative inverse of
jections are approxi- elasticity.
mate, but elasticity is
the same for all
prices.
7.4 Optimal Price For linear demand, Optimal price formu- Quickly deter-
optimal price is las are appropriate mines the price
the average of only if the variable that maximizes
variable cost and cost per unit is con- contribution.
the maximum stant, and there are no
reservation price. larger strategic con-
For constant elas- siderations.
ticity, optimal
price is a known
function of vari-
able cost and elas-
ticity. In general,
optimal price is
the price that
maximizes contri-
bution after
accounting for
how quantity
changes with
price.
7.5 Residual Residual elasticity Rests on an assump- Measures the
Elasticity is “own” elasticity tion that competitor responsiveness of
plus the product of reaction to a firm’s quantity to
competitor reac- price changes is pre- changes in price,
tion elasticity and dictable. after accounting
cross elasticity. for competitor
reactions.

Chapter 7 Pricing Strategy 221

7.1 Price Premium
Price premium, or relative price, is the percentage by which a product’s selling price
exceeds (or falls short of) a benchmark price.
[Brand A Price ($) Benchmark Price ($)]
Price Premium (%)
Benchmark Price ($)

Marketers need to monitor price premiums as early indicators of competitive pricing
strategies. Changes in price premiums can also be signs of product shortages, excess
inventories, or other changes in the relationships between supply and demand.

Purpose: To evaluate product pricing in the context
of market competition.
Although there are several useful benchmarks with which a manager can compare a
brand’s price, they all attempt to measure the “average price” in the marketplace. By
comparing a brand’s price with a market average, managers can gain valuable insight
into its strength, especially if they view these findings in the context of volume and
market share changes. Indeed, price premium—also known as relative price—is a com-
monly used metric among marketers and senior managers. Fully 63% of firms report
the Relative Prices of their products to their boards, according to a recent survey con-
ducted in the U.S., UK, Germany, Japan, and France.2
Price Premium: The percentage by which the price charged for a specified brand
exceeds (or falls short of) a benchmark price established for a similar product or bas-
ket of products. Price premium is also known as relative price.

Construction
In calculating price premium, managers must first specify a benchmark price. Typically,
the price of the brand in question will be included in this benchmark, and all prices in
the benchmark will be for an equivalent volume of product (for example, price per
liter). There are at least four commonly used benchmarks:
■ The price of a specified competitor or competitors.
■ Average price paid: The unit-sales weighted average price in the category.
■ Average price displayed: The display-weighted average price in the category.
■ Average price charged: The simple (unweighted) average price in the category.
Price of a Specified Competitor: The simplest calculation of price premium
involves the comparison of a brand’s price to that of a direct competitor.


EXAMPLE: Ali’s company sells “gO2” mineral water in its EU home market at a 12%
premium over the price of its main competitor. Ali would like to know whether the same
price premium is being maintained in the Turkish market, where gO2 faces quite differ-
ent competition. He notes that gO2 mineral water sells in Turkey for 2 (new) Lira per
liter, while its main competitor, Essence, sells for 1.9 Lira per liter.
(2.0 YTL 1.9 YTL)
Price Premium
1.9 YTL
0.1 YTL
1.9 YTL 5.3% Premium Versus Essence

When assessing a brand’s price premium vis à vis multiple competitors, managers can
use as their benchmark the average price of a selected group of those competitors.
Average Price Paid: Another useful benchmark is the average price that customers
pay for brands in a given category. This average can be calculated in at least two
ways: (1) as the ratio of total category revenue to total category unit sales, or (2) as
the unit-share weighted average price in the category. Note that the market Average
Price Paid includes the brand under consideration.
Note also that changes in unit shares will affect the average price paid. If a low-price brand
steals shares from a higher-priced rival, the average price paid will decline. This would
cause a firm’s price premium (calculated using the average price paid as a benchmark) to
rise, even if its absolute price did not change. Similarly, if a brand is priced at a premium,
that premium will decline as it gains share. The reason: A market share gain by a premium-
priced brand will cause the overall average price paid in its market to rise. This, in turn,
will reduce the price differential between that brand and the market average.

EXAMPLE: Ali wants to compare his brand’s price to the average price paid for simi-
lar products in the market. He notes that gO2 sells for 2.0 Lira per liter and has 20% of
the unit sales in market. Its up-market competitor, Panache, sells for 2.1 Lira and enjoys
10% unit market share. Essence sells for 1.9 Lira and has 20% share. Finally, the budget
brand, Besik, sells for 1.2 Lira and commands 50% of the market.

Ali calculates the weighted Average Price Paid as (20% * 2) (10% * 2.1)
(20% * 1.9) (50% * 1.2) 1.59 Lira.
(2.00 1.59)
Price Premium (%)
1.59
0.41
1.59
25.8%


To calculate the price premium using the average price paid benchmark, managers can
also divide a brand’s share of the market in value terms by its share in volume terms. If
value and volume market shares are equal, there is no premium. If value share is greater
than volume share, then there is a positive price premium.
Price Premium (%)

Average Price Charged: Calculation of the average price paid requires knowledge of
the sales or shares of each competitor. A much simpler benchmark is the average
price charged—the simple unweighted average price of the brands in the category.
This benchmark requires knowledge only of prices. As a consequence, the price pre-
mium calculated using this benchmark is not affected by changes in unit shares. For
this reason, this benchmark serves a slightly different purpose. It captures the way a
brand’s price compares to prices set by its competitors, without regard to customers’
reactions to those prices. It also treats all competitors equally in the calculation of the
benchmark price. Large and small competitors are weighted equally when calculat-
ing average price charged.

EXAMPLE: Using the previous data, Ali also calculates the average price charged in
the mineral water category as (2 2.1 1.9 1.2)/4 1.8 Lira.
Using the average price charged as his benchmark, he calculates gO2’s price premium as
(2.0 1.8)
Price Premium (%)
1.8
0.2
1.8
11.1% Premium

Average Price Displayed: One benchmark conceptually situated between average
price paid and average price charged is the average price displayed. Marketing man-
agers who seek a benchmark that captures differences in the scale and strength of
brands’ distribution might weight each brand’s price in proportion to a numerical
measure of distribution. Typical measures of distribution strength include numeric
distribution, ACV (%), and PCV (%).

EXAMPLE: Ali calculates the average price displayed using numeric distribution.
Ali’s brand, gO2, is priced at 2 Lira and is distributed in 500 of the 1,000 stores that carry bot-
tled water. Panache is priced at 2.1 Lira and stocked by 200 stores. Essence is priced at 1.9 Lira
and sold through 400 stores. Besik carries a price of 1.2 Lira and has a presence in 900 stores.


Ali calculates relative weighting on the basis of numeric distribution. The total number
of stores is 1,000. The weightings are therefore, for gO2, 500/1,000 50%; for Panache,
200/1,000 20%; for Essence, 400/1,000 40%; and for Besik, 900/1,000 90%. As the
weightings thus total 200%, in calculating average price displayed, the sum of the
weighted prices must be divided by that figure, as follows:
[(2 * 50%) (2.1 * 20%) + (1.9 * 40%) (1.2 * 90%)]
Average Price Displayed
200%

1.63 Lira
(2.00 1.63)
Price Premium (%)
1.63
0.37
1.63
22.7% premium

There are several practical aspects of calculating price premiums that deserve mention.
Managers may find it easier to select a few leading competitors and focus their analysis and
comparison on these. Often, it is difficult to obtain reliable data on smaller competitors.
Managers must exercise care when interpreting price premiums. Different benchmarks
measure different types of premiums and must be interpreted accordingly.
Can a price premium be negative? Yes. Although generally expressed in terms that
imply only positive values, a price premium can be negative. If one brand doesn’t com-
mand a positive premium, a competitor will. Consequently, except in the unlikely event
that all prices are exactly equal, managers may want to speak in terms of positive premi-
ums. When a given brand’s price is at the low end of the market, managers may want to
say that the competition holds a price premium of a certain value.
Should we use retail, manufacturer, or distributor pricing? Each is useful in under-
standing the market dynamics at its level. When products have different channel
margins, their price premiums will differ, depending on the channel under considera-
tion. When stating a price premium, managers are advised to specify the level to which
it applies.
Prices at each level can be calculated on a gross basis, or net of discounts, rebates, and
coupons. Especially when dealing with distributors or retailers, there are likely to be
substantial differences between manufacturer selling prices (retail purchase prices),
depending on whether they are adjusted for discounts and allowances.


Theoretical Price Premium: This is the price difference that would make potential
customers indifferent between two competing products. It represents a different use
of the term “price premium” that is growing in popularity. The theoretical price pre-
mium can also be discovered through a conjoint analysis using brand as an attrib-
ute. The theoretical price premium is the point at which consumers would be
indifferent between a branded and an unbranded item, or between two different
brands. We have termed this a “theoretical” price premium because there is no guar-
antee that the price premiums observed in the market will take this value. (Refer to
Section 4.5 for an explanation of conjoint analysis.)

7.2 Reservation Price and Percent Good Value
The reservation price is the value a customer places on a product. It constitutes an
individual’s maximum willingness to pay. Percent good value represents the propor-
tion of customers who believe a product is a “good value” at a specific price.
These are useful metrics in marketers’ evaluation of pricing and customer value.

Purpose
Reservation prices provide a basis for estimating products’ demand functions in situa-
tions where other data are not available. They also offer marketers insight into pricing
latitude. When it is not possible or convenient to ask customers about their reservation
prices, percent good value can provide a substitute for that metric.

Construction
Reservation Price: The price above which a customer will not buy a product. Also
known as the maximum willingness to pay.
Percent Good Value: The proportion of customers who perceive a product to repre-
sent a good value, that is, to carry a selling price at or below their reservation price.
By way of example, let’s posit a market consisting of 11 individuals with reservation
prices for a given product of $30, $40, $50, $60, $70, $80, $90, $100, $110, $120, and
$130. The manufacturer of that product seeks to decide upon its price. Clearly, it might
do better than to offer a single price. For now, however, let’s assume tailored prices are
impractical. The variable cost to produce the product is $60 per unit.


With these reservation prices, the manufacturer might expect to sell 11 units at $30 or
less, 10 units at a price greater than $30 but less than or equal to $40, and so on. It would
make no sales at a unit price greater than $130. (For convenience, we have assumed that
people buy at their reservation price. This assumption is consistent with a reservation
price being the maximum an individual is willing to pay.)
Table 7.1 shows this price-quantity relationship, together with the contribution to the
firm at each possible price.

Table 7.1 Price-Quantity Relationship

Price % Good Value Quantity Total Contribution

$20 100.00% 11 $440

$30 100.00% 11 $330
$40 90.91% 10 $200

$50 81.82% 9 $90

$60 72.73% 8 $0

$70 63.64% 7 $70

$80 54.55% 6 $120

$90 45.45% 5 $150

$100 36.36% 4 $160

$110 27.27% 3 $150

$120 18.18% 2 $120

$130 9.09% 1 $70

$140 0.00% 0 $0

$150 0.00% 0 $0

Variable Cost is $60 per unit.

A table of quantities expected at each of several prices is often called a demand sched-
ule (or curve). This example shows that one way to conceptualize a demand curve is as
the accumulation of individual reservation prices. Although it will clearly be difficult
in practice to measure individual reservation prices, the point here is simply to illus-
trate the use of reservation prices in pricing decisions. In this example, the optimal


price—that is, the price that maximizes total contribution—is $100. At $100, the man-
ufacturer expects to sell four units. Its contribution margin is $40, yielding a total con-
tribution of $160.
This example also illustrates the concept of consumer surplus. At $100, the manufac-
turer sells three items at a price point below customers’ reservation prices. The con-
sumer with the reservation price of $110 enjoys a surplus of $10. The consumer with the
reservation price of $120 receives a surplus of $20. Finally, the consumer with the high-
est reservation price, $130, receives a surplus of $30. From the manufacturer’s perspec-
tive, the total consumer surplus—$60—represents an opportunity for increased
contribution if it can find a way to capture this unclaimed value.

Finding reservation prices is no easy matter. Two techniques that are frequently used to
gain insight into this metric are as follows:
■ Second-price auctions: In a second-price auction, the highest bidder wins but
pays only the second-highest bid amount. Auction theory suggests that when
bidding on items of known value in such auctions, individuals have an incen-
tive to bid their reservation prices. Certain survey techniques have been
designed to mimic this process. In one of these, customers are asked to name
their prices for an item, with the understanding that these prices will then be
subjected to a lottery. If the price drawn in the lottery is less than the price
named, the respondent gains an opportunity to purchase the item in question
at the drawn price.
■ Conjoint analysis: In this analytical technique, marketers gain insight into cus-
tomer perceptions regarding the value of any set of attributes through the
trade-offs they are willing to make.
Such tests can, however, be difficult to construct and impractical in many circum-
stances. Consequently, as a fallback technique, marketers can measure percent good
value. Rather than seeking to learn each customer’s reservation price, they may find it
easier to test a few candidate prices by asking customers whether they consider an item
a “good value” at each of those prices.

Linear Demand
The quantity-price schedule formed by an accumulation of reservation prices can take a
variety of shapes. When the distribution of reservation prices is uniform—when reserva-
tion prices are equally spaced, as in our example—the demand schedule will be linear (see
Figure 7.1). That is, each increment in price will reduce quantity by an equal amount.As the


Quantity Maximum Willing to Buy (MWB)
Demanded

Two Points on the Linear
Demand Function

Maximum Reservation
Price (MRP)

Price
Variable Cost

Figure 7.1 Maximum Willing to Buy and Maximum Reservation Price

linear function is by far the most commonly used representation of demand, we provide a
description of this function as it relates to the distribution of underlying reservation prices.
It takes only two points to determine a straight line. Likewise, it takes only two parame-
ters to write an equation for that line. Generally, that equation is written as Y mX
b, in which m is the slope of the line and b is its Y-intercept.
A line, however, can also be defined in terms of the two points where it crosses the axes.
In the case of linear demand, these crossing points (intercepts) have useful managerial
interpretations.
The quantity-axis intercept can be viewed as a representation of the maximum willing
to buy (MWB). This is the total number of potential customers for a product. A firm
can serve all these customers only at a price of zero. Assuming that each potential
customer buys one unit, MWB is the quantity sold when the price is zero.
The price-axis intercept can be viewed as the maximum reservation price (MRP). The
MRP is a number slightly greater than the highest reservation price among all those
willing to buy. If a firm prices its product at or above MRP, no one will buy.
Maximum Reservation Price: The lowest price at which quantity demanded
equals zero.


Maximum Willing to Buy (MWB): The quantity that customers will “buy” when
the price of a product is zero. This is an artificial concept used to anchor a linear
demand function.
In a linear demand curve defined by MWB and MRP, the equation for quantity (Q) as a
function of price (P) can be written as follows:
P
Q (MWB) * [1 ]
MRP

EXAMPLE: Erin knows that the demand for her soft drink is a simple linear function
of price. She can sell 10 units at a price of zero. When the price hits $5 per unit, demand
falls to zero. How many units will Erin sell if the price is $3 (see Figure 7.2)?

Linear Demand: Price and Quantity Demanded

12

10 Maximum Willing to Buy
Quantity Demanded

8
Maximum Reservation Price
6

4

2

0

$0 $1 $2 $3 $4 $5
Price

Figure 7.2 Simple Linear Demand (Price-Quantity) Function

For Erin’s soft drink, the MRP (Maximum Reservation Price) is $5 and the MWB (Maximum
Willing to Buy ) is 10 units. At a price of $3, Erin will sell 10 * (1 $3/$5), or 4 units.


When demand is linear, any two points on the price-quantity demand function can be
used to determine MRP and MWB. If P1 and Q1 represent the first price-quantity point
on the line, and P2 and Q2 represent the second, then the following two equations can
be used to calculate MWB and MRP.
Q2 Q1
MWB Q1 ( ) P
P2 P1 * 1
P2 P1
MRP P1 ( )
Q2 Q1

EXAMPLE: Early in this chapter, we met a firm that sells five units at a price of $90
and three units at a price of $110. If demand is linear, what are MWB and MRP?

MWB 5 ( 2/$20) * $90
5 9
14
MRP $90 ($20/ 2) * 5
$90 $50
$140
The equation for quantity as a function of price is thus:
P
Q 14 * (1 )
$140
The market in this example, as you may recall, comprises 11 potential buyers with reser-
vation prices of $30, $40, . . . , $120, $130. At a price of $130, the firm sells one unit. If
we set price equal to $130 in the previous equation, our calculation does indeed result
in a quantity of one. For this to hold true, the MRP must be a number slightly higher
than $130.
A linear demand function often yields a reasonable approximation of actual demand only
over a limited range of prices. In our 11-person market, for example, demand is linear
only for prices between $30 and $130. To write the equation of the linear function that
describes demand between $30 and $130, however, we must use an MWB of 14 and an
MRP of $140. When we use this linear equation, we must remember that it reflects actual
demand only for prices between $30 and $130, as illustrated in Figure 7.3.


Quantity Linear Demand Assumption
Demanded
15
14
Quantity
13
Demanded
12
Linear
11 Demand
10
9
8
7
6
5
4
3
2
1
0
0 0 0 0 0 0 0 0 0 00 10 20 30 40
$0 $1 $2 $3 $4 $5 $6 $7 $8 $9 $1 $1 $1 $1 $1

Price

Figure 7.3 Example of Linear Demand Function

7.3 Price Elasticity of Demand
Price elasticity measures the responsiveness of quantity demanded to a small change
in price.
Change in Quantity (%)
Price Elasticity (I)
Change in Price (%)

Price elasticity can be a valuable tool, enabling marketers to set an optimal price.


Purpose: To understand market responsiveness to changes in price.
Price elasticity is the most commonly employed measure of market responsiveness to
changes in price. Many marketers, however, use this term without a clear understanding
of what it entails. This section will help clarify some of the potentially dangerous details
associated with estimates of price elasticity. This is challenging material but is well
worth the effort. A strong command of price elasticity can help managers set optimal
prices.
Price Elasticity: The responsiveness of demand to a small change in price, expressed
as a ratio of percentages. If price elasticity is estimated at 1.5, for example, then we
expect the percentage change in quantity to be approximately 1.5 times the percent-
age change in price. The fact that this number is negative indicates that when price
rises, the quantity demanded is expected to decline, and vice versa.

Construction
If we raise the price of a product, do we expect demand to hold steady or crash through
the floor? In markets that are unresponsive to price changes, we say demand is inelastic.
If minor price changes have a major impact on demand, we say demand is elastic. Most
of us have no trouble understanding elasticity at a qualitative level. The challenges come
when we quantify this important concept.

CHALLENGE ONE: QUESTIONS OF SIGN.
The first challenge in elasticity is to agree on its sign. Elasticity is the ratio of the per-
centage change in quantity demanded to the percentage change in price, for a small
change in price. If an increase in price leads to a decrease in quantity, this ratio will be
negative. Consequently, by this definition, elasticity will almost always be a negative
number.
Many people, however, simply assume that quantity goes down as price goes up, and
jump immediately to the question of “by how much.” For such people, price elasticity
answers that question and is a positive number. In their eyes, if elasticity is 2, then a
small percentage increase in price will yield twice that percentage decrease in quantity.
In this book, under that scenario, we would say price elasticity is 2.

CHALLENGE TWO: WHEN DEMAND IS LINEAR, ELASTICITY CHANGES WITH PRICE.
For a linear demand function, the slope is constant, but elasticity is not. The reason:
Elasticity is not the same as slope. Slope is the change in quantity for a small change in
price. Elasticity, by contrast, is the percentage change in quantity for a small percentage
change in price.


EXAMPLE: Consider three points on a linear demand curve: ($8, 100 units), ($9, 80
units), and ($10, 60 units) (see Figure 7.4). Each dollar change in price yields a 20-unit
change in quantity. The slope of this curve is a constant 20 units per dollar.
As price rises from $8 to $9 (a 12.5% increase), quantity declines from 100 to 80 (a 20%
decrease). The ratio of these percentages is 20%/12.5%, or 1.6. Similarly, as price rises
from $8 to $10 (a 25% increase), quantity declines from 100 to 60 (a 40% decrease).
Once again, the ratio (40%/25%) is 1.6. It appears that the ratio of percentage change
in quantity to percentage change in price is 1.6, regardless of the size of the change
made in the $8 price.

Linear Demand

140

120

100
Quantity

80

60

40

20

0
$6.00 $8.00 $10.00 $12.00

Price

Figure 7.4 Linear Demand Function

Consider, however, what happens when price rises from $9 to $10 (an 11.11% increase).
Quantity declines from 80 to 60 (a 25% decrease). The ratio of these figures, 25%/
11.11%, is now 2.25. A price decline from $9 to $8 also yields an elasticity ratio of
2.25. It appears that this ratio is 2.25 at a price of $9, regardless of the direction of
any change in price.
Exercise: Verify that the ratio of percentage change in quantity to percentage change in
price at the price of $10 is 3.33 for every conceivable price change.


For a linear demand curve, elasticity changes with price. As price increases, elasticity
gains in magnitude. Thus, for a linear demand curve, the absolute unit change in quan-
tity for an absolute dollar change in price (slope) is constant, while the percentage
change in quantity for a percentage change in price (elasticity) is not. Demand becomes
more elastic—that is, elasticity becomes more negative—as price increases.
For a linear demand curve, the elasticity of demand can be calculated in at least
three ways:
Q2 Q1
Q1
Elasticity (P1)
P2 P1
P1
Q2 Q1 P1
P2 P1 * Q1
P1
Slope *
Q1

To emphasize the idea that elasticity changes with price on a linear demand curve, we
write “Elasticity (P),” reflecting the fact that elasticity is a function of price. We also use
the term “point elasticity” to cement the idea that a given elasticity applies only to a sin-
gle point on the linear demand curve.
Equivalently, because the slope of a linear demand curve represents the change in quan-
tity for a given change in price, price elasticity for a linear demand curve is equal to the
slope, multiplied by the price, divided by the quantity. This is captured in the third
equation here.

EXAMPLE: Revisiting the demand function from earlier, we see that the slope of the
curve reflects a 20-unit decline in demand for each dollar increase in price. That is, slope
equals 20.
The slope formula for elasticity can be used to verify our earlier calculations. Calculate
price/quantity at each point on the curve, and multiply this by the slope to yield the price
elasticity at that point (see Table 7.2).
For example, at a price of $8, quantity sold is 100 units. Thus:
Elasticity ($8) 20 * (8/100)
1.6


Table 7.2 Elasticities at a Point Calculated from the Slope of a Function

Price Quantity Demanded Price/Quantity Slope Price Elasticity at Point

$8.00 100 0.08 (20.00) (1.60)

$9.00 80 0.11 (20.00) (2.25)
$10.00 60 0.17 (20.00) (3.33)

In a linear demand function, point elasticities can be used to predict the percentage change
in quantity to be expected for any percentage change in price.

EXAMPLE: Xavi manages the marketing of a toothpaste brand. He knows the brand
follows a linear demand function. At the current price of $3.00 per unit, his firm cur-
rently sells 60,000 units with an elasticity of 2.5. A proposal is floated to raise the price
to $3.18 per unit in order to standardize margins across brands. At $3.18, how many
units would be sold?
The proposed change to $3.18 represents a 6% increase over the current $3 price. Because
elasticity is 2.5, such an increase can be expected to generate a decrease in unit sales of
2.5 * 6, or 15%. A 15% reduction in current sales of 60,000 units would yield a new quan-
tity of 0.85 * 60,000, or 51,000.

Constant Elasticity: Demand Curve with a Constantly Changing Slope
A second common form of function used to estimate demand entails constant elastic-
ity.3 This form is responsible for the term “demand curve” because it is, indeed, curved.
In contrast with the linear demand function, the conditions in this scenario are
reversed: Elasticity is constant, while the slope changes at every point.
The assumption underlying a constant elasticity demand curve is that a small percent-
age change in price will cause the same percentage change in quantity, regardless of the
value of the initial price. That is, the rate of change in quantity versus price, expressed
as a ratio of percentages, is equal to a constant throughout the curve. That constant is
the elasticity.
In mathematical terms, in a constant elasticity demand function, slope multiplied by
price divided by quantity is equal to a constant (the elasticity) for all points along the
curve (see Figure 7.5). The constant elasticity function can also be expressed in an equa-
tion that is easily calculated in spreadsheets:
Q(P) A * P ELAS


Constant Elasticity Function

120

101.7
100
Quantity Demanded

80 78.0

61.5
60

40

20

0

$6 $7 $8 $9 $10 $11
Price

Figure 7.5 Constant Elasticity

In this equation, ELAS is the price elasticity of demand. It is usually a negative number.
A is a scaling factor. It can be viewed as the quantity that would be sold at a price of $1
(assuming that $1 is a reasonable price for the product under consideration).

EXAMPLE: Plot a demand curve with a constant elasticity of 2.25 and a scaling fac-
tor of 10,943.1. For every point on this curve, a small percentage increase in price will
yield a percentage decrease in quantity that is 2.25 times as great. This 2.25 ratio holds,
however, only for the very smallest percentage changes in price. This is because the slope
changes at every point. Using the 2.25 ratio to project the results of a finite percentage
increase in price is always approximate.
The curve traced in this example should look like the constant elasticity curve in Figure 7.5.
More exact figures for demand at prices $8, $9, and $10 would be 101.669, 78.000, and
61.538 units.

In its way, constant elasticity is analogous to the continuous compounding of interest.
In a constant elasticity function, every small percentage increase in price generates the
same percentage decrease in quantity. These percentage decreases compound at a con-
stant rate, leading to an overall percentage decrease that does not precisely equal the
continuous rate.


For this reason, given any two points on a constant elasticity demand curve, we can no
longer calculate elasticity using finite differences as we could when demand was linear.
Instead, we must use a more complicated formula grounded in natural logarithms:
ln(Q2 Q1)
ELAS
ln(P2 P1)

EXAMPLE: Taking any two points from the previous constant elasticity demand
curve, we can verify that elasticity is 2.25.
At $8, for example, the quantity is 101.669. Call these P1 and Q1.
At $9 the quantity is 78.000. Call these P2 and Q2.
Inserting these into our formula, we determine that
ln(78.000 101.669)
ELAS
ln(9 8)

0.265
0.118

2.25
If we had set P2 equal to $8, and P1 equal to $9, we would have arrived at the same figure for
elasticity. In fact, regardless of which two points we select on this constant elasticity curve,
and regardless of the order in which we consider them, elasticity will always be 2.25.

In summary, elasticity is the standard measure of market responsiveness to changes in
price. In general, it is the “percentage slope” of the demand function (curve) obtained by
multiplying the slope of the curve for a given price by the ratio of price to quantity.
P
Elasticity(P) Slope *
Q

Elasticity can also be viewed as the percentage change in quantity for a small percentage
change in price.
In a linear demand function, the slope is constant, but elasticity changes with price.
In this scenario, marketers can use elasticity estimates to calculate the result of an
anticipated price change in either direction, but they must use the elasticity that
is appropriate for their initial price point. The reason: In a linear demand func-
tion, elasticity varies across price points, but projections based on these elasticities
are accurate.
In a constant elasticity demand function, elasticity is the same at all price points, but
projections based on these elasticities will be approximate. Assuming they are estimated


with precision, using the constant elasticity demand function itself to make sales pro-
jections on the basis of price changes will be more accurate.

Price elasticity is generally estimated on the basis of available data. These data can be
drawn from actual sales and price changes observed in the market, conjoint studies of
customer intentions, consumer surveys about reservation prices or percent good value,
or test-market results. In deriving elasticity, price-quantity functions can be sketched on
paper, estimated from regressions in the form of linear or constant elasticity equations,
or estimated through more complex expressions that include other variables in the mar-
keting mix, such as advertising or product quality.
To confirm the validity and usefulness of these procedures, marketers must thoroughly
understand the implications of the resulting elasticity estimate for customer behavior.
Through this understanding, marketers can determine whether their estimate makes
sense or requires further validation. That done, the next step is to use it to decide
on pricing.

7.4 Optimal Prices and Linear and Constant Demand
Functions
The optimal price is the most profitable price for any product. In a linear demand
function, the optimal price is halfway between the maximum reservation price and
the variable cost of the product.
[Maximum Reservation Price ($)
Variable Cost ($)]
Optimal Price for a Linear Demand Function ($)
2
Generally, the gross margin on a product at its optimal price will be the negative
inverse of its price elasticity.
1
Gross Margin at Optimal Price (%)
Elasticity (I)
Although it can be difficult to apply, this relationship offers a powerful insight: In a
constant elasticity demand function, optimal margin follows directly from elasticity.
This greatly simplifies the determination of the optimal price for a product of known
variable cost.


Purpose: To determine the price that yields the greatest possible
contribution.
Although “optimal price” can be defined in a number of ways, a good starting point is
the price that will generate the greatest contribution by a product after deducting its
variable cost—that is, the most profitable price for the product.
If managers set price too low, they forego revenue from customers who would willingly
have paid more. In addition, a low price can lead customers to value a product less than
they otherwise might. That is, it causes them to lower their reservation prices.
By contrast, if managers set price too high, they risk losing contribution from people
who could have been served profitably.

Construction
For linear demand, the optimal price is the midpoint between the maximum reservation
price and the variable cost of the product.
In linear demand functions, the price that maximizes total contribution for a product is
always precisely halfway between the maximum reservation price (MRP) and the vari-
able cost to produce that product. Mathematically, if P* represents the optimal price of
a product, MRP is the X-intercept of its linear demand function, and VC is its variable
cost per unit:

P* (MRP VC) 2

EXAMPLE: Jaime’s business sells goods that cost $1 to produce. Demand is linear. If
priced at $5, Jaime believes he won’t sell anything. For every dollar decrease in price,
Jaime believes he will sell one additional unit.
Given that the variable cost is $1, the maximum reservation price is $5, and the demand
function is linear, Jaime can anticipate that he’ll achieve maximum contribution at a
price midway point between VC and MRP. That is, the optimal price is ($5 $1)/2, or
$3.00 (see Figure 7.6).4
In a linear demand function, managers don’t need to know the quantity of a product
demanded in order to determine its optimal price. For those who seek to examine Jaime’s
contribution figures, however, please find the details in Table 7.3.


Maximum Total Contribution When “Square” is Formed
5

Total
Contribution
4
Variable
Quantity Demanded
Cost
3

2

Quantity
1 Contribution Demanded

0
$- $1 $2 $3 $4 $5

-1

Price/Cost

Figure 7.6 Optimal Price Midway Between Variable Cost and MRP

Table 7.3 Optimal Price 1/ (MRP Variable Cost)
2

Price Quantity Variable Cost Contribution Total
Demanded per Unit per Unit Contribution
$0 5 $1 ($1) ($5)

$1 4 $1 $0 $0
$2 3 $1 $1 $3

$3 2 $1 $2 $4

$4 1 $1 $3 $3

$5 0 $1 $4 $0


The previous optimal price formula does not reveal the quantity sold at a given price or
the resulting contribution. To determine optimal contribution, managers can use the
following equation:

Contribution* (MWB/MRP) * (P* VC)2

EXAMPLE: Jaime develops a new but similar product. Its demand follows a linear
function in which the maximum willing to buy (MWB) is 200 and the maximum reser-
vation price (MRP) is $10. Variable cost is $1 per unit. Jaime knows that his optimal price
will be midway between MRP and variable cost. That is, it will be ($1 $10)/2 $5.50
per unit. Using the formula for optimal contribution, Jaime calculates total contribution
at the optimal price:

Contribution at Optimal Price for a Linear Demand Function ($)
[MWB (#)/MRP ($)] * [Price ($) Variable Costs ($)] ^ 2
(200/10) * ($5.50 $1) ^ 2
20 * $4.5 ^ 2
$405
Jaime builds a spreadsheet that supports this calculation (see Table 7.4).

Table 7.4 Contribution Maximized at the Optimal Price

Price Variable Quantity Contribution Total Contribution
Costs Demanded per Unit
$6 $1 80 $5.00 $400

$5.50 $1 90 $4.50 $405
$5 $1 100 $4.00 $400

$4 $1 120 $3.00 $360

$3 $1 140 $2.00 $280

$2 $1 160 $1.00 $160

$1 $1 180 $0.00 $0


This relationship holds across all linear demand functions, regardless of slope. For such
functions, it is therefore possible to calculate the optimal price for a product on the basis
of only two inputs: variable cost per unit and the maximum reservation price.

EXAMPLE: Brands A, B, and C each have a variable cost of $2 per unit and follow lin-
ear demand functions as shown in Table 7.5.

Table 7.5 The Optimal Price Formula Applies to All Linear Demand Functions

Price Demand Brand A Demand Brand B Demand Brand C
$2 12 20 16

$3 10 18 15
$4 8 16 14
$5 6 14 13

$6 4 12 12

$7 2 10 11

$8 0 8 10

$9 0 6 9

$10 0 4 8

$11 0 2 7

$12 0 0 6

On the basis of these inputs, we can determine the maximum reservation price—the
lowest price at which demand is zero. For Brand C, for example, we know that demand
follows a linear function in which quantity declines by one unit for each dollar increase
in price. If six units are demanded at $12, then $18 will be the lowest price at which no
one will buy a single unit. This is the maximum reservation price. We can make similar
determinations for Brands A and B (see Table 7.6).


Table 7.6 In Linear Demand Functions, the Determination of
Optimal Price Requires Only Two Inputs

Brand A Brand B Brand C
Maximum Reservation Price $8 $12 $18

Variable Costs $2 $2 $2
Optimal Price $5 $7 $10

To verify that the optimal prices so determined will generate the maximum attainable
contribution, please see Table 7.7.

Table 7.7 The Optimal Prices for Linear Demand Functions Can Be Verified

Unit Total Total Total
Contribu- Demand Contri- Demand Contri- Demand Contri-
Variable tion Brand A bution Brand B bution Brand C bution
Price costs = P - VC (Given) Brand A (Given) Brand B (Given) Brand C
P VC UC Q Q*UC Q Q*UC Q Q*UC
$2 $2 $0 12 $0 20 $0 16 $0
$3 $2 $1 10 $10 18 $18 15 $15
$4 $2 $2 8 $16 16 $32 14 $28
$5 $2 $3 6 $18 14 $42 13 $39
$6 $2 $4 4 $16 12 $48 12 $48
$7 $2 $5 2 $10 10 $50 11 $55
$8 $2 $6 0 $0 8 $48 10 $60
$9 $2 $7 0 $0 6 $42 9 $63
$10 $2 $8 0 $0 4 $32 8 $64
$11 $2 $9 0 $0 2 $18 7 $63
$12 $2 $10 0 $0 0 $0 6 $60

Because slope doesn’t influence optimal price, all demand functions with the same max-
imum reservation price and variable cost will yield the same optimal price.


EXAMPLE: A manufacturer of chair cushions operates in three different markets—
urban, suburban, and rural. These vary greatly in size. Demand is far higher in the city
than in the suburbs or the country. Variable cost, however, is the same in all markets at $4
per unit. The maximum reservation price, at $20 per unit, is also the same in all markets.
Regardless of market size, the optimal price is therefore $12 per unit in all three markets
(see Figure 7.7 and Table 7.8).
The optimal price of $12 is verified by the calculations in Table 7.9.

Different Linear Demand Functions Slopes with the Same
MWP and VC
35
Variable Cost $4
30

25
Quantity Demanded

20 Suburban
Urban
Demand
Demand
15

10

5 Rural Demand

0
$- $2 $4 $6 $8 $10 $12 $14 $16 $18 $20

Price

Figure 7.7 Linear Demand Functions with the Same MRP and Variable Cost

Table 7.8 The Slope Doesn’t Influence Optimal Price

Maximum Reservation Price $20

Variable Cost $4

Optimal Price $12


Table 7.9 Linear Demand Functions with Different Slopes

Price Contri- Suburban Rural Urban Suburban Rural Urban
bution Demand Demand Demand Contri- Contri- Contri-
bution bution bution
$0 ($4) 20 10 32 ($80) ($40) ($128)

$2 ($2) 18 9 29 ($36) ($18) ($58)

$4 $0 16 8 26 $0 $0 $0

$6 $2 14 7 22 $28 $14 $45

$8 $4 12 6 19 $48 $24 $77

$10 $6 10 5 16 $60 $30 $96

$12 $8 8 4 13 $64 $32 $102

$14 $10 6 3 10 $60 $30 $96

$16 $12 4 2 6 $48 $24 $77

$18 $14 2 1 3 $28 $14 $45

$20 $16 — — — — — —

In this example, it might help to think of the urban, suburban, and rural markets as
groups of people with identical, uniform distributions of reservation prices. In each, the
reservation prices are uniform between $0 and the maximum reservation price (MRP).
The only difference between segments is the number of people in each. That number
represents the maximum willing to buy (MWB). As might be expected, the number of
people in a segment doesn’t affect optimal price as much as the distribution of reserva-
tion prices in that segment. As all three segments here show the same distribution of
reservation prices, they all carry the same optimal price.
Another useful exercise is to consider what would happen if the manufacturer in this
example were able to increase everyone’s reservation price by $1. This would raise the
optimal price by half that amount, or $0.50. Likewise, the optimal price would rise by
half the amount of any increase in variable cost.

OPTIMAL PRICE IN GENERAL
When demand is linear, we have an easy-to-use formula for optimal price. Regardless of
the shape of the demand function, there is a simple relationship between gross margin
and elasticity at the optimal price.


Optimal Price, Relative to Gross Margin: The optimal price is the price at which
a product’s gross margin is equal to the negative of the reciprocal of its elasticity of
demand.5
1
Gross Margin at Optimal Price (%)
Elasticity at Optimal Price

A relationship such as this, which holds at the optimal price, is called an optimality condi-
tion. If elasticity is constant, then we can easily use this optimality condition to determine
the optimal price. We simply find the negative of the reciprocal of the constant elasticity.
The result will be the optimal gross margin. If variable costs are known and constant, then
we need only determine the price that corresponds to the calculated optimal margin.

EXAMPLE: The manager of a stall selling replica sporting goods knows that the
demand for jerseys has a constant price elasticity of 4. To price optimally, she sets her
gross margin equal to the negative of the reciprocal of the elasticity of demand. (Some
economists refer to the price-cost margin as the Lerner Index.)
1
Gross Margin at Optimal Price
4

25%
If the variable cost of each jersey is $5, the optimal price will be $5/(1 0.25), or $6.67.

The optimal margins for several price elasticities are listed in Table 7.10.

Table 7.10 Optimal Margins for Sample Elasticities

Price Elasticity Gross Margin

1.5 67%

2 50%

3 33%

4 25%

Thus, if a firm’s gross margin is 50%, its price will be optimal only if its elasticity at that
price is 2. By contrast, if the firm’s elasticity is 3 at its current price, then its pricing
will be optimal only if it yields a gross margin of 33%.
This relationship between gross margin and price elasticity at the optimal price is one of
the principal reasons that marketers take such a keen interest in the price elasticity


of demand. Price elasticities can be difficult to measure, but margins generally are not.
Marketers might now ask whether their current margins are consistent with estimates of
price elasticity. In the next section, we will explore this issue in greater detail.
In the interim, if elasticity changes with price, marketers can use this optimality condi-
tion to solve for the optimal price. This condition applies to linear demand functions as
well. Because the optimal price formula for linear demand is relatively simple, however,
marketers rarely use the general optimality condition in this instance.

The shortcuts for determining optimal prices from linear and constant elasticity
demand functions rest on an assumption that variable costs hold constant over the
range of volumes considered. If this assumption is not valid, marketers will likely find
that a spreadsheet model will offer the easiest way to determine optimal price.
We have explored these relationships in detail because they offer useful perspectives on
the relationship between margins and the price elasticity of demand. In day-to-day
management, margins constitute a starting point for many analyses, including those of
price. One example of this dynamic would be cost-plus pricing.
Cost-plus pricing has received bad press in the marketing literature. It is portrayed not
only as internally oriented, but also as naïve, in that it may sacrifice profits. From an alter-
nate perspective, however, cost-plus pricing can be viewed as an attempt to maintain mar-
gins. If managers select the correct margin—one that relates to the price elasticity of
demand—then pricing to maintain it may in fact be optimal if demand has constant elas-
ticity. Thus, cost-plus pricing can be more customer-oriented than is widely perceived.

Price Tailoring—a.k.a. Price Discrimination: Marketers have invented a variety of
price discrimination tools, including coupons, rebates, and discounts, for example. All
are designed to exploit variations in price sensitivity among customers. Whenever cus-
tomers have different sensitivities to price, or different costs to serve, the astute marketer
can find an opportunity to claim incremental value through price tailoring.

EXAMPLE: The demand for a particular brand of sunglasses is composed of two seg-
ments: style-focused consumers who are less sensitive to price (more inelastic), and
value-focused consumers who are more sensitive to price (more elastic) (see Figure 7.8).
The style-focused group has a maximum reservation price of $30 and a maximum will-
ing to buy of 10 units. The value-focused group has a maximum reservation price of $10
and a maximum willing to buy of 40 units.


Style Segment

60

50

40
Demand

30

20

10

-

$- $5 $10 $15 $20 $25 $30

Price

Value Segment

45

40

35

30
Demand

25

20

15

10

5

-
$- $5 $10 $15 $20 $25 $30

Price

Figure 7.8 Two Segments Form Demand


ALTERNATIVE A: ONE PRICE FOR BOTH SEGMENTS
Suppose the sunglasses manufacturer plans to offer one price to both segments. Table 7.11
shows the contribution of several candidate prices. The optimal single price (to the near-
est cent) is $6.77, generating a total contribution of $98.56.

Table 7.11 Two Segments: One Price for Both Segments

Single Price Value Quantity Style Quantity Total Demand Total
Demanded Demanded Contribution

$5 20 8.33 28.33 $85.00

$6 16 8.00 24.00 $96.00

$6.77 12.92 7.74 20.66 $98.56

$7 12 7.67 19.67 $98.33

$8 8 7.33 15.33 $92.00

ALTERNATIVE B: PRICE PER SEGMENT
If the manufacturer can find a way to charge each segment its own optimal price, it will
increase total contribution. In Table 7.12, we show the optimal prices, quantities, and
contributions attainable if each segment pays a distinct optimal price.

Table 7.12 Two Segments: Price Tailoring

MRP Variable Optimal Price Quantity Revenue Contribution
Costs

Style $30 $2 $16 4.67 $74.67 $65.33
Value $10 $2 $6 16 $96.00 $64.00
Total 20.67 $170.67 $129.33

These optimal prices were calculated as the midpoints between maximum reservation price
(MRP) and variable cost (VC). Optimal contributions were calculated with the formula

Contribution* (MWB/MRP) * (P* VC)2

In the style-focused segment, for example, this yields
Contribution* (10/30) * ($16 $2)2
(1/3) * (142) $65.33


Thus, through price tailoring, the sunglasses manufacturer can increase total contribu-
tion from $98.56 to $129.33 while holding quantity constant.

Where variable costs differ between segments, as in an airline’s costs of service in busi-
ness class versus economy class, the fundamental calculations are the same. To deter-
mine optimal prices, marketers need only change the variable cost per unit in each
segment to correspond to actual costs.

Caution: Regulation
In most industrial economies, governments have passed regulations concerning price
discrimination. In the United States, the most important of these is the Robinson-
Patman Act. According to Supreme Court interpretations of this statute (as of mid-
2009), Robinson-Patman forbids price discrimination only to the extent that it threatens
to injure competition. There are two main types of injury contemplated by the Act:

1. Primary line competitive injury: Price discrimination might be used as a
predatory tactic. That is, a firm might set prices below cost to certain customers
in order to harm competition at the supplier level. Anti-trust authorities apply
this standard to predatory pricing claims under the Sherman Act and the
Federal Trade Commission Act in order to evaluate allegations of price
discrimination.
2. Secondary line competitive injury: A seller that charges different prices to
competing buyers of the same commodity, or that discriminates in providing
“allowances”—such as compensation for advertising or other services—may be
violating the Robinson-Patman Act. Such discrimination can impair competi-
tion by awarding favored customers an edge that has nothing to do with supe-
rior efficiency.
In the United States, price discrimination is often lawful, particularly if it reflects differ-
ent costs of dealing with diverse buyers, or if it results from a seller’s attempts to meet a
competitor’s prices or services.6 Clearly, this is not intended to be a legal opinion, how-
ever. Legal advice should be sought for a company’s individual circumstances.

7.5 “Own,” “Cross,” and “Residual” Price Elasticity
The concept of residual price elasticity introduces competitive dynamics into the
pricing process. It incorporates competitor reactions and cross elasticity. This, in
turn, helps explain why prices in daily life are rarely set at the optimal level suggested


by a simpler view of elasticity. Marketers consciously or unconsciously factor com-
petitive dynamics into their pricing decisions.
Residual Price Elasticity (I) Own Price Elasticity (I) [Competitor Reaction
Elasticity (I) * Cross Elasticity (I)]
The greater the competitive reaction anticipated, the more residual price elasticity
will differ from a company’s own price elasticity.

Purpose: To account for both customers’ price elasticity and potential
competitive reactions when planning price changes.
Often, in daily life, price elasticity doesn’t quite correspond to the relationships dis-
cussed in the prior section. Managers may find, for example, that their estimates of this
key metric are not equal to the negative of the reciprocal of their margins. Does this
mean they’re setting prices that are not optimal? Perhaps.
It is more likely, however, that they’re including competitive factors in their pricing
decisions. Rather than using elasticity as estimated from current market conditions,
marketers may estimate—or intuit—what elasticity will be after competitors respond
to a proposed change in price. This introduces a new concept, residual price elasticity—
customers’ elasticity of demand in response to a change in price, after accounting for
any increase or decrease in competitors’ prices that may be triggered by the initial
change.
Residual price elasticity is the combination of three factors:

1. “Own” price elasticity—The change in units sold due to the reaction of a firm’s
customers to its changes in price.
2. “Competitor reaction” elasticity—The reaction of competitors to a firm’s price
changes.
3. “Cross” price elasticity—The reaction of a firm’s customers to price changes by
its competitors.
These factors and their interactions are illustrated in Figure 7.9.
Own Price Elasticity: How customers in the market react to our price changes.
Competitive Reaction Elasticity: How our competitors respond to our price changes.
Cross Elasticity: How our customers respond to the price changes of our competitors.
The distinction between own and residual price elasticity is not made clear in the liter-
ature. Some measures of price elasticity, for example, incorporate past competitive reac-
tions and thus are more indicative of residual price elasticity. Others principally reflect


Our Price
Change

E2: Competitor E1: Own Price
Reaction Elasticity Elasticity

Competitor E3: Cross Elasticity Our Volume
Price Change Change

E1 = Own Price Elasticity

E2 = Competitor Reaction Elasticity

E3 = Cross Elasticity

E1 + (E2*E3) = Residual Elasticity

Figure 7.9 Residual Price Elasticity

own price elasticity and require further analysis to determine where sales and income
will ultimately settle. The following sequence of actions and reactions is illustrative:

1. A firm changes price and observes the resulting change in sales. As an alterna-
tive, it may track another measure correlated with sales, such as share of choice
or preference.
2. Competitors observe the firm’s change in price and its increase in sales, and/or
their own decrease in sales.
3. Competitors decide whether and by how much to change their own prices. The
market impact of these changes will depend on (1) the direction and degree of
the changes, and (2) the degree of cross elasticity, that is, the sensitivity of the
initial firm’s sales quantity to changes in competitors’ prices. Thus, after track-
ing the response to its own price change, the initial firm may observe a further
shift in sales as competitors’ price changes take effect in the market.
Due to this dynamic, if a firm measures price elasticity only through customer response
to its initial actions, it will miss an important potential factor: competitive reactions and
their effects on sales. Only monopolists can make pricing decisions without regard
to competitive response. Other firms may neglect or decline to consider competitive


reactions, dismissing such analyses as speculation. But this generates a risk of short-
sightedness and can lead to dangerous surprises. Still other firms may embrace game
theory and seek a Nash Equilibrium to anticipate where prices will ultimately settle. (In
this context, the Nash Equilibrium would be the point at which none of the competitors
in a market have a profit-related incentive to change prices.)
Although a detailed exploration of competitive dynamics is beyond the scope of this
book, we offer a simple framework for residual price elasticity next.

Construction
To calculate residual price elasticity, three inputs are needed:

1. Own price elasticity: The change in a firm’s unit sales, resulting from its initial
price change, assuming that competitors’ prices remain unchanged.
2. Competitor reaction elasticity: The extent and direction of the price changes
that are likely to be made by competitors in response to a firm’s initial price
change. If competitor reaction elasticity is 0.5, for example, then as a firm reduces
its prices by a small percentage, competitors can be expected to reduce their own
prices by half that percentage. If competitor reaction elasticity is 0.5, then as a
firm reduces its prices by a small percentage, competitors will increase their prices
by half that percentage. This is a less common scenario, but it is possible.
3. Cross elasticity with regard to competitor price changes: The percentage and
direction of the change in the initial firm’s sales that will result from a small
percentage change in competitors’ prices. If cross elasticity is 0.25, then a small
percentage increase in competitors’ prices will result in an increase of one-
fourth that percentage in the initial firm’s sales. Note that the sign of cross
elasticity is generally the reverse of the sign of own price elasticity. When
competitors’ prices rise, a firm’s sales will usually increase, and vice versa.
Residual Price Elasticity (I) Own Price Elasticity (I) [Competitor Reaction
Elasticity (I) * Cross Elasticity (I)]
The percentage change in a firm’s sales can be approximated by multiplying its own
price change by its residual price elasticity:
Change in Sales from Residual Elasticity (%) Own Price Change (%) * Residual Price
Elasticity (I)
Forecasts of any change in sales to be generated by a price change thus should take into
account the subsequent competitive price reactions that can be reasonably expected, as
well as the second-order effects of those reactions on the sales of the firm making the ini-
tial change. The net effect of adjusting for such reactions might be to amplify, diminish,
or even reverse the direction of the change in sales that was expected from the initial
price change.


EXAMPLE: A company decides to reduce price by 10% (price change 10%). It
has estimated its own price elasticity to be 2. Ignoring competitive response, the com-
pany would expect a 10% price reduction to yield an approximately 20% increase in sales
( 2 * 10%). (Note: As observed in our earlier discussion of elasticity, projections
based on point elasticity are accurate only for linear demand functions. Because this
example does not specify the shape of the demand function, the projected 20% increase
in sales is an approximation.)
The company estimates competitor reaction elasticity to be 1. That is, in response to the
firm’s action, competitors are expected to shift pricing in the same direction and by an
equal percentage.
The company estimates cross elasticity to be 0.7. That is, a small percentage change in
competitors’ prices will result in a change in the firm’s own sales of 0.7 percent. On this basis,
Residual Elasticity Own Price Elasticity (Competitor Reaction Elasticity
* Cross Elasticity)
2 + (1 * 0.7)
2 + 0.7
1.3
Sales Increase Change in Price * Residual Elasticity
10% * 1.3
13% Increase in Sales
Competitor reactions and cross elasticity are expected to reduce the firm’s initially pro-
jected sales increase from 20% to 13%.

Accounting for potential competitive reactions is important, but there may be simpler
and more reliable methods of managing price strategy in a contested market. Game the-
ory and price leadership principles offer some guidance.
It is important for managers to distinguish between price elasticity measures that are
inherently unable to account for competitive reactions and those that may already
incorporate some competitive dynamics. For example, in “laboratory” investigations of
price sensitivity—such as surveys, simulated test markets, and conjoint analyses—
consumers may be presented with hypothetical pricing scenarios. These can measure
both own price elasticity and the cross elasticities that result from specific combinations
of prices. But an effective test is difficult to achieve.
Econometric analysis of historical data, evaluating the sales and prices of firms in a mar-
ket over longer periods of time (that is, annual or quarterly data), may be better able to
incorporate competitive changes and cross elasticities. To the extent that a firm has


changed price somewhat randomly in the past, and to the extent that competitors have
reacted, the estimates of elasticity that are generated by such analyses will measure resid-
ual elasticity. Still, the challenges and complexities involved in measuring price elasticity
from historical data are daunting.
By contrast, short-term test market experiments are unlikely to yield good estimates of
residual price elasticity. Over short periods, competitors might not learn of price
changes or have time to react. Consequently, elasticity estimates based on test markets
are much closer to own price elasticity.
Less obvious, perhaps, are econometric analyses based on transactional data, such as
scanner sales and short-term price promotions. In these studies, prices decline for a
short time, rise again for a longer period, decline briefly, rise again, and so forth. Even if
competitors conduct their own price promotions during the study period, estimates of
price elasticity derived in this way are likely to be affected by two factors. First, competi-
tors’ reactions likely will not be factored into an elasticity estimate because they won’t
have had time to react to the initial firm’s pricing moves. That is, their actions will have
been largely motivated by their own plans. Second, to the extent that consumers stock
up during price deals, any estimates of price elasticity will be higher than would be
observed over the course of long-term price changes.

Prisoner’s Dilemma Pricing
Prisoner’s dilemma pricing describes a situation in which the pursuit of self-interest by
all parties leads to sub-optimal outcomes for all. This phenomenon can lead to stability
at prices above the expected optimal price. In many ways, these higher-than-optimal
prices have the appearance of cartel pricing. But they can be achieved without explicit
collusion, provided that all parties understand the dynamics, as well as their competi-
tors’ motivations and economics.
The prisoner’s dilemma phenomenon derives its name from a story illustrating the con-
cept. Two members of a criminal gang are arrested and imprisoned. Each prisoner is
placed in solitary confinement, with no means of speaking to the other. Because the police
don’t have enough evidence to convict the pair on the principal charge, they plan to sen-
tence both to a year in prison on a lesser charge. First, however, they try to get one or both
to confess. Simultaneously, they offer each prisoner a Faustian bargain. If the prisoner tes-
tifies against his partner, he will go free, while the partner is sentenced to three years in
prison on the main charge. But there’s a catch . . . If both prisoners testify against each
other, both will be sentenced to two years in jail.7 On this basis, each prisoner reasons that
he’ll do best by testifying against his partner, regardless of what the partner does.
For a summary of the choices and outcomes in this dilemma, please see Figure 7.10,
which is drawn in the first person from the perspective of one of the prisoners. First-
person outcomes are listed in bold. Partner outcomes are italicized.


My 3 years 1 year
partner
refuses
to I go free 1 year
testify

My 2 years My partner
partner goes free
testifies
2 years 3 years

I testify I refuse to testify

Figure 7.10 Prisoner’s Dilemma Pay-off Grid

Continuing the first-person perspective, each prisoner reasons as follows: If my partner
testifies, I’ll be sentenced to two years in prison if I testify as well, or three years if I don’t.
On the other hand, if my partner refuses to testify, I’ll go free if I testify, but serve one
year in prison if I don’t. In either case, I do better if I testify. But this raises a dilemma.
If I follow this logic and testify—and my partner does the same—we end up in the
lower-left cell of the table, serving two years in prison.
Figure 7.11 uses arrows to track these preferences—a dark arrow for the first-person
narrator in this reasoning, and a light arrow for his partner.
The dilemma, of course, is that it seems perfectly logical to follow the arrows and testify.
But when both prisoners do so, they both end up worse off than they would have if
they’d both refused. That is, when both testify, both are sentenced to two years in prison.
If both had refused, they both could have shortened that term to a single year.

My 3 years 1 year
Partner
Refuses
I go free 1 year

My 2 years My partner
Partner goes free
Testifies
2 years 3 years

I testify I refuse

Figure 7.11 Pay-off Grid with Arrows Representing Preferences for Prisoners


Admittedly, it takes a good deal of time to grasp the mechanics of the prisoner’s dilemma,
and far longer to appreciate its implications. But the story serves as a powerful metaphor,
encapsulating a wide range of situations in which acting in one’s own best interest leads
to outcomes in which everyone is worse off.
In pricing, there are many situations in which a firm and its competitors face a pris-
oner’s dilemma. Often, one firm perceives that it could increase profits by reducing
prices, regardless of competitors’ pricing policies. Simultaneously, its competitors per-
ceive the same forces at work. That is, they too could earn more by cutting prices,
regardless of the initial firm’s actions. If both the initial firm and its competitors reduce
prices, however—that is, if all parties follow their own unilateral best interests—they
will, in many situations, all end up worse off. The industry challenge in these situations
is to keep prices high despite the fact that each firm will benefit by lowering them.
Given a choice between high and low prices a firm faces a prisoner’s dilemma pricing
situation when the following conditions apply:

1. Its contribution is greater at the low price when selling against both high and
low competitor prices.
2. Competitors’ contribution is greater at their low price when selling against both
the high and low prices of the initial firm.
3. For both the initial firm and its competitors, however, contribution is lower
if all parties set their price low than it would have been if all parties had
priced high.

EXAMPLE: As shown in Table 7.13, my firm faces one main competitor. Currently my
price is $2.90, their price is $2.80, and I hold a 40% share of a market that totals 20 mil-
lion units. If I reduce my price to $2.60, I expect my share will rise to 55%—unless, of
course, they also cut their price. If they also reduce price by $0.30—to $2.50—then I
expect our market shares to remain constant at 40/60. On the other hand, if my competi-
tor cuts its price but I hold steady at $2.90, then I expect they’ll increase their market
share to 80%, leaving me with only 20%.
If we both have variable costs of $1.20 per unit, and market size remains constant at 20
million units, we face four possible scenarios with eight contribution figures—four for
my firm and four for the competition:


Table 7.13 Scenario Planning Pay-off Table

My My
Pricing My Volume My Sales Variable Contribution
Scenario My Price (m) ($m) Costs ($m) ($m)
My Firm $2.90 8 $23.2 $9.6 $13.6
High.
Competition
High.
My Firm $2.90 4 $11.6 $4.8 $6.8
High.
Competition
Low.
My Firm $2.60 8 $20.8 $9.6 $11.2
Low.
Competition
Low.
My Firm $2.60 11 $28.6 $13.2 $15.4
Low.
Competition
High.

Their Their
Pricing Their Their Sales Variable Contribution
Scenario Their Price Volume (m) ($m) Costs ($m) ($m)
My Firm $2.80 12 $33.6 $14.4 $19.2
High.
Competition
High.
My Firm $2.50 16 $40.0 $19.2 $20.8
High.
Competition
Low.
My Firm $2.50 12 $30.0 $14.4 $15.6
Low.
Competition
Low.
My Firm $2.80 9 $25.2 $10.8 $14.4
Low.
Competition
High.


Are we in a prisoner’s dilemma situation?
Figure 7.12 shows the four contribution possibilities for both my firm and my competitor.

Their
Price $14.4 $19.2
= $2.80
$15.4 $13.6
High

Their
Price $15.6 $20.8
= $2.50
$11.2 $6.8
Low

My Price = $2.60 My Price = $2.90
Low High

Figure 7.12 Pay-off Grid with Expected Values (Values Are in the Millions of Dollars)

Let’s check to see whether the conditions for the prisoner’s dilemma are met:
1. My contribution is higher at the low price for both high and low competitor prices
($15.4m > $13.6m, and $11.2m > $6.8m). No matter what my competitor does, I
make more money at the low price.
2. My competitor’s contribution is higher at the low price, regardless of my price
($15.6m > $14.4m, and $20.8m > $19.2m). They, too, are better off at the low price,
regardless of my price.
3. For both my firm and my competitor, however, contribution is lower if we both price
low than it would be if we both price high ($15.6m < $19.2m, and $11.2m < $13.6m).
The conditions for the prisoner’s dilemma are met (see Figure 7.13).

Their
Price $14.4 $19.2
= $2.80
$15.4 $13.6

Their
Price $15.6 $20.8
= $2.50
$11.2 $6.8

My Price = $2.60 My Price = $2.90

Figure 7.13 Pay-off Grid with Expected Values and Preference Arrows (Values Are in the
Millions of Dollars)


The implication for my firm is clear: Although it is tempting to lower my price, seeking
increased share and a $15.4 million contribution, I must recognize that my competitor
faces the same incentives. They, too, have an incentive to cut price, grab share, and
increase their contribution. But if they lower their price, I’ll probably lower mine. If I
lower my price, they’ll probably lower theirs. If we both reduce our prices, I’ll earn only
$11.2m in contribution—a sharp decline from the $13.6m I make now.

Managerial Note: To determine whether you face a prisoner’s dilemma situation, proj-
ect the dollar contributions for both your firm and your competition at four combina-
tions of high and low prices. Projections may require assumptions about your
competitors’ economics. These, in turn, will require care. If competitors’ economics dif-
fer greatly from your projections, they may not face the decisions or motivations
ascribed to them in your model. Additionally, there are a number of reasons why the
logic of the prisoner’s dilemma won’t always hold, even if all assumptions are correct.

1. Contribution may not be the sole criterion in decision-making: In our exam-
ple, we used contribution as the objective for both firms. Market share,
however, may have importance to one or more firms, above and beyond its
immediate, direct effect on contribution. Whatever a firm’s objective may be, if
it is quantifiable, we can place it in our table to better understand the competi-
tive situation.
2. Legal issues: Certain activities designed to discourage competition and main-
tain high prices are illegal. Our purpose here is to help managers understand
the economic trade-offs involved in competitive pricing. Managers should be
aware of their legal environment and behave accordingly.
3. Multiple competitors: Pricing becomes more complicated when there are mul-
tiple competitors. The test for a multi-party prisoner’s dilemma is the logical
extension of the test described earlier. A major difference, however, arises in
practice. As a general principle, the greater the number of independent com-
petitors, the more difficult it will be to keep prices high.
4. Single versus repeated play: In our original story, two prisoners decide whether
to testify in a single investigation. In game theory terms, they play the game a
single time. Experiments have shown that in a single play of a prisoner’s
dilemma, the likely outcome is that both prisoners will testify. If the game is
played repeatedly, however, it is more likely that both prisoners will refuse to
testify. Because pricing decisions are made repeatedly, this evidence suggests
that high prices are a more likely outcome. Most businesses eventually learn to
live with their competition.
5. More than two possible prices: We have examined a situation in which each
player considers two prices. In reality, there may be a wide range of prices under


consideration. In such situations, we might extend our analysis to more boxes.
Once again, we might add arrows to track preferences. Using these more com-
plex views, one sometimes finds areas within the table in which a prisoner’s
dilemma applies (usually at the higher prices), and others where it does not
(usually at the lower prices). One might also find that the arrows lead to a partic-
ular cell in the middle of the table called the equilibrium. A prisoner’s dilemma
situation generally applies for prices higher than the set of equilibrium prices.
Applying the lessons of the prisoner’s dilemma, we see that optimal price calculations
based on own price elasticity may lead us to act in our own unilateral best interest. By
contrast, when we factor residual price elasticity into our calculations, competitive
response becomes a key element of our pricing strategy. As the prisoner’s dilemma
shows, over the long term, a firm is not always best served by acting in its apparent uni-
lateral best interest.

Dolan, Robert J., and Hermann Simon. (1996). Power Pricing: How Managing Price
Transforms the Bottom Line, New York: Free Press, 4.
Roegner, E.V., M.V. Marn, and C.C. Zawada. (2005). “Pricing,” Marketing Management,
14(1), 23–28.


8
PROMOTION

Introduction

Baseline Sales, Incremental Sales, Percent Sales on Deal, Percent Time
and Promotional Lift on Deal, and Average Deal Depth
Redemption Rates for Pass-Through and Price
Coupons/Rebates Waterfall

Price promotions can be divided into two broad categories:

■ Temporary price reductions.
■ Permanent features of pricing systems.1
With both of these, firms seek to change the behavior of consumers and trade customers
in ways that increase sales and profits over time, though a promotion’s short-term effect
on profits will often be negative. There are multiple routes to sales and profit growth
and many potential reasons for offering price promotions. Such programs might be
aimed at affecting the behavior of end users (consumers), trade customers (distributors
or retailers), competitors, or even a firm’s own salespeople. Although the goal of a
promotion is often to increase sales, these programs can also affect costs. Examples of
specific, short-term promotional objectives include the following:
■ To acquire new customers, perhaps by generating trial.
■ To appeal to new or different segments that are more price-sensitive than a
firm’s traditional customers.

263

■ To increase the purchase rates of existing customers; to increase loyalty.
■ To gain new trade accounts (that is, distribution).
■ To introduce new SKUs to the trade.
■ To increase shelf space.
■ To blunt competitive efforts by encouraging the firm’s customers to “load up”
on inventory.
■ To smooth production in seasonal categories by inducing customers to order
earlier (or later) than they ordinarily would.
The metrics for many of these interim objectives, including trial rate and percentage of
new product sales, are covered elsewhere. In this chapter, we focus on metrics for mon-
itoring the acceptance of price promotions and their effects on sales and profits.
The most powerful framework for evaluating temporary price promotions is to parti-
tion sales into two categories: baseline and incremental. Baseline sales are those that a
firm would have expected to achieve if no promotion had been run. Incremental sales
represent the “lift” in sales resulting from a price promotion. By separating baseline sales
from incremental lift, managers can evaluate whether the sales increase generated by
a temporary price reduction compensates for the concomitant decrease in prices and
margins. Similar techniques are used in determining the profitability of coupons
and rebates.
Although the short-term effect of a price promotion is almost invariably measured by
its increase in sales, over longer periods management becomes concerned about the per-
centage of sales on deal and the percentage of time during which a product is on deal.
In some industries, list price has become such a fiction that it is used only as a bench-
mark for discussing discounts.
Average deal depth and the price waterfall help capture the depth of price cuts and
explain how one arrives at a product’s net price (pocket price) after accounting for all
discounts. There are often major differences between the discounts offered to trade
customers and the extent to which those discounts are accepted. There may also be a
difference between the discounts received by the trade and those that the trade shares
with its customers. The pass-through percentage and price waterfall are analytic struc-
tures designed to capture those dynamics and thus to measure the impact of a
firm’s promotions.



8.1 Baseline Sales Intercept in Marketing To determine the
regression of sales activities also extent to which
as function of contribute to current sales are
marketing vari- baseline. independent of
ables. Baseline specific marketing
Sales Total efforts.
Sales, less incre-
mental sales
generated by a
marketing
program or
programs.
8.1 Incremental Sales, Total sales, less Need to consider To determine
or Promotional baseline sales. competitive short-term effects
Lift Regression coeffi- actions. of marketing
cient to market- effort.
ing variables
cited above.
8.2 Redemption Rates Coupons Will differ signifi- Rough measure of
redeemed divided cantly by mode coupon “lift” after
by coupons of coupon adjusting for sales
distributed. distribution. that would have
been made with-
out coupons.
8.2 Costs for Coupon face Does not consider Allows for budg-
Coupons and amount plus margins that eting of coupon
Rebates redemption would have been expense.
charges, multi- generated by
plied by the num- those willing to
ber of coupons buy product
redeemed. without coupon.
8.2 Percentage Sales Sales via coupon, Doesn’t factor in A measure of
with Coupon divided by total magnitude of dis- brand depend-
sales. count offered by ence on promo-
specific coupons. tional efforts.
Continues

Chapter 8 Promotion 265


8.3 Percent Sales on Sales with tempo- Does not make A measure of
Deal rary discounts as distinction for brand depend-
a percentage of depth of dis- ence on promo-
total sales. counts offered. tional efforts.
8.3 Pass-Through Promotional Can reflect power To measure the
discounts provid- in the channel, extent to which a
ed by the trade or deliberate manufacturer’s
to consumers, management or promotions
divided by segmentation. generate promo-
discounts tional activity
provided to the further along
trade by the the distribution
manufacturer. channel.
8.4 Price Waterfall Actual average Some discounts To indicate the
price per unit may be offered at price actually paid
divided by list an absolute level, for a product, and
price per unit. not on a per-item the sequence of
Can also be calcu- basis. channel factors
lated by working affecting that
backward from price.
list price, taking
account of poten-
tial discounts,
weighted by the
frequency with
which each is
exercised.


8.1 Baseline Sales, Incremental Sales,
and Promotional Lift
Estimates of baseline sales establish a benchmark for evaluating the incremental sales
generated by specific marketing activities. This baseline also helps isolate incremental
sales from the effects of other influences, such as seasonality or competitive promo-
tions. The following equations can be applied for defined periods of time and for the
specific element of the marketing mix that is used to generate incremental sales.
Total Sales ($,#) Baseline Sales ($,#) Incremental Sales from Marketing ($,#)
Incremental Sales from Marketing ($,#) Incremental Sales from Advertising ($,#)
Incremental Sales from Trade
Promotion ($,#)
Incremental Sales from Consumer
Promotion ($,#)
Incremental Sales from Other ($,#)

Incremental Sales ($,#)
Lift (from Promotion) (%)
Baseline Sales ($,#)

Cost of Incremental Sales ($)

The justification of marketing spending almost always involves estimating the incre-
mental effects of the program under evaluation. However, because some marketing
costs are often assumed to be fixed (for example, marketing staff and sales force
salaries), one rarely sees incremental sales attributed to these elements of the mix.

Purpose: To select a baseline of sales against which the incremental sales
and profits generated by marketing activity can be assessed.
A common problem in marketing is estimating the sales “lift” attributable to a specific
campaign or set of marketing activities. Evaluating lift entails making a comparison with
baseline sales, the level of sales that would have been achieved without the program
under evaluation. Ideally, experiments or “control” groups would be used to establish
baselines. If it were quick, easy, and inexpensive to conduct such experiments, this
approach would dominate. In lieu of such control groups, marketers often use historical
sales adjusted for expected growth, taking care to control for seasonal influences.
Regression models that attempt to control for the influence of these other changes
are often used to improve estimates of baseline sales. Ideally, both controllable and


uncontrollable factors, such as competitive spending, should be included in baseline
sales regression models. When regression is used, the intercept is often considered to be
the baseline.

Construction
In theory, determining incremental sales is as simple as subtracting baseline sales from
total sales. Challenges arise, however, in determining baseline sales.
Baseline Sales: Expected sales results, excluding the marketing programs under
evaluation.
In reviewing historical data, total sales are known. The analyst’s task then is to sepa-
rate these into baseline sales and incremental sales. This is typically done with regres-
sion analysis. The process can also involve test market results and other market
research data.
Total Sales ($,#) Baseline Sales ($,#) Incremental Sales ($,#)
Analysts also commonly separate incremental sales into portions attributable to the
various marketing activities used to generate them.
Incremental Sales ($,#) Incremental Sales from Advertising ($,#) Incremental Sales from
Trade Promotion ($,#) Incremental Sales from Consumer
Promotion ($,#) Incremental Sales from Other ($,#)
Baseline sales are generally estimated through analyses of historical data. Firms often
develop sophisticated models for this purpose, including variables to adjust for market
growth, competitive activity, and seasonality, for example. That done, a firm can use its
model to make forward-looking projections of baseline sales and use these to estimate
incremental sales.
Incremental sales can be calculated as total sales, less baseline sales, for any period of
time (for example, a year, a quarter, or the term of a promotion). The lift achieved by a
marketing program measures incremental sales as a percentage of baseline sales. The
cost of incremental sales can be expressed as a cost per incremental sales dollar or a cost
per incremental sales unit (for example, cost per incremental case).
Incremental Sales ($,#) Total Sales ($,#) Baseline Sales ($,#)
Lift (%)
Baseline Sales ($,#)


EXAMPLE: A retailer expects to sell $24,000 worth of light bulbs in a typical month
without advertising. In May, while running a newspaper ad campaign that cost $1,500,
the store sells $30,000 worth of light bulbs. It engages in no other promotions or non-
recurring events during the month. Its owner calculates incremental sales generated by
the ad campaign as follows:
Incremental Sales ($) Total Sales ($) Baseline Sales ($)
$30,000 $24,000 = $6,000
The store owner estimates incremental sales to be $6,000. This represents a lift (%) of
25%, calculated as follows:
Incremental Sales ($)
Lift (%)
Baseline Sales ($)
$6,000
25%.
$24,000

The cost per incremental sales is $0.25, calculated as follows:
Incremental Sales ($)
$1,500
0.25
$6,000

Total sales can be analyzed or projected as a function of baseline sales and lift. When
estimating combined marketing mix effects, one must be sure to determine whether lift
is estimated through a multiplicative or an additive equation. Additive equations com-
bine marketing mix effects as follows:
Total Sales ($,#) Baseline Sales [Baseline Sales ($,#) * Lift (%) from Advertising]
[Baseline Sales ($,#) * Lift (%) from Trade Promotion]
[Baseline Sales ($,#) * Lift (%) from Consumer Promotion]
[Baseline Sales ($,#) * Lift (%) from Other]
This additive approach is consistent with the conception of total incremental sales as a
sum of the incremental sales generated by various elements of the marketing mix. It is
equivalent to a statement that
Total Sales ($,#) Baseline Sales Incremental Sales from Advertising Incremental Sales
from Trade Promotion Incremental Sales from Consumer Promotion
Incremental Sales from Other


Multiplicative equations, by contrast, combine marketing mix effects by using a multi-
plication procedure, as follows:
Total Sales ($,#) Baseline Sales ($,#) * (1 Lift (%) from Advertising) * (1 Lift (%) from
Trade Promotion) * (1 Lift (%) from Consumer Promotion) * (1 Lift
(%) from Other)
When using multiplicative equations, it makes little sense to talk about the incremental
sales from a single mix element. In practice, however, one may encounter statements
that attempt to do exactly that.

EXAMPLE: Company A collects data from past promotions and estimates the lift it
achieves through different elements of the marketing mix. One researcher believes that
an additive model would best capture these effects. A second researcher believes that a
multiplicative model might better reveal the ways in which multiple elements of the mix
combine to increase sales. The product manager for the item under study receives the two
estimates shown in Table 8.1.

Table 8.1 Expected Returns to Marketing Spending

Additive Multiplicative

Trade Consumer Trade Consumer
Advertising Promotion Promotion Advertising Promotion Promotion
Spending Lift Lift Lift Lift Lift Lift
$0 0% 0% 0% 1 1 1
$100k 5.5% 10% 16.5% 1.05 1.1 1.15

$200k 12% 24% 36% 1.1 1.2 1.3

Fortunately, both models estimate baseline sales to be $900,000. The product manager
wants to evaluate the following spending plan: advertising ($100,000), trade promotion
($0), and consumer promotion ($200,000). He projects sales using each method as follows:
Additive:
Projected Sales ($) $900,000 [$900,000 * 5.5%] [$900,000 * 0] [$900,000 * 36%]
$900,000 $49,500 $0 $324,000
$1,273,500


Multiplicative:
Projected Sales Baseline * Advertising Lift * Trade Promotion Lift
* Consumer Promotion Lift
= $900,000 * 1.05 * 1 * 1.3
= $1,228,500
Note: Because these models are constructed differently, they will inevitably yield different
results at most levels. The multiplicative method accounts for a specific form of interac-
tions between marketing variables. The additive method, in its current form, does not
account for interactions.

When historic sales have been separated into baseline and incremental components, it
is relatively simple to determine whether a given promotion was profitable during
the period under study. Looking forward, the profitability of a proposed marketing
activity can be assessed by comparing projected levels of profitability with and without
the program:
Profitability of a Promotion ($) Profits Achieved with Promotion ($)
Estimated Profits without Promotion
(that is, Baseline) ($)2

EXAMPLE: Fred, the VP of Marketing, and Jeanne, the VP of Finance, receive esti-
mates that sales will total 30,000 units after erecting special displays. Because the pro-
posed promotion involves a considerable investment ($100,000), the CEO asks for an
estimate of the incremental profit associated with the displays. Because this program
involves no change in price, contribution per unit during the promotion is expected to be
the same as at other times, $12.00 per unit. Thus, total contribution during the promo-
tion is expected to be 30,000 * $12, or $360,000. Subtracting the incremental fixed cost of
specialized displays, profits for the period are projected to be $360,000 $100,000, or
$260,000.
Fred estimates that baseline sales total 15,000 units. On this basis, he calculates that con-
tribution without the promotion would be $12 * 15,000 $180,000. Thus, he projects
that the special displays can be expected to generate incremental profit of $360,000
$180,000 $100,000 $80,000.
Jeanne argues that she would expect sales of 25,000 units without the promotion, gener-
ating baseline contribution of $12 * 25,000 $300,000. Consequently, if the promotion
is implemented, she anticipates an incremental decline in profits from $300,000 to
$260,000. In her view, the promotion’s lift would not be sufficient to cover its incremen-
tal fixed costs. Under this promotion, Jeanne believes that the firm would be spending


$100,000 to generate incremental contribution of only $60,000 (that is, 5,000 units * $12
contribution per unit).
The baseline sales estimate is a crucial factor here.

EXAMPLE: A luggage manufacturer faces a difficult decision regarding whether to
launch a new promotion. The firm’s data show a major increase in product sales in
November and December, but its managers are unsure whether this is a permanent trend
of higher sales or merely a blip—a successful period that can’t be expected to continue
(see Figure 8.1).

Sales

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure 8.1 Monthly Sales Patterns

The firm’s VP of Marketing strongly supports the proposed promotion. He argues that
the increased volume can’t be expected to continue and that the firm’s historic baseline
(26,028 units) should be used as the level of sales that can be anticipated without the pro-
motion. In addition, the Marketing VP argues that only the variable cost of each sale
should be considered. “After all, the fixed costs will be with us whatever we do,” he says.
On this basis, the relevant cost per unit subject to analysis would be $25.76.
The CEO hires a consultant who has a very different opinion. In the consultant’s view,
the November-December sales increase was more than a blip. The market has grown, she
says, and the strength of the firm’s brand has grown with it. Consequently, a more appro-
priate estimate of baseline sales would be 48,960 units. The consultant also points out
that in the long term, no costs are fixed. Therefore, for purposes of analysis, fixed costs
should be allocated to the cost of the product because the product must ultimately gen-
erate a return after such expenses as factory rent are paid. On this basis, the full cost of
each unit, $34.70, should be used as the cost of incremental sales (see Table 8.2).


Table 8.2 Baseline Matters When Considering Profitability

Consultant VP Marketing
Promotion Baseline Promotion Baseline

Price $41.60 $48.00 $41.60 $48.00
Cost $34.70 $34.70 $25.76 $25.76

Margin $6.90 $13.30 $15.84 $22.24

Sales 75,174 48,960 75,174 26,028

Profit $518,701 $651,168 $1,190,756 $578,863

Profitability of
Promotion ($132,467) $611,893

The Marketing VP and the consultant make very different projections of the profitability
of the promotion. Once again, the choice of the baseline matters. Also, we can see that
establishing a shared understanding of costs and margins can be critical.

Finding a baseline estimate of what a company can be expected to sell, “all things being
equal,” is a complex and inexact process. Essentially, the baseline is the level of sales that
can be expected without significant marketing activities. When certain marketing activ-
ities, such as price promotions, have been employed for several periods, it can be espe-
cially difficult to separate “incremental” and “baseline” sales.
In many companies, it is common to measure sales performance against historic data. In
effect, this sets historic sales as the baseline level for analysis of the impact of marketing
spending. For example, retailers can evaluate their performance on the basis of same
store sales (to remove differences caused by the addition or removal of outlets). Further,
they can compare each current period to the same period in the prior year, in order to
avoid seasonality biases and to ensure that they measure periods of special activity (such
as sales events) against times of similar activity.
It is also common practice to adjust the profitability of promotions for longer-term
effects. These effects can include a decline in sales levels in periods immediately following
a promotion, as well as higher or lower sales in related product categories that are associ-
ated with a promotion. Adjustments can be negative or positive. Additional long-term


effects, such as obtaining trial by new consumers, gaining distribution with trade
customers, and increased consumption rates were discussed briefly in the chapter
introduction.

LONG-TERM EFFECTS OF PROMOTIONS
Over time, the effects of promotions may be to “ratchet” sales up or down (see Figures 8.2
and 8.3). Under one scenario, in response to one firm’s promotions, competitors may
also increase their promotional activity, and consumers and trade customers in the field
may learn to wait for deals, increasing sales for no one (see the prisoner’s dilemma in
Section 7.5).

Profits

Promotion
Aggressive
Promotion More
Aggressive
Competitors Promotion
react Competitors Customers
react Competitors learn: wait
react for deals

BASELINE

Time

Figure 8.2 Downward Spiral—Promotional Effectiveness

Profits
Promotion 3
Promotion 2 More customers
Promotion 1 learn to love the
product
Customers
become loyal
Trade stocks
successful
Customers product
try product

BASELINE

Time

Figure 8.3 Successful Promotion with Long-Term Benefits


Under a different, more heartening scenario, promotions can generate trial for new
products, build trade distribution, and encourage loyalty, thus raising the long-term
level of baseline sales.

8.2 Redemption Rates, Costs for Coupons and Rebates,
Percent Sales with Coupon
Redemption rate is the percentage of distributed coupons or rebates that are used
(redeemed) by consumers.
Coupons Redeemed (#)
Coupon Redemption Rate (%)
Coupons Distributed (#)

Cost per Redemption ($) Coupon Face Amount ($) Redemption Charges ($)
Total Coupon Cost ($) [Cost per Redemption ($) * Coupons Redeemed (#)]
Coupon Printing and Distribution Cost ($)
Sales with Coupon ($)
Percentage Sales with Coupon (%)
Sales ($)
The redemption rate is an important metric for marketers assessing the effectiveness
of their coupon distribution strategy. It helps determine whether coupons are reach-
ing the customers who are motivated to use them. Similar metrics apply to mail-in
rebates.
Cost per redemption ($) measures variable costs per coupon redeemed. Coupon
distribution costs are usually viewed as fixed costs.

Purpose: To track and evaluate coupon usage.
Some people hate coupons. Some like them. And some say they hate coupons, but really
like them. Businesses often say they hate coupons but continue to use them. Coupons
and rebates are used to introduce new products, to generate trial of existing products by
new customers, and to “load” consumers’ pantries, encouraging long-term consumption.
Almost all of the interim objectives discussed in the introduction to this chapter can apply
to coupons and rebates. Coupons can be used to offer lower prices to more price-sensitive
consumers. Coupons also serve as a form of advertising, making them dual-purpose mar-
keting vehicles. Coupon clippers will see a brand name and pay closer attention to it—
considering whether they desire the product—than would an average consumer exposed
to an advertisement without a compelling offer. Finally, both rebates and coupons can
serve as focus points for retailer promotions. To generate traffic, retailers can double or
even triple coupon amounts—generally up to a declared limit. Retailers also often adver-
tise prices “after rebates” in order to promote sales and perceptions of value.


Construction
Coupon Redemption Rate (%)
Coupons Distributed (#)

Cost per Redemption ($) Coupon Face Amount ($) Redemption Charges ($)

Total Coupon Cost: Reflects distribution, printing,3 and redemption costs to esti-
mate the total cost of a coupon promotion.
Total Coupon Cost ($) [Coupons Redeemed (#) * Cost per Redemption ($)]
Coupon Printing and Distribution Cost ($)

Total Coupon Cost ($)
Total Cost per Redemption ($)

Sales with Coupon ($,#)
Percentage Sales with Coupon (%)
Sales ($,#)

To determine the profitability of coupons and rebates, managers require approaches
similar to those used in estimating baseline and incremental sales, as discussed in the
previous section of this chapter. By themselves, redemption rates are not a good measure
of success. Under certain circumstances, even low redemption rates can be profitable.
Under other circumstances, by contrast, high redemption rates can be quite damaging.

EXAMPLE: Yvette is the Manager of Analysis for a small regional consumer packaged
goods firm. Her product has a dominant share of the retail distribution in a narrow geo-
graphic area. Her firm decides to launch a coupon campaign, and Yvette is charged with
reporting on the program’s success. Her assistant looks at the figures and realizes that of
the 100,000 coupons distributed in the local paper, 5,000 were used to buy product. The
assistant is excited when he calculates that this represents a 5% redemption rate—a much
higher figure than the company has ever previously seen.
Yvette, however, is more cautious in judging the promotion a success. She checks the
sales of the relevant product and learns that these increased by only 100 units during the
promotion period. Yvette concludes that the vast majority of coupon use was by cus-
tomers who would have bought the product anyway. For most customers, the sole impact
of the coupon was to reduce the price of the product below the level they would have
willingly paid. Until she conducts a full profitability analysis, evaluating the profit gener-
ated by the 100 incremental sales and comparing this to coupon costs and the value lost
on most coupon sales, Yvette can’t be sure that the program made an overall loss. But she
feels certain that she should curtail the celebrations.


To calculate coupon redemption rates, managers must know the number of coupons
placed in circulation (distributed) as well as the number redeemed. Companies general-
ly engage distribution services or media companies to place coupons in circulation.
Redemption numbers are usually derived from the invoices presented by coupon
clearinghouses.

MAIL-IN REBATES
The rebate, in effect, is a form of coupon that is popular with big-ticket items. Its usage
dynamics are straightforward: Customers pay the full price for a product, enabling
retailers to meet a specific price point. The customer then exercises the rebate and
receives back a specified dollar amount.
By using rebates, marketers gain information about customers, which can be useful
in remarketing and product control. Mail-in rebates also reduce the effective price of
an item for customers who are sufficiently price-conscious to take advantage of
them. Others pay full price. The “non-redemption rates” for rebates are sometimes
called “breakage.”
Breakage: The number of rebates not redeemed by customers. The breakage rate is
the percentage of rebates not redeemed.

EXAMPLE: A cell phone company sold 40,000 handsets in one month. On each pur-
chase, the customer was offered a $30 rebate. Thirty thousand rebates were successfully
claimed.
In volume terms, the rebate redemption rate can be calculated by dividing the number of
rebates successfully claimed (30,000) by number offered (40,000):
30,000
Redemption Rate (in volume terms) 75%
40,000

Managers often balk at the cost of distributing coupons. Because promotions rely on
adequate distribution, however, it is inadvisable to create arbitrary cutoffs for distribu-
tion costs. The total cost of incremental sales generated would represent a better metric
to evaluate coupon efficiency—and thus to determine the point at which diminishing
returns make further coupon distribution unattractive.


In evaluating a coupon or rebate program, companies should also consider the overall
level of benefit provided to consumers. Retailers commonly increase the value of
coupons, offering customers a discount of double or even triple the coupons’ face value.
This enables retailers to identify price-sensitive customers and offer them additional
savings. Of course, by multiplying the savings afforded consumers, the practice of dou-
bling or tripling coupons undoubtedly raises some redemption rates.

8.3 Promotions and Pass-Through
Of the promotional value provided by a manufacturer to its retailers and distribu-
tors, the pass-through percentage represents the portion that ultimately reaches
the consumer.
Sales with any Temporary Discount ($,#)
Percentage Sales on Deal (%)
Total Sales ($,#)

Value of Temporary Promotional Discounts
Provided to Consumers by the Trade ($)
Pass-Through (%)
Value of Temporary Discounts Provided to Trade
by Manufacturer ($)
Manufacturers offer many discounts to their distributors and retailers (often called
“the trade”) with the objective of encouraging them to offer their own promotions,
in turn, to their customers. If trade customers or consumers do not find promotions
attractive, this will be indicated by a decline in percentage sales on deal. Likewise, low
pass-through percentages can indicate that too many deals—or the wrong kinds of
deals—are being offered.

Purpose: To measure whether trade promotions are generating
consumer promotions.
Pass-Through: The percentage of the value of manufacturer promotions paid to
distributors and retailers that is reflected in discounts provided by the trade to their
own customers.
“Middlemen” are a part of the channel structure in many industries. Companies may
face one, two, three, or even four levels of “resellers” before their product reaches the
ultimate consumer. For example, a beer manufacturer may sell to an exporter, who sells
to an importer, who sells to a local distributor, who sells to a retail store. If each channel
adds its own margin, without regard for how others are pricing, the resulting price can
be higher than a marketer would like. This sequential application of individual margins
has been referred to as “double marginalization.”4


Construction
Percentage Sales on Deal: Measures the percentage of company sales that are sold
with a temporary trade discount of some form. Note: This usually would not include
standard discounts such as those for early payment or cooperative advertising
allowances (accruals).
Sales with Any Temporary Discount (#,$)
Percentage Sales on Deal (%)
Total Sales (#,$)
Promotional discount represents the total value of promotional discounts given
throughout the sales channel.
Promotional Discount ($) Sales with Any Temporary Discount ($)
* Average Depth of Discount As Percent of List (%)
Unit Discount ($)
Depth of Discount As Percent of List
Unit List Price ($)
Pass-through is calculated as the value of discounts given by the trade to their customers,
divided by the value of temporary discounts provided by a manufacturer to the trade.
Promotional Discounts Provided by the Trade to Consumers ($)
Pass-Through (%)
Discounts Provided to Trade by Manufacturer ($)

Manufacturers often compete with one another for the attention of retailers, distribu-
tors, and other resellers. Toward that end, they build special displays for their products,
change assortments to include new offerings, and seek to elicit increasing attention from
resellers’ sales personnel. Significantly, in their effort to increase channel “push,” manu-
facturers also offer discounts and allowances to the trade. It is important to understand
the rates and amounts of discounts provided to the trade, as well as the proportions of
those discounts that are passed along to the resellers’ customers. At times, when
resellers’ margins are thin, manufacturers’ discounts are designed to enhance them.
Market leaders often worry that trade margins are too thin to support push efforts.
Other manufacturers may be concerned that retail margins are too high, and that too
few of their discounts are being passed along. The metrics discussed in this chapter
should be interpreted with these thoughts in mind.
Resellers may decide that optimizing an entire product line is more important than
maximizing profits on any given product. If a reseller stocks multiple competing lines, it
can be difficult to find an overall solution that suits both that reseller and its suppliers.
Manufacturers strive to motivate resellers to market their goods aggressively and to
grow their shared sales through such programs as incentives for “exclusivity,” or rebates
based on increasing shares of category sales or on year-to-year growth in sales.


Resellers learn to adapt their buying and selling practices to take advantage of manufac-
turer pricing incentives. In this area, marketers must pay special attention to the law of
unforeseen consequences. For example, resellers have been known to
■ Buy larger quantities of a product than they can sell—or want to sell—in order
to qualify for volume discounts. The excess goods are then sold (diverted) to
other retailers, stored for future sales, or even destroyed or returned to the man-
ufacturer for “credit.”
■ Time their purchases at the ends of accounting periods in order to qualify for
rebates and allowances. This results in “lumpy” sales patterns for manufacturers,
making forecasting difficult, increasing problems with out-of-date products and
returns, and raising production costs.
In some instances, a particularly powerful channel “captain” can impose pricing disci-
pline on an entire channel. In most cases, however, each “link” in the distribution chain
can coordinate only its own pricing. A manufacturer, for example, may work out appro-
priate pricing incentives for wholesalers, and the wholesalers in turn may develop their
own pricing incentives for retailers.
In many countries and industries, it is illegal for suppliers to dictate the selling prices of
resellers. Manufacturers can’t dictate wholesaler selling prices, and wholesalers can’t
dictate retail prices. Consequently, members of the channel seek indirect methods of
influencing resellers’ prices.

8.4 Price Waterfall
The price waterfall is a way of describing the progression of prices from published
list price to the final price paid by a customer. Each drop in price represents a drop in
the “water level.” For example:
100
List Price
Dealer Discount
90
Cash Discount
85
Annual Rebate
82
Co-op Advertising
Net Price $80


Net Price per Unit ($)
Price Waterfall (%)
List Price per Unit ($)

In this structure, the average price paid by customers will depend on the list price of
a product, the sizes of discounts given, and the proportion of customers taking
advantage of those discounts.
By analyzing the price waterfall, marketers can determine where product value is
being lost. This can be especially important in businesses that allow the sales
channel to reduce prices in order to secure customers. The price waterfall
can help focus attention on deciding whether these discounts make sense for
the business.

Purpose: To assess the actual price paid for a product, in comparison
with the list price.
In pricing, the bad news is that marketers can find it difficult to determine the right list
price for a product. The good news is that few customers will actually pay that price any-
way. Indeed, a product’s net price—the price actually paid by customers—often falls
between 53% and 94% of its base price.5
Net Price: The actual price paid for a product by customers after all discounts and
allowances have been factored in. Also called the pocket price.
List Price: The price of a good or service before discounts and allowances are
considered.
Invoice Price: The price specified on the invoice for a product. This price will
typically be stated net of some discounts and allowances, such as dealer,
competitive, and order size discounts, but will not reflect other discounts and
allowances, such as those for special terms and cooperative advertising. Typically,
the invoice price will therefore be less than the list price but greater than the
net price.
Price Waterfall: The reduction of the price actually paid by customers for a prod-
uct as discounts and allowances are given at various stages of the sales process.
Because few customers take advantage of all discounts, in analyzing a product’s
price waterfall, marketers must consider not only the amount of each discount but
also the percentage of sales to which it applies.
As customers vary in their use of discounts, net price can fall into a wide range relative
to list price.


Construction
To assess a product’s price waterfall, one must plot the price a customer will pay at each
stage of the waterfall, specifying potential discounts and allowances in the sequence in
which those are usually taken or applied. For example, broker commissions are gener-
ally applied after trade discounts.
Net Price: The actual average price paid for a product at a given stage in its distri-
bution channel can be calculated as its list price, less discounts offered, with each
discount multiplied by the probability that it will be applied. When all discounts
are considered, this calculation yields the product’s net price.
Net Price ($) List Price ($) [Discount A ($) * Proportion of Purchases on which Discount
A is Taken (%)] [Discount B ($) * Proportion of Purchases on which
Discount B is Taken (%)] and so on . . .
Net Price per Unit ($)
Price Waterfall Effect (%)
List Price per Unit ($)

EXAMPLE: Hakan manages his own firm. In selling his product, Hakan grants
two discounts or allowances. The first of these is a 12% discount on orders of more than
100 units. This is given on 50% of the firm’s business and appears on its invoicing sys-
tem. Hakan also gives an allowance of 5% for cooperative advertising. This is not shown
on the invoicing system. It is completed in separate procedures that involve customers
submitting advertisements for approval. Upon investigation, Hakan finds that 80% of
customers take advantage of this advertising allowance.
The invoice price of the firm’s product can be calculated as the list price (50 Dinar per
unit), less the 12% order size discount, multiplied by the chance of that discount being
given (50%).
Invoice Price List Price [Discount * Proportion of Purchases on Which Discount Is Taken]
50 Dinar [(50 * 12%) * 50%]
50 Dinar 3 Dinar 47 Dinar
The net price further reduces the invoice price by the average amount of the cooperative
advertising allowance granted, as follows:
Net Price List Price [Discount * Proportion of Purchases on Which Discount Is Taken]
[Advertising Allowance * Proportion of Purchases on Which Ad Allowance Is
Taken] 50 Dinar [(50 * 12%) * 50%] [(50 * 5%) * 80%] 50 3 2
45 Dinar
To find the effect of the price waterfall, divide the net price by the list price.
45
Price Waterfall (%) 90%
50


To analyze the impact of discounts, allowances, and the overall price waterfall effect,
marketers require full information about sales, in both revenue and unit volume terms,
at an individual product level, including not only those discounts and allowances that
are formally recorded in the billing system, but also those granted without appearing
on invoices.
The major challenge in establishing the price waterfall is securing product-specific data
at all of these various levels in the sales process. In all but the smallest businesses, this is
likely to be quite difficult, particularly because many discounts are granted on an off-
invoice basis, so they might not be recorded at a product level in a firm’s financial sys-
tem. Further complicating matters, not all discounts are based on list price. Cash
discounts, for example, are usually based on net invoice price.
Where discounts are known in theory, but the financial system doesn’t fully record their
details, the problem is determining how to calculate the price waterfall. Toward that end,
marketers need not only the amount of each discount, but also the percentage of unit
sales for which customers take advantage of that discount.
The typical business offers a number of discounts from list prices. Most of these serve
the function of encouraging particular customer behaviors. For example, trade dis-
counts can encourage distributors and resellers to buy in full truckloads, pay invoices
promptly, and place orders during promotional periods or in a manner that smoothes
production. Over time, these discounts tend to multiply as manufacturers find it easier
to raise list price and add another discount than to eliminate discounts altogether.
Problems with discounts include the following:
■ Because it’s difficult to record discounts on a per-item basis, firms often record
them in aggregate. On this basis, marketers may see the total discounts provided
but have difficulty allocating these to specific products. Some discounts are
offered on the total size of a purchase, exacerbating this problem. This increases
the challenge of assessing product profitability.
■ Once given, discounts tend to be sticky. It is hard to take them away from cus-
tomers. Consequently, inertia often leaves special discounts in place, long after
the competitive pressures that prompted them are removed.
■ To the extent that discounts are not recorded on invoices, management often
loses track of them in decision-making.
As the Professional Pricing Society advises, when considering the price of a product,
“Look past the invoice price.”6


Deductions: Some “discounts” are actually deductions applied by a customer to an
invoice, adjusting for goods damaged in shipment, incorrect deliveries, late deliveries, or
in some cases, for products that did not sell as well as hoped. Deductions might not be
recorded in a way that can be analyzed, and they often are the subject of disputes.
Everyday Low Prices (EDLP): EDLP refers to a strategy of offering the same pricing
level from period to period. For retailers, there is a distinction between buying at EDLP
and selling at EDLP. For example, some suppliers offer constant selling prices to
retailers but negotiate periods during which a product will be offered on deal with
display and other retail promotions. Rather than granting temporary price discounts
to retailers, suppliers often finance these programs through “market development
funds.”
HI-LO (High-Low): This pricing strategy constitutes the opposite of EDLP. In HI-LO
pricing, retailers and manufacturers offer a series of “deals” or “specials”—times during
which prices are temporary decreased. One purpose of HI-LO pricing and other tem-
porary discounts is to realize price discrimination in the economic—not the legal—
sense of the term.

PRICE DISCRIMINATION AND TAILORING
When firms face distinct and separable market segments with different willingness to
pay (price elasticities), charging a single price means that the firm will “leave money on
the table”—not capture the full consumer value.
There are three conditions for price tailoring to be profitable:
■ Segments must have different elasticities (willingness to pay), and/or marketers
must have different costs of serving the segments (say shipping expenses) and
the incremental volume must be sufficiently large to compensate for the reduc-
tion in margin.
■ Segments must be separable—that is, charging different prices does not just
result in transfer between segments (for example, your father cannot buy your
dinner and apply the senior citizen discount).
■ The incremental profit from price tailoring exceeds the costs of implementing
multiple prices for the same product or service.
Price tailoring is clearly a euphemism for price discrimination. However, the latter term
is loaded with legal implications, and marketers understandably use it with caution.


When facing a total demand curve composed of identifiable segments with different
demand slopes, a marketer can use optimal pricing for each segment recognized, as
opposed to using the same price based upon aggregate demand. This is usually done by
■ Time: For example, subways or movie theaters charging a higher price during
rush or peak hour or products that are launched at a high price in the begin-
ning, “skimming” profits from early adopters.
■ Geography: Such as international market divisions—different prices for differ-
ent regions for DVDs, for example.
■ Tolerable discrimination: Identifying acceptable forms of segmentation, such
as discriminating between students or senior citizens and the general public.
Price differences cause gray markets; goods are imported from low-price to high-price
markets. Gray markets are common in some fashion goods and pharmaceuticals.

Caution: Regulations
Most countries have regulations that apply to price discrimination. As a marketer,
you should understand these regulations. In the U.S., the most important regulation
is the Robinson-Patman Act. It is mainly intended to control price differences
that might injure competition.7 We encourage you to visit the Federal Trade
Commission’s Web site (www.ftc.gov) for more information.

Abraham, M.M., and L.M. Lodish. (1990). “Getting the Most Out of Advertising and
Promotion,” Harvard Business Review, 68(3), 50.
Ailawadi, K., P. Farris, and E. Shames. (1999). “Trade Promotion: Essential to Selling Through
Resellers,” Sloan Management Review, 41(1), 83–92.
Christen, M., S. Gupta, J.C. Porter, R. Staelin, and D.R. Wittink. (1997). “Using Market-level Data
to Understand Promotion Effects in a Nonlinear Model,” Journal of Marketing Research (JMR),
34(3), 322.
“Roegner, E., M. Marn, and C. Zawada. (2005). “Pricing,” Marketing Management, Jan/Feb,
Vol. 14 (1).


9
ADVERTISING MEDIA
AND WEB METRICS

Introduction
Advertising: Impressions, Gross Rating Rich Media Display Time
Points, and Opportunities-to-See Rich Media Interaction Rate
Cost per Thousand Impressions Clickthrough Rates
(CPM) Rates
Cost per Impression, Cost per Click,
Reach/Net Reach and Frequency and Cost of Acquisition
Frequency Response Functions Visits, Visitors, and Abandonment
Effective Reach and Effective Frequency Bounce Rate
Share of Voice Friends/Followers/Supporters
Impressions, Pageviews, and Hits
Downloads

Advertising is the cornerstone of many marketing strategies. The positioning and com-
munications conveyed by advertising often set the tone and timing for many other sales
and promotion efforts. Advertising is not only the defining element of the marketing
mix, but it is also expensive and notoriously difficult to evaluate. This is because it is not
easy to track the incremental sales associated with advertising decisions. For many mar-
keters, media metrics are particularly confusing. A command of the vocabulary involved
in this field is needed to work with media planners, buyers, and agencies. A strong
understanding of media metrics can help marketers ensure that advertising budgets are
spent efficiently and directed toward a specific aim.
In the first part of this chapter, we discuss media metrics that reveal how many people
may be exposed to an advertising campaign, how often those people have an

287

opportunity to see the ads, and the cost of each potential impression. Toward that end,
we introduce the vocabulary of advertising metrics, including such terms as impres-
sions, exposures, OTS, rating points, GRPs, net reach, effective frequency, and CPMs.
In the second part of this chapter, we focus on metrics used in Web-based marketing
efforts. The Internet increasingly provides valuable opportunities to augment tradition-
al “broadcast” advertising with interactive media. In fact, many of the same advertising
media terms, such as impressions, are used to describe and evaluate Web-based adver-
tising. Other terms, such as clickthrough, are unique to the Web. Certain Web-specific
metrics are needed because the Internet, like direct mail, serves not only as a communi-
cations medium, but also as a direct sales channel that can provide real-time feedback
on the effectiveness of advertising in generating customer interest and sales.

9.1 Impressions An impression is As a metric, To understand
generated each impressions do how many times
time an advertise- not account for an advertisement
ment is viewed. quality of view- is viewed.
The number of ings. In this
impressions regard, a glimpse
achieved is a func- will have less effect
tion of an ad’s than a detailed
reach (the num- study. Impressions
ber of people see- are also called
ing it), multiplied exposures and
by its frequency opportunities-to-
(number of times see (OTS).
they see it).
9.1 Gross Rating Impressions Impressions To measure
Points (GRPs) divided by the expressed in rela- impressions in
number of tion to popula- relation to the
people in the tion. GRPs are number of people
audience for an cumulative across in the audience
advertisement. media vehicles, for an advertising
making it possible campaign.
to achieve GRPs
of more than
100%. Target
Rating Points
(TRPs) are meas-
ured in relation to
defined target
populations.


9.2 Cost per Cost of advertising CPM is a measure of To measure the
Thousand divided by impres- cost per advertising cost-effectiveness
Impressions sions generated (in impression, reckoning of the generation
(CPM) thousands). impressions in thou- of impressions.
sands. This makes it
easier to work with the
resulting dollar figures
than would be possible
on the basis of cost per
single impression.
9.3 Net Reach The number of Equivalent to reach. To measure the
people who receive Measures unique view- breadth of an
an advertisement. ers of an advertisement. advertisement’s
Often best mapped on spread across a
a Venn diagram. population.

9.3 Average The average num- Frequency is measured To measure how
Frequency ber of times that an only among people strongly an adver-
individual receives who have in fact seen tisement is con-
an advertisement, the advertisement centrated on a
given that he or she under study. given population.
is indeed exposed
to the ad.

9.4 Frequency Linear: All advertis- Linear model is often To model the
Response ing impressions are unrealistic, especially reaction of a
Functions equally impactful. for complex products. population to
Threshold: A cer- exposure to an
Threshold model is
tain number of advertisement.
often used, as it is sim-
impressions are ple and intuitive.
needed before an
advertising message Learning curve models
will sink in. often hypothesized, but
difficult to test for
Learning curve: An accuracy. Simpler
advertisement has models often work
little impact at first as well.
but gains force with
repetition and then
tails off as saturation
is achieved.

Continues

Chapter 9 Advertising Media and Web Metrics 289

9.5 Effective Reach achieved The effective frequen- To measure the
Reach among individuals cy rate constitutes a portion of an
who are exposed to crucial assumption in audience that is
an advertisement the calculation of this exposed to an
with a frequency metric. advertisement
greater than or equal enough times to
to the effective be influenced.
frequency.

9.5 Effective The number of As a rule of thumb in To determine
Frequency times an individual planning, marketers optimal exposure
must see an adver- often use an effective levels for an
tisement in order to frequency of 3. To the advertisement or
register its message. extent that it promises campaign, trading
to have a significant the risk of over-
impact on campaign spending against
results, this assump- the risk of failing
tion should be tested. to achieve the
desired impact.
9.6 Share of Quantifies the Market definition is To evaluate the
Voice advertising “pres- central to meaningful relative strength
ence” of a brand, results. Impressions or of advertising
campaign, or firm ratings represent a program within
in relation to total conceptually strong its market.
advertising in a basis for share of voice
market. calculations. Often,
however, such data are
unavailable.
Consequently, mar-
keters use spending, an
input, as a proxy for
output.
9.7 Pageviews The number of Represents the num- To provide a top-
times a Web page is ber of Web pages level measure of
served. served. Hits, by con- the popularity of
trast, represent a Web site.
pageviews multiplied
by the number of files
on a page, making it
as much a metric of
page design as of
traffic.


9.8 Rich Media The average time Can be heavily influ- To measure average
Display that rich media enced by unusually viewing time of
Time are displayed long display times. rich media.
per viewer. How data is gathered
is an important
consideration.
9.9 Rich Media Provides fraction The definition of Measures relative
Interaction of viewers inter- interaction should attractiveness of
Rate acting with the exclude actions unre- rich media and
rich media. lated to the rich ability to generate
media (a mouse viewer
crossing the rich engagement.
media to reach
another part of the
screen).
9.10 Clickthrough Number of click- An interactive meas- To measure the
Rate throughs as a ure of Web advertis- effectiveness of a
fraction of the ing. Has great Web advertisement
number of strengths, but clicks by counting those
impressions. represent only a step customers who are
toward conversion sufficiently
and are thus an inter- intrigued to click
mediate advertising through it.
goal.

9.11 Cost per Advertising Often used as To measure or
Click cost, divided by a billing mechanism. establish the cost-
number of clicks effectiveness of
generated. advertising.
9.11 Cost per Advertising cost, More directly related To measure or
Order divided by num- to profit than cost per establish the cost-
ber of orders click, but less effective effectiveness of
generated. in measuring pure advertising.
marketing. An adver-
tisement may generate
strong clickthrough
but yield weak conver-
sion due to a disap-
pointing product.

Continues


9.11 Cost per Advertising cost, Useful for purposes of To measure the
Customer divided by num- comparison to cus- cost-effectiveness
Acquired ber of customers tomer lifetime value. of advertising.
acquired. Helps marketers
determine whether
customers are worth
the cost of their
acquisition.

9.12 Visits The number of By measuring To measure audi-
unique viewings visits relative to ence traffic on a
of a Web site. pageviews, marketers Web site.
can determine
whether viewers are
investigating multiple
pages on a Web site.

9.12 Visitors The number of Useful in determining To measure the
unique Web site the type of traffic gen- reach of a Web
viewers in a given erated by a Web site—a site.
period. few loyal adherents, or
many occasional visi-
tors. The period over
which this metric is
measured can be
an important con-
sideration.

9.12 Abandonment The rate of pur- Can warn of weak To measure one
Rate chases started but design in an element of the
not completed. e-commerce site by close rate of
measuring the num- Internet business.
ber of potential cus-
tomers who lose
patience with a trans-
action process or are
surprised and put off
by “hidden” costs
revealed toward its
conclusion.


9.13 Bounce Rate Fraction of Web Requires a clear defi- Often used as an
site visitors who nition of when a visit indicator of site’s
view a single page. ends. Usually consid- relevance and
ers bounce rate with ability to generate
respect to visits rather visitor interest.
than visitors.

9.14 Friends/ Number of indi- Success depends on To measure size of
Followers/ viduals joining a target group and the social network,
Supporters social network. social nature of the but unlikely to
product. This metric measure
is unlikely to reflect engagement.
the ultimate aim of a
marketing campaign.
9.15 Downloads Number of times Counts the times a To determine
an application or file was downloaded, effectiveness in
file is down- not the number of getting applica-
loaded. customers who down- tions out to users.
loaded a file. It is
often useful to moni-
tor downloads started
but not completed.

9.1 Advertising: Impressions, Exposures,
Opportunities-To-See (OTS), Gross Rating Points
(GRPs), and Target Rating Points (TRPs)
Advertising impressions, exposures, and opportunities-to-see (OTS) all refer to
the same metric: an estimate of the audience for a media “insertion” (one ad) or
campaign.
Impressions = OTS = Exposures. In this chapter, we will use all these terms. It is
important to distinguish between “reach” (number of unique individuals exposed
to certain advertising) and “frequency” (the average number of times each such
individual is exposed).
Rating Point Reach of a media vehicle as a percentage of a defined population (for
example, a television show with a rating of 2 reaches 2% of the population).


Gross Rating Points (GRPs) = Total Ratings achieved by multiple media vehicles
expressed in rating points (for example, advertisements on five television shows with
an average rating of 30% would achieve 150 GRPs).
Gross rating points are impressions expressed as a percentage of a defined popula-
tion, and often total more than 100%. This metric refers to the defined population
reached rather than an absolute number of people. Although GRPs are used with a
broader audience, the term target rating points (TRPs) denotes a narrower definition
of the target audience. For example, TRPs might consider a specific segment such as
youths aged 15 to 19, whereas GRPs might be based on the total TV viewing
population.

Purpose: To measure the audience for an advertisement.
Impressions, exposures, and opportunities-to-see (OTS) are the “atoms” of media
planning. Every advertisement released into the world has a fixed number of planned
exposures, depending on the number of individuals in its audience. For example, an
advertisement that appears on a billboard on the Champs-Élysées in central Paris will
have an estimated number of impressions, based on the flow of traffic from visitors and
locals. An advertisement is said to “reach” a certain number of people on a number of
occasions, or to provide a certain number of “impressions” or “opportunities-to-see.”
These impressions or opportunities-to-see are thus a function of the number of
people reached and the number of times each such person has an opportunity to see
the advertisement.
Methodologies for estimating opportunities-to-see vary by type of media. In magazines,
for example, opportunities-to-see will not equal circulation because each copy of the
magazine may be read by more than one person. In broadcast media, it is assumed that
the quantified audience comprises those individuals available to hear or see an adver-
tisement. In print and outdoor media, an opportunity-to-see might range from a brief
glance to a careful consideration. To illustrate this range, imagine you’re walking down
a busy street. How many billboard advertisements catch your eye? You may not realize
it, but you’re contributing to the impressions of several advertisements, regardless of
whether you ignore them or study them with great interest.
When a campaign involves several types of media, marketers may need to adjust their
measures of opportunities-to-see in order to maintain consistency and allow for com-
parability among the different media.
Gross rating points (GRPs) are related to impressions and opportunities-to-see. They
quantify impressions as a percentage of the population reached rather than in absolute
numbers of people reached. Target rating points (TRPs) express the same concept but
with regard to a more narrowly defined target audience.


Construction
Impressions, Opportunities-to-See (OTS), and Exposures: The number of
times a specific advertisement is delivered to a potential customer. This is an
estimate of the audience for a media “insertion” (one ad) or a campaign.
Impressions = OTS = Exposures.
Impressions: The process of estimating reach and frequency begins with data that
sum all of the impressions from different advertisements to arrive at total “gross”
impressions.
Impressions (#) Reach (#) * Average Frequency (#)
The same formula can be rearranged as follows to convey the average number of times
that an audience was given the opportunity to see an advertisement. Average frequency
is defined as the average number of impressions per individual “reached” by an adver-
tisement or campaign.
Impressions (#)
Average Frequency (#)
Reach (#)
Similarly, the reach of an advertisement—that is, the number of people with an oppor-
tunity to see the ad—can be calculated as follows:
Impressions (#)
Reach (#)
Average Frequency (#)
Although reach can thus be quantified as the number of individuals exposed to an
advertisement or campaign, it can also be calculated as a percentage of the population.
In this text, we will distinguish between the two conceptualizations of this metric as
reach (#) and reach (%).
The reach of a specific media vehicle, which may deliver an advertisement, is often
expressed in rating points. Rating points are calculated as individuals reached by that
vehicle, divided by the total number of individuals in a defined population, and
expressed in “points” that represent the resulting percentage. Thus, a television program
with a rating of 2 would reach 2% of the population.
The rating points of all the media vehicles that deliver an advertisement or campaign
can be summed, yielding a measure of the aggregate reach of the campaign, known as
gross rating points (GRPs).
Gross Rating Points (GRPs): The sum of all rating points delivered by the media
vehicles carrying an advertisement or campaign.


EXAMPLE: A campaign that delivers 150 GRPs might expose 30% of the population
to an advertisement at an average frequency of 5 impressions per individual (150 30 *
5). If 15 separate “insertions” of the advertisement were used, a few individuals might be
exposed as many as 15 times, and many more of the 30% reached would only have 1 or
2 opportunities-to-see (OTS).

Gross Rating Points (GRPs) (%) Reach (%) * Average Frequency (#)
Impressions (#)
Gross Rating Points (GRPs) (%)
Defined Population (#)
Target Rating Points (TRPs): The gross rating points delivered by a media vehicle
to a specific target audience.

EXAMPLE: A firm places 10 advertising insertions in a market with a population of
5 people. The resulting impressions are outlined in the following table, in which “1”
represents an opportunity-to-see, and “0” signifies that an individual did not have an
opportunity to see a particular insertion.

Rating Points
Individual (Impressions/
Insertion A B C D E Impressions Population)
1 1 1 0 0 1 3 60
2 1 1 0 0 1 3 60
3 1 1 0 1 0 3 60

4 1 1 0 1 0 3 60

5 1 1 0 1 0 3 60

6 1 0 0 1 0 2 40

7 1 0 0 1 0 2 40
8 1 0 0 0 0 1 20
9 1 0 0 0 0 1 20
10 1 0 0 0 0 1 20
Totals 10 5 0 5 2 22 440


In this campaign, the total impressions across the entire population = 22.
As insertion 1 generates impressions upon three of the five members of the population,
it reaches 60% of that population, for 60 rating points. As insertion 6 generates impres-
sions upon two of the five members of the population, it reaches 40% of the population,
for 40 rating points. Gross rating points for the campaign can be calculated by adding the
rating points of each insertion.
Gross Rating Points (GRPs) Rating Points of Insertion 1 Rating Points of Insertion 2 etc.
440
Alternatively, gross rating points can be calculated by dividing total impressions by the
size of the population and expressing the result in percentage terms.
Impressions 22
Gross Rating Points (GRPs) * 100% * 100% 440
Population 5

Target rating points (TRPs), by contrast, quantify the gross rating points achieved by
an advertisement or campaign among targeted individuals within a larger population.
For purposes of this example, let’s assume that individuals A, B, and C comprise the tar-
geted group. Individual A has received 10 exposures to the campaign; individual B, 5
exposures; and individual C, 0 exposures. Thus, the campaign has reached two out of
three, or 66.67% of targeted individuals. Among those reached, its average frequency has
been 15/2, or 7.5. On this basis, we can calculate target rating points by either of the fol-
lowing methods.
Target Rating Points (TRPs) Reach (%) * Average Frequency
15
66.67% *
2
500
Impressions (#) 15
Target Rating Points (TRPs) 500
Targets (#) 3

Data on the estimated audience size (reach) of a media vehicle are typically made avail-
able by media sellers. Standard methods also exist for combining data from different
media to estimate “net reach” and frequency. An explanation of these procedures is
beyond the scope of this book, but interested readers might want to consult a company
dedicated to tracking rating points, such as Nielsen (www.nielsen.com), for further
detail.


Two different media plans can yield comparable results in terms of costs and total
exposures but differ in reach and frequency measures. In other words, one plan can
expose a larger audience to an advertising message less often, while the other delivers
more exposures to each member of a smaller audience. For an example, please see
Table 9.1.

Table 9.1 Illustration of Reach and Frequency

Reach Average Frequency* Total Exposures (Impressions, OTS)

Plan A 250,000 4 1,000,000
Plan B 333,333 3 1,000,000

*Average frequency is the average number of exposures made to each individual who has received at least
one exposure to a given advertisement or campaign. To compare impressions across media, or even within
classes of media, one must make a broad assumption: that there is some equivalency between the different
types of impressions generated by each media classification. Nonetheless, marketers must still compare
the “quality” of impressions delivered by different media.

Consider the following examples: A billboard along a busy freeway and a subway adver-
tisement can both yield the same number of impressions. Whereas the subway adver-
tisement has a captive audience, however, members of the billboard audience are
generally driving and concentrating on the road. As this example demonstrates, there
may be differences in the quality of impressions. To account for these differences, media
optimizers apply weightings to different media vehicles. When direct response data are
available, they can be used to evaluate the relative effectiveness and efficiency of impres-
sion purchases in different media. Otherwise, this weighting might be a matter of judg-
ment. A manager might believe, for example, that an impression generated by a TV
commercial is twice as effective as one made by a magazine print advertisement.
Similarly, marketers often find it useful to define audience sub-groups and generate sep-
arate reach and frequency statistics for each. Marketers might weight sub-groups differ-
ently in the same way that they weight impressions delivered through different media.1
This helps in evaluating whether an advertisement reaches its defined customer groups.
When calculating impressions, marketers often encounter an overlap of people who see
an advertisement in more than one medium. Later in this text, we will discuss how to
account for such overlap and estimate the percentage of people who are exposed to an
advertisement multiple times.


9.2 Cost per Thousand Impressions (CPM) Rates
Cost per thousand impressions (CPM) is the cost per thousand advertising impres-
sions. This metric is calculated by dividing the cost of an advertising placement by
the number of impressions (expressed in thousands) that it generates.
Advertising Cost ($)
Cost per Thousand Impressions (CPM) ($)
Impressions Generated (# in Thousands)

CPM is useful in comparing the relative efficiency of different advertising opportuni-
ties or media and in evaluating the costs of overall campaigns.

Purpose: To compare the costs of advertising campaigns within and
across different media.
A typical advertising campaign might try to reach potential consumers in multiple loca-
tions and through various media. The cost per thousand impressions (CPM) metric
enables marketers to make cost comparisons between these media, both at the planning
stage and during reviews of past campaigns.
Marketers calculate CPM by dividing advertising campaign costs by the number of
impressions (or opportunities-to-see) that are delivered by each part of the campaign.
As the impression counts are generally sizable, marketers customarily work with the
CPM impressions. Dividing by 1,000 is an industry standard.
Cost per Thousand Impressions (CPM): The cost of a media campaign, relative
to its success in generating impressions or opportunities-to-see.

Construction
To calculate CPM, marketers first state the results of a media campaign (gross impres-
sions) in thousands. Second, they divide that result into the relevant media cost:
Cost per Thousand Impressions (CPM) ($)
Impressions Generated (# in Thousands)


EXAMPLE: An advertising campaign costs $4,000 and generates 120,000 impressions.
On this basis, CPM can be calculated as follows:
Advertising Cost
Cost per Thousand Impressions
Impressions Generated (thousands)
$4,000
(120,000/1,000)
$4,000
$33.33
120

In an advertising campaign, the full cost of the media purchased can include agency
fees and production of creative materials, in addition to the cost of media space or
time. Marketers also must have an estimate of the number of impressions expected or
delivered in the campaign at an appropriate level of detail. Internet marketers (see
Section 9.7) often can easily access these data.
CPM is only a starting point for analysis. Not all impressions are equally valuable.
Consequently, it can make good business sense to pay more for impressions from some
sources than from others.
In calculating CPM, marketers should also be concerned with their ability to capture
the full cost of advertising activity. Cost items typically include the amount paid to a
creative agency to develop advertising materials, amounts paid to an organization that
sells media, and internal salaries and expenses related to overseeing the advertisement.

Cost per Point (CPP): The cost of an advertising campaign, relative to the rating
points delivered. In a manner similar to CPM, cost per point measures the cost per
rating point for an advertising campaign by dividing the cost of the advertising by
the rating points delivered.


9.3 Reach, Net Reach, and Frequency
Reach is the same as net reach; both of these metrics quantify the number or per-
centage of individuals in a defined population who receive at least one exposure to
an advertisement. Frequency measures the average number of times that each such
individual sees the advertisement.
Impressions (#) Reach (#) * Frequency (#)
Net reach and frequency are important concepts in describing an advertising cam-
paign. A campaign with a high net reach and low frequency runs the danger of being
lost in a noisy advertising environment. A campaign with low net reach but high
frequency can over-expose some audiences and miss others entirely. Reach and
frequency metrics help managers adjust their advertising media plans to fit their
marketing strategies.

Purpose: To separate total impressions into the number of people reached
and the average frequency with which those individuals are exposed
to advertising.
To clarify the difference between reach and frequency, let’s review what we learned in
Section 9.1. When impressions from multiple insertions are combined, the results are
often called “gross impressions” or “total exposures.” When total impressions are
expressed as a percentage of the population, this measure is referred to as gross rating
points (GRPs). For example, suppose a media vehicle reaches 12% of the population.
That vehicle will have a single-insertion reach of 12 rating points. If a firm advertised in
10 such vehicles, it would achieve 120 GRPs.
Now, let’s look at the composition of these 120 GRPs. Suppose we know that the 10
advertisements had a combined net reach of 40% and an average frequency of 3. Then
their gross rating points might be calculated as 40 * 3 120 GRPs.

EXAMPLE: A commercial is shown once in each of three time slots. Nielsen keeps
track of which households have an opportunity to see the advertisement. The commercial
airs in a market with only five households: A, B, C, D, and E. Time slots 1 and 2 both have
a rating of 60 because 60% of the households view them. Time slot 3 has a rating of 20.


Households with Households with no Rating Points of
Time Slot Opportunity-to-See Opportunity-to-See Time Slot
1 ABE CD 60
2 ABC DE 60

3 A BCDE 20

GRP 140

Impressions 7
GRP 140 (%)
Population 5

The commercial is seen by households A, B, C, and E, but not D. Thus, it generates
impressions in four out of five households, for a reach (%) of 80%. In the four house-
holds reached, the commercial is seen a total of seven times. Thus, its average frequency
can be calculated as 7/4, or 1.75. On this basis, we can calculate the campaign’s gross
rating points as follows:
4 7
GRP Reach (%) * Average Frequency (#) * 80% * 1.75 140 (%)
5 4

Unless otherwise specified, simple measures of overall audience size (such as GRPs or
impressions) do not differentiate between campaigns that expose larger audiences fewer
times and those that expose smaller audiences more often. In other words, these metrics
do not distinguish between reach and frequency.
Reach, whether described as “net reach” or simply “reach,” refers to the unduplicated
audience of individuals who have been exposed at least once to the advertising in ques-
tion. Reach can be expressed as either the number of individuals or the percentage of
the population that has seen the advertisement.
Reach: The number of people or percent of population exposed to an advertisement.
Frequency is calculated by dividing gross impressions by reach. Frequency is equal to the
average number of exposures received by individuals who have been exposed to at least
one impression of the advertising in question. Frequency is calculated only among indi-
viduals who have been exposed to this advertising. On this basis: Total Impressions =
Reach * Average Frequency.
Average Frequency: The average number of impressions per reached individual.
Media plans can differ in reach and frequency but still generate the same number of
total impressions.


Net Reach: This term is used to emphasize the fact that the reach of multiple
advertising placements is not calculated through the gross addition of all
individuals reached by each of those placements. Occasionally, the word “net”
is eliminated, and the metric is called simply reach.

EXAMPLE: Returning to our prior example of a 10-insertion media plan in a market
with a population of five people, we can calculate the reach and frequency of the plan by
analyzing the following data. As previously noted, in the following table, “1” represents
an opportunity-to-see, and “0” signifies that an individual did not have an opportunity
to see a particular insertion.

Individual
Rating Points
(Impressions/
Insertion A B C D E Impressions Population)
1 1 1 0 0 1 3 60
2 1 1 0 0 1 3 60
3 1 1 0 1 0 3 60

4 1 1 0 1 0 3 60

5 1 1 0 1 0 3 60

6 1 0 0 1 0 2 40

7 1 0 0 1 0 2 40
8 1 0 0 0 0 1 20
9 1 0 0 0 0 1 20
10 1 0 0 0 0 1 20
Totals 10 5 0 5 2 22 440

Reach is equal to the number of people who saw at least one advertisement. Four of the
five people in the population (A, B, D, and E) saw at least one advertisement.
Consequently, reach (#) 4.
Impressions 22
Average Frequency 5.5
Reach 4


Population needs to be excluded from the
reach of the second exposure to prevent
double counting.

Population with No Exposure

1 Exposure 1 Exposure
2

Advertisement A Advertisement B

Figure 9.1 Venn Diagram Illustration of Net Reach

When multiple vehicles are involved in an advertising campaign, marketers need infor-
mation about the overlap among these vehicles as well as sophisticated mathematical
procedures in order to estimate reach and frequency. To illustrate this concept, the fol-
lowing two-vehicle example can be useful. Overlap can be represented by a graphic
known as a Venn diagram (see Figure 9.1).

EXAMPLE: As an illustration of overlap effects, let’s look at two examples. Aircraft
International magazine offers 850,000 impressions for one advertisement. A second mag-
azine, Commercial Flying Monthly, offers 1 million impressions for one advertisement.
Example 1: Marketers who place advertisements in both magazines should not expect to
reach 1.85 million readers. Suppose that 10% of Aircraft International readers also read
Commercial Flying Monthly. On this basis, net reach (850,000 * .9) 1,000,000
1,765,000 unique individuals. Of these, 85,000 (10% of Aircraft International readers)
have received two exposures. The remaining 90% of Aircraft International readers have
received only one exposure. The overlap between two different media types is referred to
as external overlap.
Example 2: Marketers often use multiple insertions in the same media vehicle (such as
the July and August issues of the same magazine) to achieve frequency. Even if the esti-
mated audience size is the same for both months, not all of the same people will read the


magazine each month. For purposes of this example, let’s assume that marketers place
insertions in two different issues of Aircraft International, and that only 70% of readers of
the July issue also read the August issue. On this basis, net reach is not merely 850,000
(the circulation of each issue of Aircraft International) because the groups viewing the
two insertions are not precisely the same. Likewise, net reach is not 2 * 850,000, or
1.7 million, because the groups viewing the two insertions are also not completely dis-
parate. Rather, net reach 850,000 (850,000 * 30%) 1,105,000.
The reason: Thirty percent of readers of the August issue did not read the July issue
and so did not have the opportunity to see the July insertion of the advertisement. These
readers—and only these readers—represent incremental viewers of the advertisement in
August, and so they must be added to net reach. The remaining 70% of August readers were
exposed to the advertisement twice. Their total represents internal overlap or duplication.

Although we’ve emphasized the importance of reach and frequency, the impressions
metric is typically the easiest of these numbers to establish. Impressions can be aggre-
gated on the basis data originating from the media vehicles involved in a campaign. To
determine net reach and frequency, marketers must know or estimate the overlap
between audiences for different media, or for the same medium at different times. It is
beyond the capability of most marketers to make accurate estimates of reach and fre-
quency without access to proprietary databases and algorithms. Full-service advertising
agencies and media buying companies typically offer these services.
Assessing overlap is a major challenge. Although overlap can be estimated by perform-
ing customer surveys, it is difficult to do this with precision. Estimates based on man-
agers’ judgment occasionally must suffice.

9.4 Frequency Response Functions
Frequency response functions help marketers to model the effectiveness of multiple
exposures to advertising. We discuss three typical assumptions about how people
respond to advertisements: linear response, learning curve response, and threshold
response.
In a linear response model, people are assumed to react equally to every exposure
to an advertisement. The learning curve response model assumes that people are
initially slow to respond to an advertisement and then respond more quickly for a
time, until ultimately they reach a point at which their response to the message tails
off. In a threshold response function, people are assumed to show little response
until a critical frequency level is reached. At that point, their response immediately
rises to maximum capacity.


Frequency response functions are not technically considered metrics. Understanding
how people respond to the frequency of their exposure to advertising, however, is a
vital part of media planning. Response models directly determine calculations of
effective frequency and effective reach, metrics discussed in Section 9.5.

Purpose: To establish assumptions about the effects
of advertising frequency.
Let’s assume that a company has developed a message for an advertising campaign, and
that its managers feel confident that appropriate media for the campaign have been
selected. Now they must decide: How many times should the advertisement be placed?
The company wants to buy enough advertising space to ensure that its message is effec-
tively conveyed, but it also wants to ensure that it doesn’t waste money on unnecessary
impressions.
To make this decision, a marketer will have to make an assumption about the
value of frequency. This is a major consideration: What is the assumed value of repeti-
tion in advertising? Frequency response functions help us to think through the value
of frequency.
Frequency Response Function: The expected relationship between advertising
outcomes (usually in unit sales or dollar revenues) and advertising frequency.
There are a number of possible models for the frequency response functions used in
media plans. A selection among these for a particular campaign will depend on the
product advertised, the media used, and the judgment of the marketer. Three of the
most common models are described next.
Linear Response: The assumption behind a linear response function is that each
advertising exposure is equally valuable, regardless of how many other exposures to
the same advertising have preceded it.
Learning Curve Response: The learning or S curve model rests on the assumption
that a consumer’s response to advertising follows a progression: The first few times
an advertisement is shown, it does not register with its intended audience. As repeti-
tion occurs, the message permeates its audience and becomes more effective as people
absorb it. Ultimately, however, this effectiveness declines, and diminishing returns
set in. At this stage, marketers believe that individuals who want the information
already have it and can’t be influenced further; others simply are not interested.
Threshold Response: The assumption behind this model is that advertising has
no effect until its exposure reaches a certain level. At that point, its message
becomes fully effective. Beyond that point, further advertising is unnecessary and
would be wasted.


These are three common ways to value advertising frequency. Any function that accu-
rately describes the effect of a campaign can be used. Typically, however, only one func-
tion will apply to a given situation.

Construction
Frequency response functions are most useful if they can be used to quantify the effects
of incremental frequency. To illustrate the construction of the three functions described
in this section, we have tabulated several examples.
Tables 9.2 and 9.3 show the assumed incremental effects of each exposure to a certain
advertising campaign. Suppose that the advertisement will achieve maximum effect
(100%) at eight exposures. By analyzing this effect in the context of various response
functions, we can determine when and how quickly it takes hold.
Under a linear response model, each exposure below the saturation point generates one-
eighth, or 12.5%, of the overall effect.
The learning curve model is more complex. In this function, the incremental effective-
ness of each exposure increases until the fourth exposure and declines thereafter.
Under the threshold response model, there is no effect until the fourth exposure. At that
point, however, 100% of the benefit of advertising is immediately realized. Beyond that
point, there is no further value to be obtained through incremental advertising.
Subsequent exposures are wasted.
The effects of these advertising exposures are tabulated cumulatively in Table 9.3. In this
display, maximum attainable effectiveness is achieved when the response to advertising
reaches 100%.

Table 9.2 Example of the Effectiveness of Advertising

Exposure Frequency Linear Learning or S Curve Threshold Value

1 0.125 0.05 0
2 0.125 0.1 0

3 0.125 0.2 0

4 0.125 0.25 1
5 0.125 0.2 0
6 0.125 0.1 0

7 0.125 0.05 0

8 0.125 0.05 0


Table 9.3 Assumptions: Cumulative Advertising Effectiveness

Exposure Frequency Linear Learning or S Curve Threshold Value

1 12.5% 5% 0%
2 25.0% 15% 0%

3 37.5% 35% 0%

4 50.0% 60% 100%

5 62.5% 80% 100%

6 75.0% 90% 100%

7 87.5% 95% 100%

8 100.0% 100% 100%

We can plot cumulative effectiveness against frequency under each model (see
Figure 9.2). The linear function is represented by a simple straight line. The Threshold
assumption rises steeply at four exposures to reach 100%. The cumulative effects of the
learning curve model trace an S-shaped curve.
Frequency Response Function; Linear: Under this function, the cumulative effect
of advertising (up to the saturation point) can be viewed as a product of the
frequency of exposures and effectiveness per exposure.
Frequency Response Function; Linear (I) Frequency (#) * Effectiveness per Exposure (I)
Frequency Response Function; Learning Curve: The learning curve function
can be charted as a non-linear curve. Its form depends on the circumstances of a
particular campaign, including selection of advertising media, target audience,
and frequency of exposures.
Frequency Response Function; Threshold: The threshold function can be
expressed as a Boolean “if ” statement, as follows:
Frequency Response Function; Threshold Value (I) If (Frequency (#) Threshold (#), 1, 0)
Stated another way: In a threshold response function, if frequency is greater than or
equal to the threshold level of effectiveness, then the advertising campaign is 100%
effective. If frequency is less than the threshold, there is no effect.


Conceptions of Advertising Effectiveness
100.0%

90.0%

80.0%

70.0%
Cumulative Effectiveness

60.0%

50.0%

40.0%

30.0%

20.0%

10.0%

0.0%

1 2 3 4 5 6 7 8

Frequency of Exposure

Linear Learning or S-Curve Threshold Value

Figure 9.2 Illustration of Cumulative Advertising Effectiveness

A frequency response function can be viewed as the structure of assumptions made by
marketers in planning for the effects of an advertising campaign. In making these
assumptions, a marketer’s most useful information can be derived from an analysis of
the effects of prior ad campaigns. Functions validated with past data, however, are most
likely to be accurate if the relevant circumstances (such as media, creative, price, and
product) have not significantly changed.
In comparing the three models discussed in this section, the linear response function has
the benefit of resting on a simple assumption. It can be unrealistic, however, because it is
hard to imagine that every advertising exposure in a campaign will have the same effect.
The learning curve has intuitive appeal. It seems to capture the complexity of life better
than a linear model. Under this model, however, challenges arise in defining and


predicting an advertisement’s effectiveness. Three questions emerge: At what point does
the curve begin to ramp up? How steep is the function? When does it tail off? With
considerable research, marketers can make these estimates. Without it, however, there
will always be the concern that the learning curve function provides a spurious level of
accuracy.
Any implementation of the threshold response function will hinge on a firm’s estimate
of where the threshold lies. This will have important ramifications. If the firm makes a
conservative estimate, setting the tipping point at a high number of exposures, it may
pay for ineffective and unneeded advertising. If it sets the tipping point too low, howev-
er, it may not buy enough advertising media, and its campaign may fail to achieve the
desired effect. In implementation, marketers may find that there is little practical differ-
ence between using the threshold model and the more complicated learning curve.

Wear-in: The frequency required before a given advertisement or campaign
achieves a minimum level of effectiveness.
Wear-out: The frequency at which a given advertisement or campaign begins to
lose effectiveness or even yield a negative effect.

9.5 Effective Reach and Effective Frequency
The concept of effective frequency rests on the assumption that for an advertisement
or campaign to achieve an appreciable effect, it must attain a certain number of
exposures to an individual within a specified time period.
Effective reach is defined as the number of people or the percentage of the audience
that receives an advertising message with a frequency equal to or greater than the
effective frequency. That is, effective reach is the population receiving the “mini-
mum” effective exposure to an advertisement or campaign.

Purpose: To assess the extent to which advertising audiences are being
reached with sufficient frequency.
Many marketers believe their messages require repetition to “sink in.” Advertisers, like
parents and politicians, therefore repeat themselves. But this repetition must be moni-
tored for effectiveness. Toward that end, marketers apply the concepts of effective fre-
quency and effective reach. The assumptions behind these concepts run as follows: The
first few times people are exposed to an ad, it may have little effect. It is only when more
exposures are achieved that the message begins to influence its audience.


With this in mind, in planning and executing a campaign, an advertiser must determine
the number of times that a message must be repeated in order to be useful. This num-
ber is the effective frequency. In concept, this is identical to the threshold frequency in
the threshold response function discussed in Section 9.4. A campaign’s effective fre-
quency will depend on many factors, including market circumstances, media used, type
of ad, and campaign. As a rule of thumb, however, an estimate of three exposures per
purchase cycle is used surprisingly often.

Effective Frequency: The number of times a certain advertisement must be
exposed to a particular individual in a given period to produce a desired response.
Effective Reach: The number of people or the percentage of the audience that
receives an advertising message with a frequency equal to or greater than the
effective frequency.

Construction
Effective reach can be expressed as the number of people who have seen a particular
advertisement or the percentage of the population that has been exposed to that adver-
tisement at a frequency greater than or equal to the effective frequency.
Effective Reach (#, %) Individuals Reached with Frequency Equal to or
Greater Than Effective Frequency

EXAMPLE: An advertisement on the Internet was believed to need three view-
ings before its message would sink in. Population data showed the distribution in
Table 9.4.

Table 9.4 Number of Views of Advertisement

Number of Views Population
0 140,000

1 102,000

2 64,000

3 23,000

4 or more 11,000

Total 340,000


Because the effective frequency is 3, only those who have seen the advertisement three or
more times have been effectively reached. The effective reach is thus 23,000 11,000
34,000.
In percentage terms, the effective reach of this advertisement is 34,000/340,000 = 10% of
the population.

The Internet has provided a significant boost to data gathering in this area. Although
even Internet campaigns can’t be totally accurate with regard to the number of adver-
tisements served to each customer, data on this question in Web campaigns are far supe-
rior to those available in most other media.
Where data can’t be tracked electronically, it’s difficult to know how many times a
customer has been in a position to see an advertisement. Under these circumstances,
marketers make estimates on the basis of known audience habits and publicly available
resources, such as TV ratings.
Although test markets and split-cable experiments can shed light on the effects of adver-
tising frequency, marketers often lack comprehensive, reliable data on this question. In
these cases, they must make—and defend—assumptions about the frequency needed
for an effective campaign. Even where good historical data are available, media planning
should not rely solely on past results because every campaign is different.
Marketers must also bear in mind that effective frequency attempts to quantify the
average customer’s response to advertising. In practice, some customers will need more
information and exposure than others.

9.6 Share of Voice
Share of voice quantifies the advertising “presence” that a specific product or brand
enjoys. It is calculated by dividing the brand’s advertising by total market advertising,
and it is expressed as a percentage.
Brand Advertising ($, #)
Share of Voice (%)
Total Market Advertising ($, #)
For purposes of share of voice, there are at least two ways to measure “advertising”:
in terms of dollar spending; or in unit terms, through impressions or gross rating
points (GRPs). By any of these measures, share of voice represents an estimate of a
company’s advertising, as compared to that of its competitors.


Purpose: To evaluate the comparative level of advertising committed
to a specific product or brand.
Advertisers want to know whether their messages are breaking through the “noise” in
the commercial environment. Toward that end, share of voice offers one indication of a
brand’s advertising strength, relative to the overall market.
There are at least two ways to calculate share of voice. The classic approach is to divide
a brand’s advertising dollar spend by the total advertising spend in the marketplace.
Alternatively, share of voice can be based on the brand’s share of GRPs, impressions,
effective reach, or similar measures (see earlier sections in this chapter for more details
on basic advertising metrics).

Construction
Share of Voice: The percentage of advertising in a given market that a specific
product or brand enjoys.
Brand Advertising ($, #)
Share of Voice (%)
Total Market Advertising ($, #)

When calculating share of voice, a marketer’s central decision revolves around defining
the boundaries of the market. One must ensure that these are meaningful to the
intended customer. If a firm’s objective is to influence savvy Web users, for example, it
would not be appropriate to define advertising presence solely in terms of print media.
Share of voice can be computed at a company level, but brand- and product-level
calculations are also common.
In executing this calculation, a company should be able to measure its total advertising
spend fairly easily. Determining the ad spending for the market as a whole can be
fraught with difficulty, however. Complete accuracy will probably not be attainable. It is
important, however, that marketers take account of the major players in their market.
External sources such as annual reports and press clippings can shed light on competi-
tors’ ad spending. Services such as leading national advertisers (LNA) can also provide
useful data. These services sell estimates of competitive purchases of media space and
time. They generally do not report actual payments for media, however. Instead, costs
are estimated on the basis of the time and space purchased and on published “rate
cards” that list advertised prices. In using these estimates, marketers must bear in mind
that rate cards rarely cite the discounts available in buying media. Without accounting


for these discounts, published media spending estimates can be inflated. Marketers are
advised to deflate them by the discount rates they themselves receive on advertising.
A final caution: Some marketers might assume that the price of advertising is equal to
the value of that advertising. This is not necessarily the case. With this in mind, it can be
useful to augment a dollar-based calculation of share of voice with one based on
impressions.

9.7 Impressions, Pageviews, and Hits
As noted in Section 9.1, impressions represent the number of opportunities that have
been presented to people to see an advertisement. The best available measures of
this figure use technology in an effort to judge whether a given advertisement was
actually seen. But this is never perfect. Many recorded impressions are not actually
perceived by the intended viewer. Consequently, some marketers refer to this metric
as opportunities-to-see.
In applying this concept to Internet advertising and publishing, pageviews represent
the number of opportunities-to-see for a given Web page. Every Web page is
composed of a variety of individual objects and files, which can contain text, images,
audio, and video. The total number of these files requested in a given period is the
number of hits a Web site or Web server receives. Because pages composed of many
small files generate numerous hits per pageview, one must take care not to be overly
impressed by large hit counts.

Purpose: To assess Web site traffic and activity.
To quantify the traffic a Web site generates, marketers monitor pageviews—the number
of times a page on a Web site is accessed.
In the early days of e-commerce, managers paid attention to the number of hits a Web
site received. Hits measure file requests. Because Web pages are composed of numerous
text, graphic, and multimedia files, the hits they receive are a function not only of
pageviews, but also of the way those pages were composed by their Web designer.
As marketing on the Internet has become more sophisticated, better measures of Web
activity and traffic have evolved. Currently, it is more common to use pageviews as the
measure of traffic at a Web location. Pageviews aim to measure the number of times a
page has been displayed to a user. It thus should be measured as close to the end user as
possible. The best technology counts pixels returned to a server, confirming that a page
was properly displayed. This pixel2 count technique yields numbers closer to the end
user than would a tabulation of requests to the server, or of pages sent from the server


in response to a request. Good measurement can mitigate the problems of inflated
counts due to servers not acting on requests, files failing to serve on a user’s machine, or
users terminating the serving of ads.

Hits: A count of the number of files served to visitors on the Web. Because Web
pages often contain multiple files, hits is a function not only of pages visited, but
also of the number of files on each page.
Pageviews: The number of times a specific page has been displayed to users. This
should be recorded as late in the page-delivery process as possible in order to get
as close as possible to the user’s opportunity to see. A page can be composed of
multiple files.
For marketing purposes, a further distinction needs to be made as to how many
times an advertisement was viewed by unique visitors. For example, two individuals
entering a Web page from two different countries might receive the page in their
respective languages and might not receive the same ad. One example of an advertise-
ment that changes with different visitors is an embedded link with a banner ad.
Recognizing this potential for variation, advertisers want to know the number of times
that their specific advertisement was displayed to visitors, rather than a site’s number
of pageviews.
With this in mind, Internet advertisers often perform their analyses in terms of
impressions—sometimes called ad impressions or ad views. These represent the num-
ber of times an advertisement is served to visitors, giving them opportunities to see it.
(Many of the concepts in this section are in line with the terms covered in the advertis-
ing section, Section 9.1.)
For a single advertisement served to all visitors on a site, impressions are equal to the
number of pageviews. If a page carries multiple advertisements, the total number of all
ad impressions will exceed the number of pageviews.

Construction
Hits: The number of hits on a Web site is a function of the number of pageviews mul-
tiplied by the number of files comprising each page. Hit counts are likely to be more rel-
evant to technicians responsible for planning server capacity than to marketers
interested in measuring visitor activity.
Hits (#) Pageviews (#) * Files on the Page (#)
Pageviews: The number of pageviews can be easily calculated by dividing the number of
hits by the number of files on the page.
Hits (#)
Pageviews (#)
Files on the Page (#)


EXAMPLE: There are 250,000 hits on a Web site that serves five files each time a page
is accessed. Pageviews 250,000/5 50,000.
If the Web site served three files per page and generated 300,000 pageviews, then hits
would total 3 * 300,000 900,000.

Pageviews, page impressions, and ad impressions are measures of the responses of a
Web server to page and ad requests from users’ browsers, filtered to remove robotic
activity and error codes prior to reporting. These measures are recorded at a point as
close as possible to the user’s opportunity to see the page or ad.3
A count of ad impressions can be derived from pageviews if the percentage of pageviews
that contain the ad in question is known. For example, if 10% of pageviews receive the
advertisement for a luxury car, then the impressions for that car ad will equal 10% of
pageviews. Web sites that serve the same advertisement to all Web users are much easier
to monitor because only one count is required.
These metrics quantify opportunities-to-see without taking into account the number of
ads actually seen or the quality of what is shown. In particular, these metrics do not
account for the following:
■ Whether the message appeared to a specific, relevant, defined audience.
■ Whether the people to whom the pages appeared actually looked at them.
■ Whether those who looked at the pages had any recall of their content, or of the
advertising messages they contained, after the event.
Despite the use of the term impression, these measures do not tell a business manager
about the effect that an advertisement has on potential customers. Marketers can’t be
sure of the effect that pageviews have on visitors. Often, pageview results will consist of
data that include duplicate showings to the same visitor. For this reason, the term gross
impressions might be used to suggest a key assumption—that opportunities-to-see can
be delivered to the same viewer on multiple occasions.


9.8 Rich Media Display Time

Marketers use the rich media display time metric to monitor how long their adver-
tisements are holding the attention of potential customers.
Total Rich Media Display Time (#)
Average Rich Media Display Time (#) =
Total Rich Media Impressions (#)

Rich media display time represents an important way of tracking the success of
Internet advertising.

Purpose: To determine how long an advertisement is viewed.
Rich media is a term used for interactive media that allows consumers to be more
actively engaged than they might be with a billboard, a TV advertisement, or even a tra-
ditional display Web advertisement. Rich media metrics, or Audience Interaction
Metrics, are very similar in principle to other advertising metrics. Marketers want to
track whether the advertisement is effective at grabbing and maintaining the attention
of potential customers and so they track how long people spend “viewing” the adver-
tisement as a proxy for how interested they are in the content of the advertisement. The
rich media display time shows how long, on average, people spend engaged with the rich
media.

Construction
Rich media display time is simply the average time that viewers spent with the rich
media of an advertisement. For this the marketer will need the total amount of time
spent with the rich media and the total number of times that the rich media was dis-
played. It is a simple matter to create an average time in seconds spent with the rich
media by dividing the total amount of time in seconds spent by the total number of
impressions.
Total Rich Media Display Time (#)
Average Rich Media Display Time (#) =

As with many Web-based metrics, data often seem abundant to marketers who come
from the offline world. However, there are several measurement issues the marketer
must address in order to convert the abundance of data into useful metrics. For exam-
ple, marketers usually cut display times off at some upper bound, that is, if the piece of


rich media has been displayed for five minutes, it is safe to assume the viewer has prob-
ably gone to make a cup of coffee or been otherwise distracted. The question of how
long a displayed piece of rich media was actually viewed is similar to the question offline
marketers face with respect to whether an offline advertisement was viewed. A slight
advantage here goes to the rich online media in that most displays of rich media begin
because of an active request of the viewer…whereas no such action is required offline.
This metric, because it usually deals with short periods of time, can be influenced by
unusual events. Take a simplified example: If five people see the rich media display for
one second each and one person sees it for 55 seconds, the (average) rich media display
time is ten seconds. There is no way to distinguish this average display time from the
average time generated by six moderately interested viewers each viewing the advertise-
ment for ten seconds. Such is the case with any average.
Marketers should be clear that they understand how the data were gathered and be espe-
cially aware of any changes in the way the data were gathered. Changes in the way the
data were gathered and the metric constructed may be necessary for technological rea-
sons, but will limit the usefulness of the metric as longitudinal comparisons are no
longer valid. At a minimum, the marketer must be aware of and account for measure-
ment changes when interpreting the metric.

9.9 Rich Media Interaction Rate
Marketers use the rich media interaction rate to assess the effectiveness of a single
rich media advertisement in generating engagement from its viewers.

Total Rich Media Impressions with Interactions (#)
Rich Media Interaction Rate (%) =

Rich media interaction rate represents an important way of tracking the success of
Internet advertising in that it monitors the fraction of impressions that generate
interaction on the part of the viewer.

Purpose: To measure and monitor active involvement with an
advertisement.
The rich media interaction rate tracks how actively involved potential consumers are
with an advertisement. The big advantage of rich media is the ability of viewers to inter-
act with it. Marketers using rich media can have a much better idea of potential cus-
tomers’ reactions to an advertisement simply because these interactions are counted.
They can monitor whether potential customers are simply passively “viewing” the media


on their screen or are actively engaged by taking some traceable action. A user who
interacts is showing evidence of being more actively engaged and is thus probably more
likely to move toward purchase.

Construction
This metric is the number of impressions of an advertisement that were interacted with
divided by the total number of impressions of that advertisement. It tells the marketers
how successful any advertisement was at getting potential customers to engage with it in
some way, (mouse rollover, click on, etc.). As an example, a rich media advertisement
that was displayed 100 times with an interaction rate of 15% would mean that 15 of the
impressions resulted in some kind of interaction whereas 85 resulted in no interaction.
Total Rich Media Impressions with Interactions (#)
Rich Media Interaction Rate (%) =

Data for this metric will typically be available. Indeed the metric itself might be report-
ed as part of a standard reporting package. One important decision that has to be made
in generating the metric is what counts as an interaction. This will depend upon the
potential actions that the viewers could take, which in turn depends upon the precise
form of the advertisement. What counts as an interaction will usually have some lower
bound. For example, an interaction is only counted if the visitor spends more than one
second with his mouse over the impression. (This is designed to exclude movements of
the mouse unrelated to the advertisement such as moving the mouse to another part of
the page.)
As is true of any advertising, marketers should not forget the goal of their advertising.
Interaction is unlikely to be an end in itself. As such, a larger interaction rate, which
might be secured by gimmicks that appeal to people who will never buy the product,
may be no better than a smaller rate if the larger rate doesn’t move the visitor closer to
a sale (or some other high order objective).

Related Metrics
Rich Media Interaction Time: This metric captures the total amount of time that a vis-
itor spends interacting with an advertisement. This is an accumulation of the total time
spent interacting per visit on a single page. So on a visit to a page a user might interact
with the rich media for two interactions of two seconds each and so have an interaction
time of four seconds.


Video Interactions: Video metrics are very similar to rich media metrics. Indeed video
can be classified as rich media depending upon the way it is served to the viewer. Similar
principles apply, and the marketer should track how long viewers engage with the video
(the amount of time the video plays), what viewers do with the video (pause it, mute it),
and the total and specific interactions with the video (which show evidence of attention
to the video). Such metrics are then summarized across the entire pool of visitors, (for
instance the average visit led to the video being played for 12 seconds).

9.10 Clickthrough Rates
Clickthrough rate is the percentage of impressions that lead a user to click on an ad.
It describes the fraction of impressions that motivate users to click on a link, causing
a redirect to another Web location.
Clickthroughs (#)
Clickthrough Rate (%)
Impressions (#)

Most Internet-based businesses use clickthrough metrics. Although these metrics are
useful, they should not dominate all marketing analysis. Unless a user clicks on a
“Buy Now” button, clickthroughs measure only one step along the path toward a
final sale.

Purpose: To capture customers’ initial response to Web sites.
Most commercial Web sites are designed to elicit some sort of action, whether it be to
buy a book, read a news article, watch a music video, or search for a flight. People gen-
erally don’t visit a Web site with the intention of viewing advertisements, just as people
rarely watch TV with the purpose of consuming commercials. As marketers, we want to
know the reaction of the Web visitor. Under current technology, it is nearly impossible
to fully quantify the emotional reaction to the site and the effect of that site on the firm’s
brand. One piece of information that is easy to acquire, however, is the clickthrough
rate. The clickthrough rate measures the proportion of visitors who initiated action
with respect to an advertisement that redirected them to another page where they might
purchase an item or learn more about a product or service. Here we have used “clicked
their mouse” on the advertisement (or link) because this is the generally used term,
although other interactions are possible.

Construction
Clickthrough Rate: The clickthrough rate is the number of times a click is made
on the advertisement divided by the total impressions (the times an advertisement
was served).


Clickthroughs (#)
Clickthrough Rate (%)
Impressions (#)
Clickthroughs: If you have the clickthrough rate and the number of impressions,
you can calculate the absolute number of clickthroughs by multiplying the click-
through rate by the impressions.
Clickthroughs (#) Clickthrough Rate (%) * Impressions (#)

EXAMPLE: There are 1,000 clicks (the more commonly used shorthand for click-
throughs) on a Web site that serves up 100,000 impressions. The clickthrough rate is 1%.
1,000
Clickthrough Rate 1%
100,000

If the same Web site had a clickthrough rate of 0.5%, then there would have been 500
clickthroughs:
Clickthrough Rate 100,000 * 0.5% 500
If a different Web site had a 1% clickthrough rate and served up 200,000 impressions,
there would have been 2,000 clicks:
# of Clicks 1% * 200,000 2,000

The number of impressions is a necessary input for the calculation. On simpler Web sites,
this is likely to be the same as pageviews; every time the page is accessed, it shows the same
details. On more sophisticated sites, different advertisements can be shown to different
viewers. In these cases, impressions are likely to be some fraction of total pageviews. The
server can easily record the number of times the link was clicked (see Figure 9.3).
First, remember that clickthrough rate is expressed as a percentage. Although high click-
through rates might in themselves be desirable and help validate your ad’s appeal, com-
panies will also be interested in the total number of people who clicked through.
Imagine a Web site with a clickthrough rate of 80%. It may seem like a highly successful
Web site until management uncovers that only a total number of 20 people visited the
site with 16 clicking through compared with an objective of 500 visitors.
Also remember that a click is a very weak signal of interest. Individuals who click on an
ad might move on to something else before the new page is loaded. This could be
because the person clicked on the advertisement by accident or because the page took
too long to load. This is a problem that is of greater significance with the increase in
richer media advertisements. Marketers should understand their customers. Using large


video files is likely to increase the number of people abandoning the process before the
ad is served, especially if the customers have slower connections.

Clickthrough rate
New Page
captures numbers
Served
clicking on ad.
Clicked
Potential
Customer
Cancelled Before Ad Served

Didn’t Click

Out of Process

Figure 9.3 Clickthrough Process

As with impressions, try to ensure that you understand the measures. If the measure is
of clicks (the requests received from client machines to the server to send a file), then
there may be a number of breakage points between the clickthrough rate and the
impressions of the ad generated from a returned pixel count. Large discrepancies should
be understood—is it technical (the size/design of the advertisement) or weak interest
from clickers?
Clicks are the number of times the advertisement was interacted with, not the number
of customers who clicked. An individual visitor can click on an ad several times—either
in a single session or across multiple sessions. Only the most sophisticated Web sites
control the number of times they show a specific advertisement to the same customer.
This means that most Web sites can only count the number of times the ad was clicked,
not the number of visitors who clicked on an ad. Finally, the clickthrough rate must be
interpreted relative to an appropriate baseline. Clickthrough rates for banner ads are
very low and continue to fall. In contrast, clickthrough rates for buttons that simply take
visitors to the next page on a site should be much higher. An analysis of how click-
through rates change as visitors navigate through various pages can help identify “dead
end” pages that visitors rarely move beyond.


9.11 Cost per Impression, Cost per Click,
and Cost per Order
These three metrics measure the average cost of impressions, clicks, and customers.
All three are calculated in the same way—as the ratio of cost to the number of result-
ing impressions, clicks, or customers.
Cost per Impression
Number of Impressions (#)
Cost per Click ($)
Number of Clicks (#)
Cost per Order ($)
Orders (#)
These metrics are the starting point for assessing the effectiveness of a company’s
Internet advertising and can be used for comparison across advertising media and
vehicles and as an indicator of the profitability of a firm’s Internet marketing.

Purpose: To assess the cost effectiveness of Internet marketing.
In this section, we present three common ways of measuring the cost effectiveness of
Internet advertising. Each has benefits depending upon the perspective and end goal of
the advertising activity.

Cost per Impression: The cost to offer potential customers one opportunity to see
an advertisement.
Cost per Click: The amount spent to get an advertisement clicked.
Cost per click has a big advantage over cost per impression in that it tells us something
about how effective the advertising was. Clicks are a way to measure attention and inter-
est. Inexpensive ads that few people click on will have a low cost per impression and a
high cost per click. If the main purpose of an ad is to generate a click, then cost per click
is the preferred metric.
Cost per Order: The cost to acquire an order.
If the main purpose of the ad is to generate sales, then cost per order is the preferred
metric.
Once a certain number of Web impressions are achieved, the quality and placement of
the advertisement will affect clickthrough rates and the resulting cost per click (see
Figure 9.4).


Further along, measures are better tied to overall business objectives.

Earlier in the process, measures are less affected by noise.

Potential Sees Ad Follows Order
Customer Cost per Link Placed
Impressions Cost per Cost per
Click Order

Doesn’t See Ad Doesn’t Click Doesn’t Buy

Customer Out of Process

Figure 9.4 The Order Acquisition Process

Construction
The formulas are essentially the same for the alternatives; just divide the cost by the
appropriate number, for example, impressions, clicks, or orders.
Cost per Impression: This is derived from advertising cost and the number of impres-
sions.
Cost per Impression ($)
Number of Impressions (#)
Remember that cost per impression is often expressed as cost per thousand impressions
(CPM) in order to make the numbers easier to manage (for more on CPM, refer to
Section 9.2).
Cost per Click: This is calculated by dividing the advertising cost by the number of
clicks generated by the advertisement.
Cost per Click ($)
Clicks (#)
Cost per Order: This is the cost to generate an order. The precise form of this cost
depends on the industry and is complicated by product returns and multiple sales chan-
nels. The basic formula is
Cost per Order ($)
Orders Placed (#)


EXAMPLE: An Internet retailer spent $24,000 on online advertising and generated
1.2 million impressions, which led to 20,000 clicks, with 1 in 10 clicks resulting in a
purchase.
$24,000
Cost per Impression $0.02
1,200,000

$24,000
Cost per Click $1.20
20,000

If 1 in 10 of the clicks resulted in a purchase
$24,000
Cost per Order $12.00
2,000

This last calculation is also called “cost per purchase.”

The Internet has provided greater availability of advertising data. Consequently,
Internet advertising metrics are likely to rely on data that is more readily obtainable
than data from conventional channels. The Internet can provide more information
about how customers move through the system and how individual customers behave
at the purchase stage of the process.
For advertisers using a mix of online and “offline” media, it will be difficult to categorize
the cause and effect relationships between advertising and sales from both online and
offline sources. Banner ads might receive too much credit for an order if the customer
has also been influenced by the firm’s billboard advertisement. Conversely, banner ads
might receive too little credit for offline sales.
The calculations and data we have discussed in this section are often used in contracts
compensating advertisers. Companies may prefer to compensate media and ad agencies
on the basis of new customers acquired instead of orders.

SEARCH ENGINES
Search engine payments help determine the placement of links on search engines. The
most important search engine metric is the cost per click, and it is generally the basis for
establishing the search engine placement fee. Search engines can provide plenty of data
to analyze the effectiveness of a campaign. In order to reap the benefits of a great Web
site, the firm needs to get people to visit it. In the previous section, we discussed how
firms measure traffic. Search engines help firms create that traffic.


Although a strong brand helps drive traffic to a firm’s site, including the firm’s Web
address in all of its offline advertising might not increase traffic count. In order to
generate additional traffic, firms often turn to search engines. It was estimated that
over $2.5 billion was spent on paid search marketing, which made up approximately
36% of total online spending of $7.3 billion in 2003.4 Other online spending was com-
posed of the following categories: 50% as impressions, 12% as banner ads, and 2% as
email advertising.
Paid search marketing is essentially paying for the placement of ads on search engines
and content sites across the Internet. The ads are typically small portions of text (much
like newspaper want ads) made to look like the results of an unpaid or organic search.
Payment is usually made only when someone clicks on the ad. It is sometimes possible
to pay more per click in return for better placement on the search results page. One
important subset of paid search is keyword search in which advertisers can bid to be dis-
played whenever someone searches for the keyword(s). In this case, companies bid on
the basis of cost per click. Bidding a higher amount per click gets you placed higher.
However, there is an added complexity, which is if the ad fails to generate several clicks,
its placement will be lowered in comparison to competing ads.
The measures for testing search engine effectiveness are largely the same as those used
in assessing other Internet advertising.
Cost per Click: The most important concept in search engine marketing is cost per
click. Cost per click is widely quoted and used by search engine companies in charging
for their services. Marketers use cost per click to build their budgets for search engine
payments.
Search engines ask for a “maximum cost per click,” which is a ceiling whereby the mar-
keter imposes the maximum amount they are willing to pay for an individual click. A
search engine will typically auction the placement of links and only charge for a click at
a rate just above the next highest bid. This means the maximum cost per click that a
company would be willing to pay can be considerably higher than the average cost per
click they end up paying.
Marketers often talk about the concept of daily spend on search engines—just as it
sounds, this is the total spent on paid search engine advertising during one day. In order
to control spending, search engines allow marketers to specify maximum daily spends.
When the maximum is reached, the advertisement receives no preferential treatment.
The formula is the multiple of average cost per click and the number of clicks:
Daily Spend ($) Average Cost per Click ($) * Number of Clicks (#)


EXAMPLE: Andrei, the Internet marketing manager of an online music retailer,
decides to set a maximum price of $0.10 a click. At the end of the week he finds that the
search engine provider has charged him a total of $350.00 for 1,000 clicks per day.
His average cost per click is thus the cost of the advertising divided by the number of
clicks generated:
Cost per Week
Cost per Click
Clicks per Week
$350
7,000
$0.05 a Click

Daily spend is also calculated as average cost per click times the number of clicks:
Daily Spend $0.05 * 1,000
$50.00

ADVICE FOR SEARCH ENGINE MARKETERS
Search engines typically use auctions to establish a price for the search terms they sell.
Search engines have the great advantage of having a relatively efficient market; all users
have access to the information and can be in the same virtual location. They tend to
adopt a variant on the second price auction. Buyers only pay the amount needed for
their requested placement.
Cost per Customer Acquired: Similar to cost per order when the order came from
a new customer. Refer to Chapter 5, “Customer Profitability,” for a discussion on
defining customer and acquisition costs.

9.12 Visits, Visitors, and Abandonment
Visits measures the number of sessions on the Web site. Visitors measures the num-
ber of people making those visits. When an individual goes to a Web site on Tuesday
and then again on Wednesday, this should be recorded as two visits from one visitor.
Visitors are sometimes referred to as “unique visitors.” Visitors and unique visitors
are the same metric.


Abandonment usually refers to shopping carts. The total number of shopping carts
used in a specified period is the sum of the number abandoned and the number that
resulted in complete purchases. The abandonment rate is the ratio of the number of
abandoned shopping carts to the total.

Purpose: To understand Web site user behavior.
Web sites can easily track the number of pages requested. As we saw earlier in Section 9.7,
the pageviews metric is useful but far from complete. In addition to counting the num-
ber of pageviews a Web site delivers, firms will also want to count the number of times
someone visits the Web site and the number of people requesting those pages.
Visits: The number of times individuals request a page on the firm’s server for the
first time. Also known as sessions.
The first request counts as a visit. Subsequent requests from the same individual do
not count as visits unless they occur after a specified timeout period (usually set at
30 minutes).
Visitors: The number of individuals requesting pages from the firm’s server during
a given period. Also known as unique visitors.
To get a better understanding of traffic on a Web site, companies attempt to track the
number of visits. A visit can consist of a single pageview or multiple pageviews, and one
individual can make multiple visits to a Web site. The exact specification of what con-
stitutes a visit requires an accepted standard for a timeout period, which is the number
of minutes of inactivity from the time of entering the page to the time of requesting a
new page.
In addition to visits, firms also attempt to track the number of individual visitors to
their Web site. Because a visitor can make multiple visits in a specified period, the num-
ber of visits will be greater than the number of visitors. A visitor is sometimes referred
to as a unique visitor or unique user to clearly convey the idea that each visitor is only
counted once.
The measurement of users or visitors requires a standard time period and can be dis-
torted by automatic activity (such as “bots”) that classify Web content. Estimation of
visitors, visits, and other traffic statistics are usually filtered to remove this activity by
eliminating known IP addresses for “bots,” by requiring registration or cookies, or by
using panel data.
Pageviews and visits are related. By definition, a visit is a series of pageviews grouped
together in a single session, so the number of pageviews will exceed the number of visits.


Consider the metrics as a series of concentric ovals as shown in Figure 9.5. In this view,
the number of visitors must be less than or equal to the number of visits, which must
be less than or equal to the number of pageviews, which must be equal to or less than
the number of hits. (Refer to Section 9.7 for details of the relationship between hits
and pageviews.)

Hits
Pageviews

Visits

Visitors

Figure 9.5 Relationship of Hits to Pageviews to Visits to Visitors

Another way to consider the relationship between visitors, visits, pageviews, and hits is
to consider the following example of one visitor entering a Web site of an online news-
paper (see Figure 9.6). Suppose that the visitor enters the site on Monday, Tuesday, and
Friday. In her visits she looks at a total of 20 pageviews. Those pages are made up of a
number of different graphic files, word files, and banner ads.
The ratio of pageviews to visitors is sometimes referred to as the average pages per visit.
Marketers track this average to monitor how the average visit length is changing over
time.
It is possible to dig even deeper and track the paths visitors take within a visit. This path
is called the clickstream.
Clickstream: The path of a user through the Internet.
The clickstream refers to the sequence of clicked links while visiting multiple sites.
Tracking at this level can help the firm identify the most and least appealing pages (see
Figure 9.7).


20 Pageviews 200 Hits

3 Visits 10 News 160 graphic files
1 Visitor Monday 5 Sports 20 word files
Tuesday 5 Business 20 banner ads
Friday

Figure 9.6 Example of Online Newspaper Visitor

News In Depth Sales

Welcome

Features Member Member
Benefits Login

Links Clickstream, the actual path taken by a customer

Figure 9.7 A Clickstream Documented

The analysis of clickstream data often yields significant customer insights. What path is
a customer most likely to take prior to purchase? Is there a way to make the most pop-
ular paths even easier to navigate? Should the unpopular paths be changed or even elim-
inated? Do purchases come at the end of lengthy or short sessions? At what pages do
sessions end?
A portion of the clickstream that deserves considerable attention is the subset of clicks
associated with the use of shopping carts. A shopping cart is a piece of software on the
server that allows visitors to select items for eventual purchase. Although shoppers in
brick and mortar stores rarely abandon their carts, abandonment of virtual shopping
carts is quite common. Savvy marketers count how many of the shopping carts used in
a specified period result in a completed sale versus how many are abandoned. The ratio
of the number of abandoned shopping carts to the total is the abandonment rate.


Abandonment Rate: The percentage of shopping carts that are abandoned.
To decide whether a visitor is a returning visitor or a new user, companies often employ
cookies. A cookie is a file downloaded onto the computer of a person surfing the Web
that contains identifying information. When the person returns, the Web server reads
the cookie and recognizes the visitor as someone who has been to the Web site previ-
ously. More advanced sites use cookies to offer customized content, and shopping carts
make use of cookies to distinguish one shopping cart from another. For example,
Amazon, eBay, and EasyJet all make extensive use of cookies to personalize the Web
views to each customer.
Cookie: A small file that a Web site puts on the hard drive of visitors for the
purpose of future identification.

Construction
Visitors: Cookies can help servers track unique visitors, but this data is never 100%
accurate (see the next section).
Abandoned Purchases: The number of purchases that were not completed.

EXAMPLE: An online comics retailer found that of the 25,000 customers who loaded
items into their electronic baskets, only 20,000 actually purchased:

Purchases Not Completed Purchases Initiated Less Purchases Completed
25,000 20,000 5,000
Not Completed 5,000
Abandonment Rate
Customer Initiation 25,000

20% Abandonment Rate

Visits can be estimated from log file data. Visitors are much more difficult to measure.
If visitors register and/or accept cookies, then at least the computer that was used for the
visit can be identified.
Meaningful results are difficult to get for smaller or more narrowly focused Web sites.
It is possible to bring in professionals in competitive research and user behavior.
Nielsen, among other services, runs a panel in the U.S. and a number of major
economies.5


9.13 Bounce Rate
Bounce Rate is a measure of the effectiveness of a Web site in encouraging visitors to
continue their visit. It is expressed as a percentage and represents the proportion of
visits that end on the first page of the Web site that the visitor sees.
Visits That Access Only a Single Page (#)
Bounce Rate (%)
Total Visits (#) to the Web site

High bounce rates typically indicate that the Web site isn’t doing a good job of
attracting the continuing interest of visitors.

Purpose: To determine the effectiveness of the Web site at generating the
interest of visitors.
Bounce rate is a commonly reported metric that reflects the effectiveness of Web sites at
drawing the continuing attention of visitors. The assumption behind the usefulness of
the metric is that the owner of the Web site wants visitors to visit more than just the
landing page. For most sites this is a reasonable assumption. For example, sites that are
seeking to sell goods want visitors to go to other pages to view the goods and ultimate-
ly make a purchase. Bounce rate is also a measure of how effective the company is at
generating relevant traffic. The more the Web site is relevant to the traffic coming to it,
the lower will be the bounce rate. This becomes particularly important when traffic is
generated through paid search. Money spent to generate traffic for whom the Web site
is not relevant (as reflected in a high bounce rate) is money wasted. The bounce rate is
a particularly useful measure in respect of the entry pages to Web sites. An entry page
with a very low bounce rate is doing its job of driving traffic to other pages. As Google
analytics explains; “The more compelling your landing pages, the more visitors will stay
on your site and convert.”6
Having a low bounce rate is often a prerequisite of having a successful e-commerce
presence.

Construction
Bounce Rate: The number of visits that access only a single page of a Web site divided
by the total number of visits to the Web site.
Visits that Access Only a Single Page (#)
Bounce Rate (%) =
Total Visits (#) to the Web site


Data to construct this metric, or even the metric itself, will usually come from the Web
site’s host as part of the normal reporting procedure. Given how common it is that
bounce rate is reported by default, it is a metric that is difficult to ignore. Construction
of the metric requires a precise definition of when a visit ends. Leaving the site may
come from closing the window, entering a new URL, clicking on a link off the site, hit-
ting the Back button or being timed out. After a timeout a new session is usually start-
ed if the visitor returns to the Web site. A lower timeout period results in increased
bounce rates, all else equal.
Reports may use the term visitors instead of visits. You should be clear what data is actu-
ally reported. Visits are much easier to track because when the same visitor makes
return visits, especially to different entry pages, it can be difficult to connect the return
visit to the original visitor. As such visits, rather than visitors, are most likely used to cal-
culate bounce rates.
This metric can also be defined and constructed for individual pages rather than the site
as a whole. Indeed the bounce rate for each page allows for more precise diagnosis of
problem areas on a Web site. One must interpret page bounce rates, however, in light of
the purpose of the page. For some pages, such as directions pages, a high bounce rate is
to be expected. The value of this metric will depend upon the objective of the organiza-
tion. Informational sites may develop a strong bond with their users through frequent
short interaction, such as checking sports scores. The organization may be comfortable
if many users do not visit other parts of the site, and may not be too concerned about
high bounce rates. However, most companies will probably want low bounce rates and
will actively monitor this important metric.

9.14 Friends/Followers/Supporters
Friends/Followers/Supporters is a very simple metric that measures the number of
individuals who join an organization’s social network.
Friends (#) = Number of friends of the entity registered on a social networking page (#)
A high number of friends signifies an active interest in the owner of the page. If a
brand has a high number of friends, this indicates a stronger brand with a loyal
customer base.


Purpose: To determine the effectiveness of a social networking presence.
We use the term friends to encompass followers, supporters, and other similar concepts.
Friends are members of a social networking site who register that they know, like and/or
support the owner of the social networking page. For instance a strong brand may have
many customers who want to publicly signal their love of the brand. Social networking
sites hold great benefits in allowing companies to develop customer relationships and
can help a company identify and communicate with committed customers.

Construction
Friends (#) = Number of friends of the entity registered on a social networking page (#)

Success in recruiting friends is likely to depend heavily on the group of people who
identify with the entity (e.g., individuals, brands, companies, or other groups). In the
case of brands, some customer segments are more reluctant to reveal their brand loyal-
ty than others, and as such two brands of equivalent strength may have very different
levels of social network presence. Similarly the product involved is likely to influence the
likelihood of registering as a friend at the social networking site. It is easy to think of
some vitally important but more private products that are relied upon by their users but
are less likely to gain public expressions of support than brands that are more related to
public consumption.
It is very hard to objectively judge the effectiveness of social networking activities.
Generally having more followers is an excellent sign of customer engagement. The more
customers who have an ongoing relationship with a brand that they are willing to pub-
licly support, the more likely the brand is to have strong customer awareness and loyal-
ty. It is worth noting, however, that Friends, as with many metrics, is most often an
intermediate metric rather than an aim of the organization itself. It is unlikely that most
organizations exist with the explicit objective of generating friends. As such it is rarely
sufficient to report the number of friends as a successful outcome of a marketing strat-
egy without any additional information. It is often appropriate to construct metrics
around the downstream outcomes and cost effectiveness of such strategies. A marketer
would be well advised to pay attention to the costs and ultimate benefits of social net-
working presence as well as the clear potential to engage with customers.
Cost per Friend: The cost to the organization per friend recruited.
Total Cost to Provide Social Networking Presence ($)
Cost per Friend =
Number of Friends (#)


Often the direct costs of having a social networking site are very low. This should not,
however, lead the marketer to conclude the cost is effectively zero. Sites have to be
designed, staff have to update the site, and marketers have to devise strategies.
Remember when calculating the cost of having a social network presence that the costs
should include all costs incurred in the provision of the social network presence.
Outcomes Per Friend: A similar attempt might be made to clarify the precise down-
stream outcomes gained by the presence of friends. (“Did we sell more ketchup?”) It is
often very hard to track outcomes to specific social networking actions. This does not
mean that an active social networking presence is not a vital part of an Internet market-
ing strategy, but when designing a presence the ultimate objective of the company needs
to be borne in mind. For example, friends are often recruited to “vote” in polls. The per-
centage of friends participating is a simple example of an “outcome per friend” metric
but probably not the ultimate objective.

9.15 Downloads
Monitoring downloads is a way of tracking engagement with the organization.
Downloads (#) = Number of times that an application or file is downloaded (#)
Downloads reflect the success of organizations at getting their applications distributed
to users.

Purpose: To determine effectiveness in getting applications out to users.
Downloads are a common way for marketers to gain a presence with consumers. This
includes applications for mobile phones, for MP3-style devices, and computers.
Apps for iPhones, software trials, spreadsheets, ring tones, white papers, pictures, and
widgets are examples of downloads. These downloads typically provide a benefit to the
consumer in return for a presence on the device of the user. For instance a weather app
might be branded with the Web site of the weather channel and provide updates on
atmospheric conditions. A consumer packaged goods company might supply an app
that suggests recipes that could use its products in novel ways.

Construction
Downloads (#) = Number of times that an application or file is downloaded (#)


Downloads is a simple count of the number of times an application or file is down-
loaded, regardless of who requested the download. It does not distinguish 10 identical
downloads to a given individual from 10 separate downloads to 10 separate individuals,
although these two situations may have dramatically different consequences for the
company. In this way downloads is akin to impressions where a given number of
impressions can be obtained by a variety of combinations of reach and frequency (see
section 9.3).
A consideration in the counting of downloads is how to handle downloads that are
started but not completed. The alternative to keeping track of both (allowing the con-
struction of a bounce-rate-like metric with respect to downloads) is to pick one or the
other (starts or completions). As always, it is imperative for the user to know which con-
vention was used in construction of the download metric.

Farris, Paul W., David Reibstein, and Ervin Shames. (1998). “Advertising Budgeting: A Report
from the Field,” monograph, New York: American Association of Advertising Agencies.
Forrester, J.W. (1959). “ADVERTISING: A Problem in Industrial Dynamics,” Harvard Business
Review, 37(2), 100.
Interactive Advertising Bureau. (2004). Interactive Audience Measurement and Advertising
Campaign Reporting and Audit Guidelines. United States Version 6.0b.
Lodish, L.M. (1997). “Point of View: J.P. Jones and M.H. Blair on Measuring Ad Effects: Another
P.O.V,” Journal of Advertising Research, 37(5), 75.
Net Genesis Corp. (2000). E-metrics Business Metrics for the New Economy. Net Genesis and
Target Marketing of Santa Barbara.
Tellis, G.J., and D.L. Weiss. (1995). “Does TV Advertising Really Affect Sales? The Role of
Measures, Models, and Data Aggregation,” Journal of Advertising, 24(3), 1.


10
MARKETING AND FINANCE

Introduction

Net Profit and Return on Sales (ROS) Project Metrics: Payback, NPV, IRR
Return on Investment (ROI) Return on Marketing Investment
Economic Profit (aka, EVA®)

As marketers progress in their careers, it becomes increasingly necessary to coordinate
their plans with other functional areas. Sales forecasts, budgeting, and estimating
returns from proposed marketing initiatives are often the focus of discussions between
marketing and finance. For marketers with little exposure to basic finance metrics, a
good starting point is to gain a deeper understanding of “rate of return.” “Return” is
generally associated with profit, or at least positive cash flow. “Return” also implies that
something has left—cash outflow. Almost all business activity requires some cash out-
flow. Even sales cost money that is only returned when bills are paid. In this chapter we
provide a brief overview of some of the more commonly employed measures of prof-
itability and profits. Understanding how the metrics are constructed and used by
finance to rank various projects will make it easier to develop marketing plans that meet
the appropriate criteria.
The first section covers net profits and return on sales (ROS). Next, we look at return on
investment (ROI), the ratio of net profit to amount of investment. Another metric that
accounts for the capital investment required to earn profits is economic profits (also
known as economic value added—EVA), or residual income. Because EVA and ROI
provide snapshots of the per-period profitability of firms, they are not appropriate for
valuing projects spanning multiple periods. For multi-period projects, three of the

337

most common metrics are payback, net present value (NPV), and internal rate of
return (IRR).
The last section discuses the frequently mentioned but rarely defined measure, return
on marketing investment (ROMI). Although this is a well-intentioned effort to measure
marketing productivity, consensus definitions and measurement procedures for “mar-
keting ROI” or ROMI have yet to emerge.

10.1 Net Profit Sales revenue less Revenue and costs The basic profit
total costs. can be defined in equation.
a number of ways
leading to confu-
sion in profit
calculations.

10.1 Return on Sales Net profit as a Acceptable level Gives the percent-
(ROS) percentage of of return varies age of revenue
sales revenue. between indus- that is being cap-
tries and business tured in profits.
models. Many
models can be
described as high
volume/low
return or vice
versa.

10.1 Earnings Before Earnings Before Strips out the Rough measure of
Interest, Taxes, Interest, Taxes, effect of account- operating cash
Depreciation, and Depreciation, and ing and financing flow.
Amortization Authorization. polices from
(EBITDA) profits. Ignores
important factors,
such as deprecia-
tion of assets.

10.2 Return on Net profits over Often meaningless A metric that
Investment the investment in the short term. describes how
(ROI) needed to gener- Variations such as well assets are
ate the profits. return on assets being used.
and return on
investment capital
analyze profits
in respect of
different inputs.


10.3 Economic Profit Net operating Requires a cost of Shows profit
(aka EVA®, profit after tax capital to be pro- made in dollar
Economic Value (NOPAT) less the vided/calculated. terms. Gives a
Added) cost of capital. clearer distinction
between the sizes
of returns than
does a percentage
calculation.

10.4 Payback The length of Will favor proj- Simple return
time taken to ects with quick calculation.
return the initial returns more
investment. than long-term
success.

10.4 Net Present Value The value of a The discount rate To summarize the
(NPV) stream of future used is the vital value of cash
cash flows after consideration flows over multi-
accounting for and should ple periods.
the time value of account for the
money. risk of the
investment
too.

10.4 Internal Rate of The discount rate IRR does not An IRR will typi-
Return (IRR) at which the NPV describe the mag- cally be compared
of an investment nitude of return; to a firm’s hurdle
is zero. $1 on $10 is the rate. If IRR is
same as $1 mil- higher than
lion on $10 mil- hurdle rate,
lion. invest; if lower,
pass.

10.5 Return on Incremental rev- Marketers need Compares the
Marketing enue attributable to establish an sales generated in
Investment to marketing over accurate Baseline revenue terms
(ROMI); Revenue the marketing to be able to with the market-
spending. meaningfully ing spending that
state what rev- helped generate
enue is attributa- the sales. The
ble to marketing. percentage term
helps comparison
across plans
of varying
magnitude.

Chapter 10 Marketing and Finance 339

10.1 Net Profit and Return on Sales
Net profit measures the profitability of ventures after accounting for all costs. Return
on sales (ROS) is net profit as a percentage of sales revenue.

Net Profit ($) Sales Revenue ($) Total Costs ($)
Net Profit ($)
Return on Sales—ROS (%)
Sales Revenue ($)

EBITDA ($) = Net Profit ($) + Interest Payments ($) + Taxes ($) +
Depreciation and Authorization Charges ($)
ROS is an indicator of profitability and is often used to compare the profitability of
companies and industries of differing sizes. Significantly, ROS does not account for
the capital (investment) used to generate the profit.
Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) is a rough
measure of operating cash flow, which reduces the effect of accounting, financing,
and tax polices on reported profits.

Purpose: To measure levels and rates of profitability.
How does a company decide whether it is successful or not? Probably the most common
way is to look at the net profits of the business. Given that companies are collections of
projects and markets, individual areas can be judged on how successful they are at
adding to the corporate net profit. Not all projects are of equal size, however, and one
way to adjust for size is to divide the profit by sales revenue. The resulting ratio is return
on sales (ROS), the percentage of sales revenue that gets “returned” to the company as
net profits after all the related costs of the activity are deducted.

Construction
Net profit measures the fundamental profitability of the business. It is the revenues of
the activity less the costs of the activity. The main complication is in more complex
businesses when overhead needs to be allocated across divisions of the company (see
Figure 10.1). Almost by definition, overheads are costs that cannot be directly tied to any
specific product or division. The classic example would be the cost of headquarters staff.
Net Profit: To calculate net profit for a unit (such as a company or division), subtract
all costs, including a fair share of total corporate overheads, from the gross revenues.


Sales Revenues for the Firm

Total Variable Costs Line Specific Overhead
Fixed Costs Business
Net Profit

Simple View of Business – Revenues and Costs

Figure 10.1 Profits Revenues Less Costs

Net Profit ($) Sales Revenue ($) Total Costs ($)
Return on Sales (ROS): Net profit as a percentage of sales revenue.
Net Profit ($)
Return on Sales (%)
Sales Revenue ($)

Earning before interest taxes, depreciation, and amortization (EBITDA) is a very popu-
lar measure of financial performance. It is used to assess the “operating” profit of the
business. It is a rough way of calculating how much cash the business is generating and
is even sometimes called the “operating cash flow.” It can be useful because it removes
factors that change the view of performance depending upon the accounting and
financing policies of the business. Supporters argue it reduces management’s ability to
change the profits they report by their choice of accounting rules and the way they gen-
erate financial backing for the company. This metric excludes from consideration
expenses related to decisions such as how to finance the business (debt or equity) and
over what period to depreciate fixed assets. EBITDA is typically closer to actual cash
flow than is NOPAT (discussed later in the chapter).
EBITDA can be calculated by adding back the costs of interest, depreciation, and amor-
tization charges and any taxes incurred.
EBITDA ($) = Net Profit ($) + Interest Payments ($) + Taxes Incurred ($) + Depreciation
and Amortization Charges ($)

Although it is theoretically possible to calculate profits for any sub-unit, such as a prod-
uct or region, often the calculations are rendered suspect by the need to allocate overhead
costs. Because overhead costs often don’t come in neat packages, their allocation among
the divisions or product lines of the company can often be more art than science.


For return on sales, it is worth bearing in mind that a “healthy” figure depends on the
industry and capital intensity (amount of assets per sales dollar). Return on sales is similar
to margin (%), except that ROS accounts for overheads and other fixed costs that are often
ignored when calculating margin (%) or contribution margin (%). (Refer to Section 3.1.)

Net operating profit after tax (NOPAT) deducts relevant income taxes but excludes
some items that are deemed to be unrelated to the main (“operating”) business.

10.2 Return on Investment
Return on investment is one way of considering profits in relation to capital invested.
Net Profit ($)
Return on Investment—ROI (%)
Investment ($)
Return on assets (ROA), return on net assets (RONA), return on capital (ROC), and
return on invested capital (ROIC) are similar measures with variations on how
“investment” is defined.
Marketing not only influences net profits but also can affect investment levels too.
New plants and equipment, inventories, and accounts receivable are three of the
main categories of investments that can be affected by marketing decisions.

Purpose: To measure per period rates of return on dollars invested
in an economic entity.
ROI and related metrics (ROA, ROC, RONA, and ROIC) provide a snapshot of prof-
itability adjusted for the size of the investment assets tied up in the enterprise.
Marketing decisions have obvious potential connection to the numerator of ROI (prof-
its), but these same decisions often influence assets usage and capital requirements (for
example, receivables and inventories). Marketers should understand the position of
their company and the returns expected. ROI is often compared to expected (or
required) rates of return on dollars invested.

Construction
For a single period review just divide the return (net profit) by the resources that were
committed (investment):
Net Profit ($)
Return on Investment (%)
Investment ($)


Averaging the profits and investments over periods such as one year can disguise wide
swings in profits and assets, especially inventories and receivables. This is especially true
for seasonal businesses (such as some construction materials and toys). In such busi-
nesses it is important to understand these seasonal variations to relate quarterly and
annual figures to each other.

Return on assets (ROA), return on net assets (RONA), return on capital employed
(ROCE), and return on invested capital (ROIC) are commonly used variants of ROI.
They are also calculated using net profit as the numerator, but they have different
denominators. The relatively subtle distinctions between these metrics are beyond the
scope of this book. Some differences are found in whether payables are subtracted from
working capital and how borrowed funds and stockholder equity are treated.

10.3 Economic Profit—EVA
Economic profit has many names, some of them trademarked as “brands.” Economic
value added (EVA) is Stern-Stewart’s trademark. They deserve credit for popularizing
this measure of net operating profit after tax adjusted for the cost of capital.

Economic Profit ($) Net Operating Profit After Tax (NOPAT) ($) Cost of Capital ($)
Cost of Capital ($) Capital Employed ($) * WACC (%)
Unlike percentage measures of return (for example, ROS or ROI), Economic profit is
a dollar metric. As such, it reflects not only the “rate” of profitability, but also the size
of the business (sales and assets).

Purpose: To measure dollar profits while accounting
for required returns on capital invested.
Economic profit, sometimes called residual income, or EVA, is different from “account-
ing” profit—in that economic profit also considers the cost of invested capital—the
opportunity cost (see Figure 10.2). Like the discount rate for NPV calculations, this
charge should also account for the risk associated with the investment. A popular (and
proprietary) way of looking at economic profit is economic value added.1
Increasingly, marketers are being made aware of how some of their decisions influence
the amount of capital invested or assets employed. First, sales growth almost always


requires additional investment in fixed assets, receivable, or inventories. Economic
profit and EVA help determine whether these investments are justified by the profit
earned. Second, the marketing improvements in supply chain management and channel
coordination often show up in reduced investments in inventories and receivables.
In some cases, even if sales and profit fall, the investment reduction can be worth-
while. Economic profit is a metric that will help assess whether these trade-offs are
being made correctly.

After-Tax
Operating Profit

EVA Minus

A Charge for
Capital Used

Figure 10.2 EVA Is After-Tax Profit Minus a Charge for Capital Usage

Construction
Economic profit/economic value added can be calculated in three stages. First, deter-
mine NOPAT (net operating profit after tax). Second, calculate the cost of capital by
multiplying capital employed by the weighted average cost of capital.2 The third stage is
to subtract the cost of capital from NOPAT.
Economic Profit ($) Net Operating Profit After Tax (NOPAT) ($) Cost of Capital ($)
Cost of Capital ($) Capital Employed ($) * WACC (%)
Economic Profit: If your profits are less than the cost of capital, you have
lost value for the firm. Where economic profit is positive, value has been
generated.

EXAMPLE: A company has profits—NOPAT—of $145,000.
They have a straightforward capital structure, half of which is supplied by shareholders.
This equity expects a 12% return on the risk the shareholders are taking by investing in
this company. The other half of the capital comes from a bank at a charge of 6%:


Weighted average cost of capital (WACC) therefore
Equity (12% * 50%) Debt (6% * 50%) 9%
The company employs total capital of $1 million. Multiplying the capital employed by
the weighted average cost for the capital employed will give us an estimate of the profit
(return) required to cover the opportunity cost of capital used in the business:
Cost of Capital Capital Employed * WACC
$1,000,000 * 9%
$90,000
Economic profit is the surplus of profits over the expected return to capital.
Economic Profit NOPAT Cost of Capital
$145,000 $90,000
$55,000

Economic profit can give a different ranking for companies than does return on invest-
ment. This is especially true for companies such as Wal-Mart and Microsoft that have
experienced (achieved) high rates of growth in sales. Judging the results of the giant U.S.
retailer Wal-Mart by many conventional metrics will disguise its success. Although the
rates of return are generally good, they hardly imply the rise to dominance that the
company achieved. Economic profit reflects both Wal-Mart’s rapid sales growth and its
adequate return on the capital invested. This metric shows the magnitude of profits
after the cost of capital has been subtracted. This combines the idea of a return on
investment with a sense of volume of profits. Simply put, Wal-Mart achieved the trick of
continuing to gain decent returns on a dramatically increasing pool of capital.

10.4 Evaluating Multi-period Investments
Multi-period investments are commonly evaluated with three metrics.

Payback (#) The number of periods required to “pay back”
or “return” the initial investment.
Net Present Value (NPV) ($) The discounted value of future cash flows
minus the initial investment.
Internal Rate of Return (IRR) (%) The discount rate that results in an NPV of zero.


These three metrics are designed to deal with different aspects of the risk and returns
of multi-period projects.

Purpose: To evaluate investments with financial consequences
spanning multiple periods.
Investment is a word business people like. It has all sorts of positive connotations of future
success and wise stewardship. However, because not all investments can be pursued, those
available must be ranked against each other. Also, some investments are not attractive even
if we have enough cash to fund them. In a single period, the return on any investment is
merely the net profits produced in the time considered divided by the capital invested.
Evaluation of investments that produce returns over multiple periods requires a more
complicated analysis—one that considers both the magnitude and timing of the returns.

Payback (#): The time (usually years) required to generate the (undiscounted)
cash flow to recover the initial investment.
Net Present Value—NPV ($): The present (discounted) value of future cash inflows
minus the present value of the investment and any associated future cash outflows.
Internal Rate of Return—IRR (%): The discount rate that results in a net present
value of zero for a series of future cash flows after accounting for the initial
investment.

Construction
Payback: The years required for an investment to return the initial investment.
Projects with a shorter payback period by this analysis are regarded more favorably
because they allow the resources to be reused quickly. Also, generally speaking, the
shorter the payback period, the less uncertainty is involved in receiving the returns. Of
course the main flaw with payback period analysis is that it ignores all cash flows after
the payback period. As a consequence, projects that are attractive but that do not pro-
duce immediate returns will be penalized with this metric.

EXAMPLE: Harry is considering buying a small chain of hairdressing salons. He esti-
mates that the salons will produce a net income of $15,000 a year for at least five years.
Harry’s payback on this investment is $50,000/$15,000, or 3.33 years.


NET PRESENT VALUE
Net present value (NPV) is the discounted value of the cash flows associated with the
project.
The present value of a dollar received in a given number of periods in the future is
Cash Flow ($) * 1
Discounted Value ($)
[(1 Discount Rate (%)) ^ Period (#)]

This is easiest to see when set out in spreadsheet form.
A 10% discount rate applied to $1 received now and in each of the next three years
reduces in value over time as shown in Table 10.1.

Table 10.1 Discounting Nominal Values

Year 0 Year 1 Year 2 Year 3

Discount 1 1/(1 10%)^1 1/(1 10%)^2 1/(1 10%)^3
Formula
Discount 1 90.9% 82.6% 75.1%
Factor
Undiscounted $1.00 $1.00 $1.00 $1.00
Cash Flows
Present Value $1.00 $0.91 $0.83 $0.75

Spreadsheets make it easy to calculate the appropriate discount factors.

EXAMPLE: Harry wants to know the dollar value of his business opportunity.
Although he is confident about the success of the venture, all future cash flows have a
level of uncertainty. After receiving a friend’s advice, he decides a 10% discount rate on
future cash flows is about right.
He enters all the cash flow details into a spreadsheet (see Table 10.2).3 Harry works out
the discount factor using the formula and his discount rate of 10%:


Cash Flow * 1
Discounted Value
[(1 Discount Rate) ^ Year]

$15,000 * 1 $15,000 * 1
For Year 1 Cashflows
[(1 10%) ^ 1)] 110%

$15,000 * 90.9% 13,636

Table 10.2 Discounted Cashflow (10% Discount Rate)

Year 0 Year 1 Year 2 Year 3 Year 4 Year 5 Total

Investment ($50,000) ($50,000)
Income $15,000 $15,000 $15,000 $15,000 $15,000 $75,000
Undiscounted ($50,000) $15,000 $15,000 $15,000 $15,000 $15,000 $25,000
Cashflow
Discount 1/(1 1/(1 1/(1 1/(1 1/(1 1/(1
Formula DR)^0 DR)^1 DR)^2 DR)^3 DR)^4 DR)^5
Discount Factor 100.0% 90.9% 82.6% 75.1% 68.3% 62.1%
Present Value ($50,000) $13,636 $12,397 $11,270 $10,245 $9,314 $6,862

The NPV of Harry’s project is $6,862. Of course the NPV is lower than the sum of the
undiscounted cash flows. NPV accounts for the fact that on a per-dollar basis, cash flows
received in the future are less valuable than cash in the hand.

INTERNAL RATE OF RETURN
The internal rate of return is the percentage return made on the investment over a peri-
od of time. The internal rate of return is a feature supplied on most spreadsheets and
thus is relatively easy to calculate.
Internal Rate of Return (IRR): The discount rate for which the net present value
of the investment is zero.
The IRR is especially useful because it can be compared to a company’s hurdle rate. The
hurdle rate is the necessary percentage return to justify a project. Thus a company might
decide only to undertake projects with a return greater than 12%. Projects that have an
IRR greater than 12% get the green light; all others are thrown in the bin.


EXAMPLE: Returning to Harry, we can see that IRR is an easy calculation to perform
using a software package. Enter the values given in the relevant periods on the spread-
sheet (see Table 10.3).
Year 0—now—is when Harry makes the initial investment; each of the next five years
sees a $15,000 return. Applying the IRR function gives a return of 15.24%.

Table 10.3 Five-Year Cashflow

Cell ref A B C D E F G

1 Year 0 Year 1 Year 2 Year 3 Year 4 Year 5
2 Cashflows ($50,000) $15,000 $15,000 $15,000 $15,000 $15,000

In Microsoft Excel, the function is IRR(B2:G2)
which equals 15.24%.
The cell references in Table 10.3 should help in re-creating this function. The function is
telling Excel to perform an IRR on the range B2 (cashflow for year 0) to G2 (cashflow for year 5).

IRR AND NPV ARE RELATED
The internal rate of return is the percentage discount rate at which the net present value
of the operation is zero.
Thus companies using a hurdle rate are really saying that they will only accept projects
where the net present value is positive at the discount rate they specify as the hurdle rate.
Another way to say this is that they will accept projects only if the IRR is greater than the
hurdle rate.

Payback and IRR calculations require estimates of cash flows. The cash flows are the
monies received and paid out that are associated with the project per period, including
the initial investment. Topics that are beyond the scope of this book include the time
frame over which forecasts of cash flows are made and how to handle “terminal values”
(the value associated with the opportunity at the end of the last period).4 Net present
value calculations require the same inputs as payback and IRR, plus one other: the


discount rate. Typically, the discount rate is decided at the corporate level. This rate has
a dual purpose to compensate for the following:
■ The time value of money
■ The risk inherent in the activity
A general principle to employ is that the riskier the project, the greater the discount rate
to use. Considerations for setting the discounts rates are also beyond the scope of this
book. We will simply observe that, ideally, separate discount rates would be assessed for
each individual project because risk varies by activity. A government contract might be
a fairly certain project—not so for an investment by the same company in buying a fash-
ion retailer. The same concern occurs when companies set a single hurdle rate for all
projects assessed by IRR analysis.
Cashflows and Net Profits: In our examples cash flow equals profit, but in many
cases they will be different.

A Note for Users of Spreadsheet Programs
Microsoft Excel has an NPV calculator, which can be very useful in calculating NPV.
The formula to use is NPV(rate,value1,value2, etc.) where the rate is the discount
rate and the values are the cash flows by year, so year 1 value 1, year 2 value 2,
and so on.
The calculation starts in period one, and the cash flow for that period is discounted.
If you are using the convention of having the investment in the period before, i.e.
period 0, you should not discount it but add it back outside the formula.
Therefore Harry’s returns discounted at 10% would be
NPV(Rate, Value 1, Value 2, Value 3, Value 4, Value 5)
NPV(10%, 15000, 15000, 15000, 15000, 15000) or $56,861.80 less
the initial investment of $50,000.
This gives the NPV of $6,861.80 as demonstrated fully in the example.

10.5 Return on Marketing Investment
Return on marketing investment (ROMI) is a relatively new metric. It is not like the
other “return-on-investment” metrics because marketing is not the same kind of
investment. Instead of moneys that are “tied” up in plants and inventories, marketing
funds are typically “risked.” Marketing spending is typically expensed in the current
period. There are many variations in the way this metric has been used, and although


no authoritative sources for defining it exist, we believe the consensus of usage justi-
fies the following:

[Incremental Revenue Attributable to Marketing ($) *
Contribution Margin % Marketing Spending ($)]
Return on Marketing
Investment (ROMI) (%) Marketing Spending ($)

The idea of measuring the market’s response in terms of sales and profits is not new,
but terms such as marketing ROI and ROMI are used more frequently now than in
past periods. Usually, marketing spending will be deemed as justified if the ROMI
is positive.

Purpose: To measure the rate at which spending on marketing
contributes to profits.
Marketers are under more and more pressure to “show a return” on their activities.
However, it is often unclear exactly what this means. Certainly, marketing spending is
not an “investment” in the usual sense of the word. There is usually no tangible asset and
often not even a predictable (quantifiable) result to show for the spending, but mar-
keters still want to emphasize that their activities contribute to financial health. Some
might argue that marketing should be considered an expense and the focus should be
on whether it is a necessary expense. Marketers believe that many of their activities gen-
erate lasting results and therefore should be considered “investments” in the future of
the business.5
Return on Marketing Investment (ROMI): The contribution attributable to
marketing (net of marketing spending), divided by the marketing “invested”
or risked.

Construction
A necessary step in calculating ROMI is the estimation of the incremental sales
attributable to marketing. These incremental sales can be “total” sales attributable to
marketing or “marginal.” The following example, in Figure 10.3, should help clarify
the difference:
Y0 Baseline Sales (with $0 Marketing spending),

Y1 Sales at Marketing spending level X1, and
Y2 Sales at Marketing spending level X2,


where the difference between X1 and X2 represents the cost of an incremental marketing
budget item that is to be evaluated, such as an advertising campaign or a trade show.

1. Revenue Return to Incremental Marketing (Y2 Y1)/(X2 X1): The addi-
tional revenue generated by an incremental marketing investment, such as a spe-
cific campaign or sponsorship, divided by the cost of that marketing investment.
2. Revenue Attributable to Marketing Y2 Y0: The increase in sales attributa-
ble to the entire marketing budget (equal to sales minus baselines sales).
3. Revenue Return to Total Marketing (Y2 Y0)/(X2): The revenue attributable
to marketing divided by the marketing budget.
4. Return on Marketing Investment (ROMI) [(Y2 Y0) * Contribution
Margin (%) X2]/X2: The additional net contribution from all marketing
activities divided by the cost of those activities.
5. Return on Incremental Marketing Investment (ROIMI) [(Y2 Y1) *
Contribution Margin (%) (X2 X1)]/(X2 X1): The incremental net con-
tribution due to the incremental marketing spending divided by the amount of
incremental spending.

Y2
Y1

Sales ($)

Y0

X1 X2


Figure 10.3 Evaluating the Cost of an Incremental Marketing Budget Item

EXAMPLE: A farm equipment company was considering a direct mail campaign to
remind customers to have tractors serviced before spring planting. The campaign is expected
to cost $1,000 and to increase revenues from $45,000 to $50,000. Baseline revenues for


tractor servicing (with no marketing) were estimated at $25,000. The direct mail campaign
was in addition to the regular advertising and other marketing activities costing $6,000.
Contribution on tractor servicing revenues (after parts and labor) averages 60%.

For some industries the revenue-based metrics might be useful, but for most situations
these metrics are liable to be very misleading. ROMI or ROIMI (see following examples)
are generally more useful. However, for most situations this metric is liable to be very
misleading. There is no point in spending $20,000 on advertising to generate $100,000
of sales—a respectable 500% return to revenue—if high variable costs mean the mar-
keting only generates a contribution of $5,000.
[Revenue Attributable to Marketing * Contribution
Margin (%) Marketing Cost ($)]
Return on Marketing Investment
ROMI (%) Marketing Cost ($)

EXAMPLE: Each of the metrics in this section can be calculated from the informa-
tion in the example.
($50,000 $45,000)
Revenue Return to Incremental Marketing
($7,000 $6,000)
$5,000
= 500%
$1,000

Revenue Attributable to Marketing $50,000 $25,000 $25,000 [Note this figure
applies if the additional direct mail campaign
is used; otherwise it would be $20,000
($45,000 $25,000).]
Revenue Returns to Total Marketing $25,000/$7,000 357% [Or, if the direct mail
campaign is not used ($20,000/$6,000), 333%.]
Return on Marketing Investment (ROMI) ($25,000 * 60% $7,000)/ $7,000 114%
[Or, if the direct mail campaign is not used
($20,000 * .6 $6,000)/ $6,000 100%.]

($5,000 * 60% $1,000)
Return on Incremental Marketing Investment (ROIMI) 200%
$1,000


The first piece of information needed for marketing ROI is the cost of the marketing
campaign, program, or budget. Although defining which costs belong in marketing can
be problematic, a bigger challenge is estimating the incremental revenue, contribution,
and net profits attributable to marketing. This is similar to the distinction between base-
line and lift discussed in Section 8.1.
A further complication of estimating ROMI concerns how to deal with important
interactions between different marketing programs and campaigns. The return on
many marketing “investments” is likely to show up as an increase in the responses
received for other types of marketing. For example, if direct mail solicitations show
an increase in response because of television advertising, we could and should
calculate that those incremental revenues had something to do with the TV cam-
paign. As an interaction, however, the return on advertising would depend on what
was being spent on other programs. The function is not a simple linear return to the
campaign costs.
For budgeting, one key element to recognize is that maximizing the ROMI would
probably reduce spending and profits. Marketers typically encounter diminishing
returns, in which each incremental dollar will yield lower and lower incremental
ROMI, and so low levels of spending will tend to have very high return rates.
Maximizing ROMI might lead to reduced marketing and eliminating campaigns or
activities that are, on balance, profitable, even if the return rates are not as high. This
issue is similar to the distinction between ROI (%) and EVA ($) discussed in Sections
10.2 and 10.3. Additional marketing activities or campaigns that bring down average
percentage returns but increase overall profits can be quite sensible. So, using ROMI
or any percentage measure of profit to determine overall budgets is questionable. Of
course, merely eliminating programs with a negative ROMI is almost always a
good idea.
The previous discussion intentionally does not deal with carryover effect, that is, mar-
keting effects on sales and profits that extend into future periods. When marketing
spending is expected to have effects beyond the current period, other techniques will be
needed. These include payback, net presented value, and internal rate of return. Also, see
customer lifetime value (Section 5.3) for a more disaggregated approach to evaluating
marketing spending designed to acquire long-lived customer relationships.

Related Metrics
Media Exposure Return on Marketing Investment: In an attempt to evaluate the value
of marketing activities such as sponsorships, marketers often commission research to
gauge the number and quality of media exposures achieved. These exposures are then


valued (often using “rate cards” to determine the cost of equivalent advertising
space/time) and a “return” is calculated by dividing the estimated value by the costs.
(Estimated Value of Media Exposures Achieved ($)
Cost of Marketing Campaign, Sponsorship,
or Promotion ($))
Media Exposure Return on Marketing
Investment (MEROMI) (%) Cost of Marketing Campaign, Sponsorship,
or Promotion ($)

This is most appropriate where there isn’t a clear market rate for the results of the cam-
paign and so marketers want to be able to illustrate the equivalent cost for the result for
a type of campaign that has an established market rate.

EXAMPLE: A travel portal decides to sponsor a car at a Formula 1 event. They assume
that the logo they put on the car will gain the equivalent of 500,000 impressions and will
cost 10,000,000 yen. The cost per impression is thus 10 million yen/500,000 = or 20 yen
per impression. This can be compared to the costs of other marketing campaigns.

Hawkins, D. I., Roger J. Best, and Charles M. Lillis. (1987). “The Nature and Measurement of
Marketing Productivity in Consumer Durables Industries: A Firm Level Analysis,” Journal of the
Academy of Marketing Science, 1(4), 1–8.


11
THE MARKETING
METRICS X-RAY

11.1 The Marketing Metrics X-Ray
Our purpose in this chapter is to give some examples of how marketing metrics can
augment and complement traditional financial metrics when used to assess firm and
brand performance. In particular, marketing metrics can serve as leading indicators of
problems, opportunities, and future financial performance. Just as x-rays (now MRIs)
are designed to provide deeper views of our bodies, marketing metrics can show prob-
lems (and opportunities) that would otherwise be missed.

Put Your Money Where Your Metrics Are
Table 11.1 shows common summary financial information for two hypothetical compa-
nies, Boom and Cruise. Income statement data from five years provide the basis for
comparing the companies on several dimensions.

ON WHICH FIRM WOULD YOU BET YOUR GRANDPARENTS’ SAVINGS?
We have used this example with MBA students and executives many times—usually, we
ask them “Assume that your grandparent wants to buy a partnership in one of these
firms, using limited retirement savings. If these financial statements were the only data
you had available or could obtain, which firm would you recommend?” These data are
the metrics traditionally used to evaluate firm performance.
The table shows that gross margins and profits are the same for both firms. Although
Boom’s sales and marketing spending are growing faster, its return on sales (ROS) and
return on investment (ROI) are declining. If this decline continues, Boom will be in
trouble. In addition, Boom’s marketing/sales ratio is increasing faster than Cruise’s. Is
this a sign of inefficient marketing?

357

Table 11.1 Financial Statements

Boom
All $ in (Thousands) Year 1 Year 2 Year 3 Year 4 Year 5

Revenue $833 $1,167 $1,700 $2,553 $3,919

Margin Before Marketing $125 $175 $255 $383 $588

Marketing $100 $150 $230 $358 $563

Profit $25 $25 $25 $25 $25

Margin (%) 15% 15% 15% 15% 15%

Marketing/Sales 12% 13% 14% 14% 14%

ROS 3.0% 2.1% 1.5% 1.0% 0.6%

Year on Year Revenue Growth — 40% 46% 50% 53%

CAGR Revenue from Year 1 — 40% 43% 45% 47%

Invested Capital $500 $520 $552 $603 $685

ROI 5.0% 4.8% 4.8% 4.1% 3.6%

Cruise
All $ in (Thousands) Year 1 Year 2 Year 3 Year 4 Year 5

Revenue $1,320 $1,385 $1,463 $1,557 $1,670

Margin Before Marketing $198 $208 $219 $234 $251

Marketing $173 $183 $194 $209 $226

Profit $25 $25 $25 $25 $25

Margin (%) 15% 15% 15% 15% 15%

Marketing/Sales 13% 13% 13% 13% 14%

ROS 1.9% 1.8% 1.7% 1.6% 1.5%

Year on Year Revenue Growth — 5% 6% 6% 7%

CAGR Revenue from Year 1 — 5% 5% 6% 6%

Invested Capital $500 $501 $503 $505 $507

ROI 5.0% 5.0% 5.0% 5.0% 4.9%


On the basis of the information in Table 11.1, most people chose Cruise. Cruise is doing
more with less. It’s more efficient. Its trend in ROS looks much better, and Cruise has
maintained a fairly consistent ROI of about 5%. About the only thing Boom has going
for it is size and growth of the “top line” (sales revenue). Let’s look deeper at the mar-
keting metrics x-ray.

USING THE MARKETING METRICS X-RAY
Table 11.2 presents the results of our marketing metrics x-ray of Boom and Cruise. It
shows the number of customers each firm is serving and separates these into “old”
(existing customers) and “new” customers.
This table allows us to see not only the rate at which the firm acquired new customers
but also their retention (loyalty) rates. Now, Boom’s spending on marketing looks a
lot better because we now know that spending was used to generate new customers
and keep old ones. In addition, Boom acquires new customers at a lower cost than
Cruise. And although Cruise’s customers spend more, Boom’s stay around longer.
Perhaps we should order another set of x-rays to examine customer profitability and
lifetime value?
Table 11.3 uses the information in the previous table to calculate some additional cus-
tomer metrics. Under an assumption of constant margins and retention rates and a
15% discount rate, we can calculate the customer lifetime value (CLV) for the cus-
tomers of each firm and compare this CLV with what the firms are spending to acquire
the customers. The CLV represents the discounted margins a firm will earn from its
customers over their life buying from the firm. Refer to Section 5.3 for details about the
estimation of CLV and the process for using the number to value the customer base as
an asset. The asset value is merely the number of ending customers times their remain-
ing lifetime value (CLV minus the just-received margin). For these examples, we have
assumed that all marketing is used to acquire new customers, so the customer acquisi-
tion cost is obtained by dividing marketing spending by the new customers in year
period.
Boom’s aggressive marketing spending looks even better in this light. The difference
between the CLV and acquisition cost is only $3.71 for Cruise but is $48.21 for Boom.
From the viewpoint of the customer asset value at the end of year five, Boom is worth
almost five times as much as Cruise.
Table 11.4 gives us even more information on customers. Customer satisfaction is much
higher for Boom, and Boom’s customers are more willing to recommend the firm to
others. As a consequence, we might expect Boom’s acquisition costs to decline in the
future. In fact, with such a stable and satisfied customer base, we could expect that
brand equity (refer to Section 4.4) measures would be higher too.

Chapter 11 The Marketing Metrics X-Ray 359

360
MARKETING METRICS

Table 11.2 Marketing Metrics

Boom Cruise
Year 1 Year 2 Year 3 Year 4 Year 5 Year 1 Year 2 Year 3 Year 4 Year 5
New Customers (Thousands) 1.33 2.00 3.07 4.77 7.50 1.86 1.97 2.09 2.24 2.43
Total Customers (Thousands) 3.33 4.67 6.80 10.21 15.67 3.86 4.05 4.28 4.55 4.88
Sales/Customer $250 $250 $250 $250 $250 $342 $342 $342 $342 $342
Marketing/New Customer $75 $75 $75 $75 $75 $93 $93 $93 $93 $93
Retention Rate — 80% 80% 80% 80% — 54% 54% 54% 54%

Table 11.3 Customer Profitability

Customer Value Metric Boom Cruise

Customer CLV $123.21 $96.71

Customer Acquisition Cost $75.00 $93.00

Customer Count (Thousands) 15.67 4.88

Customer Asset Value (Thousands) $1,344 $222


Table 11.4 Customer Attitudes and Awareness

Boom Cruise
Year 1 Year 2 Year 3 Year 4 Year 5 Year 1 Year 2 Year 3 Year 4 Year 5
Chapter 11 The Marketing Metrics X-Ray

Awareness 30% 32% .31% 31% 33% 20% 22% 22% 23% 23%
Top of Mind 17% 18% 20% 19% 20% 12% 12% 11% 11% 10%
Satisfaction 85% 86% 86% 87% 88% 50% 52% 52% 51% 53%
Willingness to Recommend 65% 66% 68% 67% 69% 42% 43% 42% 40% 39%
361


Hiding Problems in the Marketing Baggage?
The income statement for another example firm, Prestige Luggage, is depicted in
Table 11.5. The company seems to be doing quite well. Unit and dollar sales are growing
rapidly. Margins before marketing are stable and quite robust. Marketing spending and
marketing to sales ratios are growing, but so is the bottom line. So what is not to like?

Table 11.5 Prestige Luggage Income

Statement
Sales Revenue (Thousands) $14,360 $18,320 $23,500 $30,100
Unit Sales (Thousands) 85 115 159 213
Market Share (Unit) 14% 17% 21% 26%
Gross Margin 53% 53% 52% 52%
Marketing $1,600 $2,143 $2,769 $3,755
Profit $4,011 $5,317 $7,051 $9,227
ROS 27.9% 29.0% 30.0% 30.7%
Marketing/Sales 11.1% 11.7% 11.8% 12.5%

USING THE MARKETING METRICS X-RAY
Let’s take a deeper look at what’s going on with Prestige Luggage by examining their
retail customers. When we do, we’ll get a better view of the marketing mechanics that
underlie the seemingly pleasant financials in Table 11.5.
Table 11.6 (refer to Section 6.6 for distribution measures) shows that Prestige Luggage’s
sales growth comes from two sources: an expanding number of outlets stocking the
brand and an increase (more than four-fold) in price promotions. Still, there are plenty
of outlets that do not stock the brand. So there may be room to grow.
Table 11.7 reveals that although the overall sales are increasing, they are not keeping
pace with the number of stores stocking the brand. (Sales per retail store are already
declining.) Also, the promotional pricing by the manufacturer seems to be encouraging
individual stores’ inventories to grow. Soon, retailers may become irritated that the
GMROII (gross margin return on inventory investment) has declined considerably.
Future sales may continue to slow further and put pressure on retail margins. If retailer
dissatisfaction causes some retailers to drop the brand from their assortment, manufac-
turer sales will decline precipitously.


Table 11.6 Prestige Luggage Marketing and Channel Metrics

Retail Dollar Sales (Thousands) $24,384 $27,577 $33,067 $44,254
Retail Unit Sales (Thousands) 87 103 132 183
Number Stocking Outlets 300 450 650 900
Price Premium 30.0% 22.3% 15.1% 8.9%
ACV Distribution2 30% 40% 48% 60%
% Sales on Deal 10% 13% 20% 38%
Advertising Spending (Thousands) $700 $693 $707 $721
Promotion Spending (Thousands) $500 $750 $1,163 $2,034

Table 11.7 Luggage Manufacturer Retail Profitability Metrics

Retail Margin $ $9,754 $11,169 $13,557 $18,366
Retail Margin % 40% 41% 41% 42%
Retail Inventory (Thousands) 15 27 54 84
Inventory Per Store 50 60 83 93
Sales/Outlet (Thousands) $81 $61 $51 $49
Stores per Point of AVC % 10 11 14 15
GMROII 385% 260% 170% 155%

In addition, the broadening of distribution and the increase of sales on deal suggest a
possible change in how potential consumers view the previously exclusive image of the
Prestige Luggage brand. The firm might want to order another set of x-rays to see if and
how consumer attitudes about the brand have changed. Again, if these changes are by
design, then maybe Prestige Luggage is okay. If not, then Prestige Luggage should be
worried that its established strategy is falling apart. Add that to the possibility that some
retailers are using deep discounts to unload inventory after they’ve dropped the brand,
and suddenly Prestige Luggage faces a vicious cycle from which they may never recover.
Some things you can’t make up, and this example is one. The actual company was
“pumped up” through a series of price promotions, distribution was expanded, and
sales grew rapidly. Shortly after being bought by another company looking to add to
their luxury goods portfolio of brands, the strategy unraveled. Many stores dropped the
line, and it took years to rebuild the brand and sales.


These two examples illustrate the importance of digging behind the financial statements
using tools such as the marketing x-ray. More numbers, in and of themselves, are only
part of the answer. The ability to see patterns and meaning behind the numbers is even
more important.

Smoking More But Enjoying It Less?
Table 11.8 displays marketing metrics reported by a major consumer-products company
aimed at analyzing the trends in competition by lower-priced discount brands. A declin-
ing market size, stagnant company market share, and a growing share of firm sales
accounted for by discount brands all made up a baleful picture of the future. The firm
was replacing premium sales with discount brand sales. To top it off, the advertising and
promotion budgets had almost doubled. In the words of Erv Shames, Darden Professor,
it would be easy to conclude that the marketers had “run out of ideas” and were resort-
ing to the bluntest of instruments: price.

Table 11.8 Market Trends for Discount Brands and Spending; Big Tobacco Company

Year 1987 1992
Market Size (Units) 4,000 3,850
Company Unit Share 25% 24%
Unit Sales 1000 924
Premium Brand Units 925 774
Discount Brand Units 75 150
Advertising & Promotion Spend $600 $1,225

Table 11.9 Additional Metrics

Year 1987 1992
Revenue (Thousands) $1,455 $2,237
Average Unit Price $1.46 $2.42
Average Premium Price $1.50 $2.60
Average Discount Price $0.90 $1.50
Operating Profit (Thousands) $355 $550


The picture looks much brighter, however, after examining the metrics in Table 11.9. It
turns out that in the same five years during which discount brands had become more
prominent, sales revenue and operating income had both grown by over 50%. The rea-
son is clear: Prices had almost doubled, even though a large portion of these price
increases had been “discounted back” through promotions. Overall, the net impact was
positive on the firm’s bottom line.
Now you might be thinking that the messages in Table 11.9 are so obvious that no one
would ever find the metrics in Table 11.8 to be as troubling as we made them out to be. In
fact, our experience in teaching a case that contains all these metrics is that experienced
marketers from all over the world tend to focus on the metrics in Table 11.8 and pay little
or no attention to the additional metrics—even when given the same level of prominence.
The situation described by the two tables is a close approximation to the actual market
conditions just before the now-famous “Marlboro Friday.” Top management took action
because they were concerned that the series of price increases that led to the attractive
financials in 1992 would not be sustainable because the higher premium prices gave
competitive discount brands more latitude to cut prices. On what later became known as
“Marlboro Friday,” the second of April 1993, Phillip Morris cut Marlboro prices by $0.40
a pack, reducing operating earnings by almost 40%. The stock price tumbled by 25%.
Note in this example the contrast from the preceding example. Prestige Luggage was
increasing promotion expenditures to expand distribution. Prices were falling while
promotion, or sales on deal, were increasing—an ominous sign. With Marlboro, they
were constantly raising the price and then discounting back—a very different strategy.

Marketing Dashboards
The presentation of metrics in the form of management “dashboards” has received a
substantial amount of attention in the last several years. The basic notion seems to be
that the manner of presenting complex data can influence management’s ability to rec-
ognize key patterns and trends. Would a dashboard, a graphical depiction of the same
information, make it easier for managers to pick up the ominous trends?
The metaphor of an automobile dashboard is appropriate because there are numerous
metrics that could be used to measure a car’s operation.The dashboard is to provide a reduced
set of the vital measures in a form that is easy for the operator to interpret and use. Un-
fortunately,although all automobiles have the same key metrics, it is not as universal across
all businesses. The set of appropriate and critical measures may differ across businesses.
Figure 11.1 presents a dashboard of five critical measures over time. It reveals
strong sales growth while maintaining margins even though selling less expensive
items. Disturbingly, however, the returns for the retailer (GMROII) have fallen precipi-
tously while store inventories have grown. Sales per store have similarly dropped.
The price premium that Prestige Luggage can command has fallen, and more of the


Revenue and Margins

Prestige Luggage Headline Financials
$35,000 60%
Revenue (thousands)

$30,000 58%
56%

Gross Margin
$25,000 54%
◆ ◆
$20,000 ◆ ◆ 52%
50%
$15,000 48%
$10,000 46%
44%
$5,000 42%
$0 40%

The financial metrics look healthy; revenue showing good growth while margins are almost unchanged.
Manufacturer Prices to Store Prices Store Inventory and GMROII

Prestige Luggage Prices and Retail Prices 100 Inventory Per Store 450%
90 ◆ GMROII
$330 ◆ 400%
80 350%
Avg. Retail
$280 ■ 70 300%
■ Price
Average Price

■ 60 ◆
■ 250%
$230 50
200%
40 ◆
$180 ◆ 150%
◆ ◆
30
Avg. ◆ ◆ 20 100%
$130
Prestige 10 50%
Price
$80 0 0%
Year 1 Year 2 Year 3 Year 4 Year 1 Year 2 Year 3 Year 4

Prestige Luggage is selling less expensive items. Prestige Luggage is making diminishing returns for retailer.
Distribution Pricing and Promotions

Prestige Luggage Store Distribution Prestige Luggage Pricing and Promotion

35%
$90
30% ◆ Year 1
Year 1
$80 25%
Price Premium

■ Year 2
$70 20%
Year 2
$60 15% ▲ Year 3
Year 3 Year 4 Year 4
10%
$50 ¥
5%
$40
0%
$30 0% 10% 20% 30% 40%
Sales per Store
% on Deal

We are moving into smaller stores. Prestige Luggage is becoming reliant on promotion.

Figure 11.1 Prestige Luggage: Marketing Management Dashboard


company’s sales are on deal. This should be a foreboding picture for the company and
should raise concerns about the ability to maintain distribution.

Summary: Marketing Metrics Financial Metrics Deeper Insight
Dashboards, scorecards, and what we have termed “x-rays” are collections of marketing
and financial metrics that management believes are important indicators of business
health. Dashboards are designed to provide depth of marketing understanding con-
cerning the business. There are many specific metrics that may be considered impor-
tant, or even critical, in any given marketing context. We do not believe it is generally
possible to provide unambiguous advice on which metrics are most important or which
management decisions are contingent on the values and trends in certain metrics. These
recommendations would have be of the “if, then” form, such as “If relative share is
greater than 1.0 and market growth is higher than change in GDP, then invest more in
advertising.” Although such advice might be valuable under many circumstances, our
aims were more modest—simply to provide a resource for marketers to achieve a deeper
understanding of the diversity of metrics that exist.
Our examples, Boom versus Cruise, Prestige Luggage, and Big Tobacco, showed how
selected marketing metrics could give deeper insights into the financial future of compa-
nies. In situations such as these, it is important that a full array of marketing and finan-
cial metrics inform the decision. Examining a complete set of x-rays does not necessarily
make the decisions any easier (the Big Tobacco example is debated by knowledgeable
industry observers to this day!), but it does help ensure a more comprehensive diagnosis.

Ambler, Tim, Flora Kokkinaki, and Stefano Puntonni (2004). “Assessing Marketing
Performance: Reason for Metric Selection,” Journal of Marketing Management, 20,
pp. 475–498.
McGovern, Gail, David Court, John A. Quelch, and Blair Crawford (2004). “Bringing
Customers into the Boardroom,” Harvard Business Review, November, pp. 1–10.
Meyer, C. (1994). “How the Right Measures Help Teams Excel,” Harvard Business Review.
72(3), 95.


12
SYSTEM OF METRICS

“There are three kinds of economists: those who can count and those who can’t.”
—Unknown source

Modeling Firm Performance
To better understand the factors contributing to overall firm success, managers and ana-
lysts often decompose return on assets (ROA) into the product of two ratios with each
ratio reflecting a different aspect of the business. One popular approach or “model” for
this decomposition is the DuPont Model.
Net Profit Sales
ROA =
Sales Assets

The first ratio in this simplified DuPont Model is called either the profit margin or
return on sales. It measures how profitable is each dollar of sales. To the extent that mar-
keters create products that customers value, claim that value through intelligent pricing,
drive down costs by paying attention to manufacturing and channel costs, and optimize
their marketing spending, marketers can increase the firm’s return on sales. The second
ratio in the DuPont Model is known as asset turnover. Asset turnover can be thought of
as the number of dollars of sales each dollar of assets generates. Here the job of mar-
keters is even more focused—on generating dollars of sales but with an eye toward man-
aging assets such as inventory and receivables captured in the denominator.
Notice that the DuPont Model is an identity.1 It is always true regardless of the values
taken on by the various ratios. It is always true mostly because we have defined the ratios
in such a way so as to make it always true. So it makes no sense to argue with or take
exception to the DuPont Model.
But if it is simply an equation that is true by definition, what good is it?

369

It is useful to the extent that the decomposition of ROA into the two component ratios
helps firms maximize ROA by focusing (separately) on the two components. It is also
useful in that it reminds marketers that their job is not simply to generate sales, but to
generate profitable sales and to do so efficiently (with respect to assets used).
The DuPont Model has demonstrated its usefulness in practice. A Google search
resulted in 4.4 million results for “DuPont Model” compared to 2.9 million results for
“DuPont Chemicals.” In some circles, the company is now more famous for its model
than its chemicals.
Figure 12.1 illustrates how the DuPont Model is often expanded to include components
affecting the two input ratios.

The DuPont Model

Cost of Goods
Sales
Sold
Measures of the
S, G & A -
effectiveness with Expenses
which assets are Total Net
Interest Costs Profit
used to produce Expense
revenues. -/-
Income Net Profit
Taxes Sales to Sales

Cash

Accounts x Return on
Measures of Receivable Assets
investments in
Current
working capital assets Inventories Assets
needed for sustaining
Marketable
ongoing operations. Securities Sales to
Sales
Assets
Other -/-
Total
+ Assets

Land

Measures of Buildings
investments in long- Non-Current
Assets
term, revenue- Machinery and
producing assets. Equipment

Intangibles

Figure 12.1 An Extended DuPont Model (adapted from http://www.12manage.com/
methods_dupont_model.html)

Notice that the three rightmost columns of boxes in Figure 12.1 represent the DuPont
Model. The two leftmost columns of boxes represent a particular method of breaking
apart net profit and total assets into smaller components. Our purpose here is not to


critique the above representation of the components of firm performance, but simply to
offer a few observations. First, we note that the decompositions of total costs, current
assets, and non-current assets should be familiar to most readers. The categories of
components used are consistent with what one finds on income statements (where total
cost appears) and balance sheets (where total assets appears). Second, we note that the
assets that marketing creates (brands and customer relationships, for example) get
lumped together as intangibles signaling that they are difficult to measure (which we
agree with) and perhaps an afterthought or “other” category (which we disagree with).
Finally, and most importantly, we observe that although total cost, current assets, and
non-current assets all get broken out into smaller, well-understood components, sales
does not. It is as if Costs and Assets deserve a lot of attention but the components of sales
do not. This is perhaps not surprising since this particular model was designed by
finance and accounting executives. As marketers, however, much of our focus is on how
sales are generated. We also care about costs and asset utilization, of course, but we care
more about sales and the components of sales. Figure 12.1 reflects the inward focus of a
firm whose success depended on making things, minimizing costs, and using assets effi-
ciently. For today’s firms whose success depends at least as much on marketing and sales
as production, we need a different model. We need our own “DuPont Model,” with at
least the same amount of detail and clarity for breaking down the components of sales
as the commonly used breakdowns of costs and assets.
Of course, as we begin to think about how to break down sales into its components, we
quickly come to understand why there is no commonly used breakdown across different
types of businesses. As all marketers know, there are multiple ways to decompose or
break down sales simply because several entities (most of them outside the firm) are
involved in the creation of revenue: sales force, customers, dealers, and even our com-
petition. With a multitude of insightful ways to break down sales, it is no wonder there
is not one commonly accepted way.
To illustrate, Figure 12.2 shows four (of many) separate and valid ways to break down
sales into smaller components.
■ SALES = Number Salespersons * Avg. Sales/Salesperson
■ SALES = Number dealers ACV% * Avg. Sales per dealer ACV%
■ SALES = Our Dollar Share * Total Market Sales
■ SALES = Number customers * Sales per customer
As with the DuPont Model, each of four ways to compute sales is an identity. Sales will
always equal the number of customers times the average sales per customer. But even
though they are identities, they can still lead to valuable insights as we will attempt to
demonstrate.

Chapter 12 System of Metrics 371

Number
Dealers
Number ACV%
Salespersons
Avg. Sales
Average per Dealer
Sales/Salesperson ACV%

Sales

Our
Number Dollar
Customers Share

Market
Sales per
Sales
Customer

Figure 12.2 A Sales Model

We also point out that there will be other ways to break down sales. Figure 12.2 simply
illustrates four ways. Also know that the outer ring of components of sales in Figure 12.2
can themselves be decomposed. For example: Sales per customer can be calculated as
Purchases per customer (per period) * Average Sales per purchase. And, not unexpect-
edly, there will be multiple ways to decompose each of the outer-ring components. For
example Sales per customer can also be decomposed into Units purchased per customer
* Average price per unit. Decomposing the components of sales can be thought of as
expanding the diagram in Figure 12.2 outward. We might also think of expanding the
model “upward” with separate pages (decompositions) for each product or each cus-
tomer group or each vendor.

Three Reasons for Using Systems of Identities in Marketing
There are three primary reasons for formulating marketing DuPont-like component
models of your marketing decisions and objectives:

1. Decomposing the metric of interest into components can make it possible to
identify problems and opportunities for improvement in more detail. For exam-
ple, did share drop because our sales were down or competitors’ sales were up?
If our sales were down, was that due to fewer customers buying, lower unit sales
per customer, lower average prices, or some combination? Decomposition may


also help by separating identities from empirical relationships. Although identi-
ties are easy (just arithmetic), empirical relationships require difficult judg-
ments about the form of the relationship, causality, and the future.
2. Decomposing metrics may also allow us to estimate, indirectly, other compo-
nent metrics that are difficult to measure directly. Using multiple identities can
help eliminate measurement error with multiple “checks” on the value of any
specific metrics. In the same way, individual marketing metrics may be regarded
as part of a network or “web” of relationships. If each link in the network is
valid, even if individual values are estimated with error, the entire structure will
be more robust.
3. Selecting and organizing the right network of marketing metrics often helps
formulate models of marketing mix decisions. Like the DuPont Model, using
models with interim components can make such models and dashboards more
managerially transparent and help managers make and monitor the effects of
their decisions.

Decomposing for Diagnostic Purposes
As mentioned previously, a primary purpose for using one or more identities to decom-
pose any marketing metric of interest is to gain a deeper understanding (or at least a dif-
ferent perspective) on the reasons for changes and differences observed. Although
identities may be developed with a view to understanding the sources of changes and
differences, they do not require calibration or estimation. They are true by definition,
and we will designate these with an (ID).
An example of an identity is the relationship between Sales, Quantity, and Price:
Sales = Quantity × Price (ID)
This identity tells us that Sales declines whenever quantity decreases (as a percentage)
more than price increases. If we witness declining sales, the identity helps us see, first,
whether the decline was due to declining quantity or price or both. And next it helps us
understand that if quantity declined, price increased, but sales declined, that quantity
must have declined by a larger percentage than did the price increase.
In contrast to identities are empirical relationships—relationships between variables for
which the exact equation is not known and/or for which the relationship holds only
imperfectly. Empirical relationships are required, for example, to help us decide whether
we should increase or decrease prices. We designate these with an (EM). For example,
we might consider the relationship between quantity sold to be a direct, linear function
of price charged:
Quantity = b × Price + error (EM)


This empirical relationship between quantity and price necessarily contains an error to
account for measuring price or quantity imperfectly or influences on quantity sold
other than price (our competitors’ prices, for example). Also note that the parameter “b”
in this empirical relationship is, itself, a variable. It is an unknown constant—one that
we might, for example, be able to estimate from available data. But one of the key dif-
ferences between identities (ID) and empirical relationships (EM) is that empirical rela-
tionships are more flexible. They apply to the tough and important questions such as
“how many more units will we sell if we lower the price by $1?”
Dashboards of metrics often reflect underlying management logic about how marketing
works to influence sales and profits. Dashboards include both identities and empirical
relationships. As illustrated in Figure 12.2, sales can be decomposed many ways. Some
of the components of sales might themselves be decomposed using one or more identi-
ties. Each firm needs to identify its primary performance measures. This is what should
appear on their dashboards. There should be the capability to drill down on each of
these performance measures (using identities) to diagnose and explain changes across
time. But if dashboards are to be more than monitoring devices, we should have some
idea of causal connections (step on the brake to slow down the vehicle, step on the accel-
erator to make it go faster). Before long, they can become complicated as we start to take
into consideration the multiple effects of some of the variables, e.g., step on the acceler-
ator to make the car go faster and the fuel gauge drops. Sometimes we also need a sys-
tem of metrics to help infer (or forecast) values that are difficult to measure directly
(e.g., how much farther can we drive before the gas tank is empty?).

Eliminating Error by Harnessing the Law of Large (and Not So Large)
Numbers
There is the classic story of the physics professor whose final exam asked students to
explain how to use a barometer to measure the height of a building. In addition to the
“obvious” answer to measure the barometric pressures at the top and bottom of the
building and use the difference to calculate the building’s height, the professor purport-
edly received several other creative answers. Drop the barometer from the top of the
building, time how long it takes to hit the ground, and use the appropriate physics for-
mula to infer the height. Tie the barometer to a string, lower it to the ground, and meas-
ure the length of the string. Measure the length of the shadow cast by the building, the
length of the shadow cast by the barometer, the height of the barometer, and use pro-
portions to calculate the height of the building. By far the most creative solution pur-
portedly offered was to knock on the door of the building’s janitor and offer to give the
janitor the barometer in exchange for revealing the height of the building.
The multiple ways to calculate sales shown in Figure 12.2 are similar to the multiple
ways students came up with to measure the height of the building. Rather than argue


over which single method to use, we propose to look for a way to use them all. When
faced with a dilemma of which of two methods to use, why not do both? For the barom-
eter problem, why not use several different methods and then combine the many esti-
mates into one final estimate—perhaps by doing something as simple as taking the
average of the estimates. If we wanted to do a little bit better, we could calculate a
weighted average with weights depending on some measure of how “accurate” each esti-
mate was. We might put more weight on the string-based estimate and less on the esti-
mate from timing the fall of the barometer if we thought our watch and wind made the
timing-based estimate less accurate. The relative weight to put on the janitor’s estimate
would depend on our confidence in the estimate. If the janitor claims to “know” the
height, we should give the estimate more weight than if the janitor admits the number
is something of a guess.
Using the average of the estimates instead of any one of the estimates takes advantage of
the law of large (and not so large) numbers. The average is expected to be closer to the
true value and become closer the more estimates that we have to average together.
Ideally we want “independent” estimates such as might be the case with the barometer
example (unless, of course, the janitor got his number using the string method).
In the barometer example, we were mostly interested in measuring the height of the
building. In our example, marketers are probably just as interested in the measuring
components as we are in measuring sales itself. In fact, it often is the case that the firm
has a good handle on sales and would like to get a better measure of some of the com-
ponents such as share or the sales per customer or any of the other metrics in the outer
ring or outer-outer ring. In extreme cases, the firm may have no separate measure of
one of the components and will have to “back into it” based on the measurements of all
the others. (In the barometer example, use the height of the building and the length of
the barometer’s shadow to estimate the length of the building’s shadow—to measure
how far away the building is, for example, without having to travel to the building.)
What this means is that every initial estimate (and the associated standard deviation)
will combine to determine our final estimates. Our estimate of the length of the string
will be used to help revise our estimate of the time it took the barometer to hit the
ground and the length of the building’s shadow and vice versa. We think it is easy to see
that the more separate estimates and identities we have in the model, the more confi-
dent we are with the final estimate.
Whereas the carpenter adage is to measure twice and cut once, here we say measure
many times and many ways and put them all together in a systematic, logical way. Use
not only a square to check for a right angle, also measure 3 feet and 4 feet along each
side and check to see if the diagonal measures 5 feet. That’s the idea behind the pro-
posed process for fine tuning a system of marketing metrics. (See Appendix 1 at the end
of this chapter for a numerical example.)


Using Identities to Estimate Metrics that are Difficult to Measure
Directly
“Decomposition involves figuring out how to compute something very uncertain from
other things that are a lot less uncertain or at least easier to measure.” (Hubbard,
2007)

Marketing models can often make use of our ability to infer missing variables through
construction of the appropriate identity. First, let’s take an example from the physical
world and use that to draw a parallel to marketing problems. If you wanted to calculate
directly the average depth of your local swimming pool, that would involve a series of
complicated and difficult measurements (either measuring the depth repeatedly as one
moved across the length and width of the pool or somehow capturing the curve of the
bottom with a functional form and using calculus and algebra). An indirect method
might be easier. Record the volume of water required to fill the pool and divide by the
pool’s surface area.
Marketers are also often interested in estimating the values that are conceivably directly
measurable, yet might be more efficiently estimated from combinations of other met-
rics. An example is a firm’s average Share of Requirements or Share of Wallet either in
dollars or in units. To measure this directly would require a database of customer pur-
chases that included its own firm purchases and all other purchases in the same cate-
gory. Further, the customers included in the database would need to be representative of
the entire category or weighted in an appropriate way. Instead of a direct measurement,
marketers might find it easier and more efficient to estimate share of requirements from
the equation included in Sections 2.3 and 2.5:
Market Share (%)
(Penetration Share (dollars or units) * Heavy Usage Index
(dollars or units))

The latter three variables might be directly measurable from reported sales, a count of
known customers, and an estimate of the degree to which the firm’s own customers are
heavy or light users of the category. Of course, the metric estimated in this manner is an
average and will not give insight into the variation in customer loyalty behavior repre-
sented by the metric.

Marketing Mix Models—Monitoring Relationships between
Marketing Decisions and Objectives
As Neil Borden, Sr., the author of the term “marketing mix” noted a half-century ago,
“Several characteristics of the marketing environment make it difficult to predict and


control the effect of marketing actions.”2 A system of marketing identities can help with
this problem by providing integrated frameworks and structures for monitoring the
outcomes from marketing decisions. Marketing models must often trade off compre-
hensiveness with comprehensibility; completeness with simplicity.
The complexities include these: First several potential marketing actions may affect sales
and profits. These potential actions include pricing, price promotion, advertising, per-
sonal selling, and distribution changes, to name just a few. Second, the effects of any one
of these actions on sales, even holding all of the other actions equal, are often non-
linear. The infamous S-curve is an example of this non-linearity (a little advertising pro-
duces no effect, somewhat more stimulates sales, and at some point effectiveness
diminishes and disappears altogether). Third, the effects of one marketing decision
often depend upon other marketing decisions. For example, the effects of advertising on
sales depend not only on the product design, but also on price and product availability.
Fourth, there are also “feedback” and lagged effects in marketing. Over time, our invest-
ments in advertising might build brand equity that allows our brand to charge higher
prices. Or, if competitors introduce a better product and sales fall to the point that sales-
people are earning too little, the same salespeople may resign or spend less time on a
particular product line, causing sales to fall again. The potential complexity resulting
from specifying a large number of marketing mix elements, non-linearities of effects,
interactions among elements, lagged and feedback effects, and competitive behavior is
mind-numbing. Further, these potential complexities seem to be limited only by the
imagination—and marketing people are (by definition?) creative! It is simply not possi-
ble, we assert, to capture all of these complexities with any empirical model.
In the face of such potential for complexity, it is important that marketers find
approaches that will help them, in the words of Arnold Zellner, keep it sophisticatedly
simple (KISS—we know you thought it stood for something else).3 Careful selection of
marketing metrics frameworks that are constructed around a few important identities
has several benefits. One is that they enable us to specify the most important interac-
tions and feedback loops at the level of structural identities instead of empirical
relationships.
Let’s begin by distinguishing between marketing decisions (actions), objectives (for
example, profits), and intervening metrics that help us understand the connections. A
simple marketing mix model might be the following: profits = f (unit price, advertising,
sales force, and trade promotion), which written out in English means profits are a
function of unit price, advertising, sales force, and trade promotion (see Figure 12.3).
Many marketers would reject the model in Figure 12.3 as not sufficiently detailed as
concerns the multiple effects of marketing mix decisions. A $1 increase in Unit Price, for
example, would result in a $1 increase in unit margin while, probably, decreasing unit
sales. Estimating the empirical relationship between unit price and unit sales separately,
and then making use of identities involving unit price, unit cost, and unit sales to


calculate gross profit (as illustrated in Figure 12.4), is generally preferred Thus, we sep-
arate what can be calculated (using an identity) from what must be estimated (using an
empirical relationship). Similarly, knowing the causal effect of advertising, sales force,
and trade promotion spending on unit sales allows the marketer to calculate the effect
on profits and determine whether an increase or decrease is justified (see Figure 12.4).
The usefulness rests on the assumption that we will do a better job of understanding
marketing mix effects by separating those that must be empirically estimated from
others that are governed by accounting identities.

Unit Price

Advertising
Net Profit EM
Sales Force

Trade Promotion
EM Empirical Relationship

ID Identity Relationship

Objectives Decision

Figure 12.3 Empirical Relationships between Marketing Decisions and Objectives

Unit Cost

Unit
ID
Margin
ID Unit Price

Gross
Profit Advertising
Unit Sales EM

Sales Force

Net Profit ID
Trade Promotion

Marketing ID
$


Objectives Metrics Decision

Figure 12.4 Empirical Relationship with Components of Marketing Outcomes


Marketing mix models are used to estimate the effects of marketing levers on marketing
objectives and make decisions about how to allocate resources. One of the most fre-
quently applied marketing mix models is the one underlying simulated test markets and
depicted in Figure 12.5. With only minor variations, these models are used to forecast
new product sales (see Section 4.1 for more detail). The structure of this model is
straightforward, even if some would argue it is not simple. Forecast unit sales are calcu-
lated in a multiplicative identity from the metrics below. The multiplicative nature of
the identity captures the most significant interactions of the marketing mix without
resorting to (even more) complex equations. It is, we assert, more managerially trans-
parent and useful because of this well-structured system of metrics that defines and sep-
arates identities from empirical relationships.
Forecast Unit Sales = Number of Consumer Prospects *Awareness * Availability
*( Trial Rate * Trial Units + Repeat Rate * Repeat Units).
The input estimates for the components are obtained from the results of the simulated
test, surveys, management judgment, and/or empirical models.
One of the advantages of the model in Figure 12.5 is that it also provides clear and sep-
arate paths by which the different marketing mix elements are believed to impact unit
sales. Advertising affects Consumer Awareness but not Availability. Of course, in reality
“everything affects everything,” but the KISS structure affords a transparency and utility
that might be destroyed if management didn’t impose the discipline of focusing on the
most important empirical relationships that the identity relationships suggest.
In the case of the new product forecasting model in Figure 12.5 we have decomposed
(defined) the forecast sales to be a function of the metrics listed. The way we choose to
decompose the objective may be more or less suitable for separating marketing mix
empirical effects. For example, breaking down a share goal into share of requirements,
heavy usage index, and penetration share would not have an obvious relationship to
individual mix elements. Everything would still affect everything. So, not every identity
will be helpful in a model of marketing mix effects.


Advertising
Number of
Consumer
Prospects
Consumer
EM
Awareness
Sales Force

Availability EM Trade Promotion
(ACV%)

Forecast
ID
Unit Sales
Trial Rate%* Product Concept
EM
Trial Units

Unit Price

Repeat Rate%* EM
Repeat Units
Product Quality


Objectives
Metrics
Decision Levers

Figure 12.5 Simulated Test Markets Combine Empirical and Identity Relationships

Also, depending on how the data are collected, some identities may be strongly sug-
gested by the data, even if they are not directly measured. For example, in consumer
packaged goods markets, data on distribution (see Section 6.6) and channel promotion
activity (incremental sales lift %—see Section 8.1) are regularly collected and reported
to marketing managers. The availability of these two metrics strongly suggests the need
for a third metric, “preference,” to create an attractive identity that may be useful in sep-
arating empirical effects and allowing for important interactions. Figure 12.6 shows
how marketers might be able to “back into” values of preference by combining share, lift
%, and distribution metrics. Of course, this approach means that the marketers are
defining preference in a way that is consistent with relative choice under scenarios of
equal distribution and lift %.


Unit Price

Consumer
EM Advertising
Preference
%
Sales Force

Promotion
Share ID EM
Lift %
Trade Promotion

Distribution EM
PCV %



Metrics Constructs*
Decision Levers

Figure 12.6 Empirical Relationship with Marketing Components
and Intermediate Metrics and Constructs

By definition, accounting identities always hold. It is simply a matter of getting the cor-
rect values for the component parts. Other identities, such as those found in theoretical
models of finance and economics, are true “in theory” or assuming certain conditions.
For example, as discussed in Sections 7.3 and 7.4, at profit-maximizing levels of price,
this identity should be true:
Margin on Sales [(Price – Variable Cost)/Price] = 1
price elasticity for constant elasticity demand, or

Price = Variable Cost + 1⁄2 (Maximum Willingness to Pay
– Variable Cost) for linear demand functions
These identities identify relationships that are unlikely to be precise, but are vaguely
right.


Hubbard, Douglas W. (2007). How to Measure Anything: Finding the Value of “Intangi-
bles” in Business, John Wiley & Sons, Hoboken, New Jersey.

Appendix 1
Numerical Example
Consider a firm with estimated sales of $25,677 million last year. Although this is the
number stated in the annual report, marketing managers know that this number is an
estimate and not the actual sales. They judge the error in the estimate of sales to have a
standard deviation of $3,000 million. This means they judge there to be about a 68%
chance that actual sales is somewhere between $22,677 and $28,677 million. Keep in
mind that if the managers wanted to assume that $25,677 million was, indeed, the actual
sales figure, they would simply set the standard deviation of that estimate to zero.

Variable Initial Estimate stdev
Sales $25,677 $3,000
Salespersons $1,012 $5
Sales per salesperson $22 422
Our share 0.4 0.1
Market Sales $60,000 $1,000
Customers 15 1
Sales per customer $5,000 $5,000

Similarly, the marketing managers came up with estimates and standard deviations for
six outer-ring components of sales. In this particular example, we ignore vendor-related
metrics. Note that both sales per salesperson and sales per customer have high standard
deviations relative to their initial estimates. This reflects the fact that managers were not
certain about these two metrics and would expect their initial estimates to be off by
quite a bit.
Notice that we now have four ways to estimate sales: the initial estimate of $25,677 from
the managers and three other pairs of initial estimates of components that can be com-
bined (multiplied in this example) to also estimate sales. One way to proceed would be
to calculate those three other estimates and average all four estimates to get our final
estimate of sales. But we can do better than that. The unweighted average of the four
estimates does not take advantage of the information we have on the quality of each of


the initial estimates. Since sales per salesperson and sales per customer are very uncer-
tain, we might want to pay more attention to (give more weight to) the estimate we get
using the share and total sales estimates.
The process we propose for combining the initial estimates (and their quality measures)
into one set of final estimates is logical and straightforward. First, we want to find a set
of final estimates that satisfy the three identities (our final estimate of sales should equal
our final estimate of salespersons times our final estimate of sales per salesperson, for
example). And from among the many sets of final estimates that satisfy all the identities,
we want to find the one that is “closest” to the managers’ initial estimates—where close-
ness is measured in units of standard deviation.
In summary, our final estimates will be the set of metrics “closest” to our initial esti-
mates that satisfy all the identities in our model. Our final estimates will be internally
consistent and as close as possible to the initial set of estimates (which were not inter-
nally consistent).

Conclusion
“. . . metrics should be necessary (i.e., the company cannot do without them), precise,
consistent, and sufficient (i.e., comprehensive) for review purposes.” 4
Understanding metrics will allow marketers to choose the right input data to give
them meaningful information. They should be able to pick and choose from a variety of
metrics depending upon the circumstances and create a dashboard of the most vital
metrics to aid them in managing their business. After reading this work, we hope you
agree that no one metric is going to give a full picture. It is only when you can use mul-
tiple viewpoints that you are likely to obtain anything approaching a full picture.

“. . . results measures tell us where we stand in efforts to achieve goals, but not how we
go there or what to do differently”. 5
Marketing metrics are needed to give a complete picture of a business’s health. Financial
metrics focus on dollars and periods of time, telling us how profits, cash, and assets are
changing. However, we also need to understand what is happening with our customers,
products, prices, channels, competitors, and brands.
The interpretation of marketing metrics requires knowledge and judgment. This book
helps give you the knowledge so that you can know more about how metrics are con-
structed and what they measure. Knowing the limitations of individual metrics is
important. In our experience, businesses are usually complex, requiring multiple met-
rics to capture different facets—to tell you what is going on.


Because of this complexity, marketing metrics often raise as many questions as they
answer. Certainly, they rarely provide easy answers about what managers should do.
Having a set of metrics based on a limited, faulty, or outmoded view of the business can
also blind you. Such a set of metrics can falsely reassure you that the business is fine
when in fact trouble is developing. Like the ostrich with his head in the sand, it might
be more comfortable to know less.
We don’t expect that a command of marketing metrics will make your job easier. We do
expect that such knowledge will help you do your job better.


APPENDIX—SURVEY OF
MANAGERS’ USE OF METRICS
Job Title Industry Market
Q1. Which best describes what your business sells?

Products

Services
Relatively even mix of both products and services
Other

Q2. Purchase relationship with customers can best be defined as

Contractual for a specified period which customers can renew (e.g., magazines)
Contractual for an indefinite period which customers can cancel (e.g., news-
papers)
Frequent purchases (e.g., consumables, restaurant meals)
Infrequent purchase with little/no service/repair/supplies (e.g., digital cameras)
Infrequent purchase with service/repair/supplies relationship (e.g., automobiles,
printers)

Q3. Are your customers best understood as

Consumers (e.g., breakfast cereal)
Business or other organizational buying units (e.g., steel)
Relatively even mix of both consumers and business customers (e.g., UPS)

Q4. How does your business go to market?
Q5. What are the major influencers of the purchase decision?

Individual choice, little in the way of group dynamics (e.g., soft drinks, express
services)

385

Consumers rely heavily on recommendations of professionals (e.g., doctors,
plumbers)
Separate buying organization with multiple influences (e.g., corporate purchasing
organizations)
Other (please explain)

Q6. Total sales of your company are

Below $10 million $10-$100 million $101-$500 million
$501 million–$1 billion Over $1 billion
Q7. Over the last three years, the growth rate in sales at my company has been

Below 1% 1-3% 3-10% Over 10%

For the following questions, please tell us how useful you find each of the metrics
below in managing and monitoring your business.
Q8.1. How useful in managing and monitoring your business are the following
Market Share Measures?
Choices: Very Useful, Somewhat Useful, Not at All Useful, Don’t Know, N/A

1. Dollar (revenue) market share
2. Unit market share
3. Relative market share
4. Brand development index
5. Category development index
6. Market penetration
7. Brand penetration
8. Penetration share
9. Share of requirements
10. Heavy usage index
11. Hierarchy of effects


Hierarchy of Effect Metrics? (Consumer awareness, attitude, belief, trial, repeat, etc.
of product)

1. Brand awareness
2. Top of mind
3. Ad awareness
4. Consumer knowledge
5. Consumer beliefs
6. Purchase intentions
7. Purchase habits
8. Loyalty
9. Likeability
10. Willingness to recommend
11. Net promoter score
12. Customer satisfaction
13. Willingness to search

Margins and Cost Metrics?

1. Unit margin
2. Margin %
3. Channel margin
4. Average price per unit
5. Price per statistical unit
6. Variable and fixed costs
7. Marketing spending
8. Contribution per unit
9. Contribution margin %
10. Break-even sales

Appendix Survey of Managers’ Use of Metrics 387

Forecasting and New Product Metrics?

1. Target volumes
2. Target revenues
3. Trial volume
4. Repeat volume
5. Penetration
6. Volume projections
7. Annual growth %
8. Growth CAGR
9. Cannibalization rate
10. Brand equity metrics
11. Conjoint utilities
12. Conjoint utilities & volume projection

Customer Metrics?

1. Customers #
2. Recency
3. Retention rate
4. Customer profit
5. Customer lifetime value
6. Prospect lifetime value
7. Average acquisition cost
8. Average retention cost


Q8.6. How useful in managing and monitoring your business are the following Sales
Force Metrics?

1. Workload
2. Sales potential forecast
3. Sales total
4. Sales force effectiveness
5. Compensation
6. Break-even number of employees
7. Sales funnel, sales pipeline

Distribution and Retail Metrics?

1. Numeric distribution (%)
2. All commodity volume
3. Product category volume
4. Total distribution
5. Facings
6. Out of stock %
7. Inventories
8. Markdowns
9. Direct product profitability
10. GMROII

Pricing and Promotion Metrics?

1. Price premium
2. Reservation price


3. Percent good value
4. Price elasticity
5. Optimal price
6. Residual elasticity
7. Baseline sales
8. Incremental sales, or promotional lift
9. Redemption rates
10. Cost of coupons/rebates
11. Percentage sales with coupon
12. Percentage sales on deal
13. Percent time on deal
14. Average deal depth
15. Pass-through

Advertising Media and Web Metrics?

1. Impressions
2. Gross rating points
3. Cost per thousand impressions
4. Net reach
5. Average frequency
6. Effective reach
7. Effective frequency
8. Share of voice
9. Pageviews
10. Clickthrough rate
11. Cost per click
12. Cost per order


13. Cost per customer acquired
14. Visit (# Web site views)
15. Visitors (# Web site viewers)
16. Abandonment rate

Finance and Profitability Metrics?

1. Net profit
2. Return on sales
3. Return on investment
4. Economic profit (EVA)
5. Payback
6. Net present value
7. Internal rate of return
8. Return on marketing investment ROMI


BIBLIOGRAPHY
Aaker, David A. (1996). Building Strong Brands, New York: The Free Press.
Aaker, David A. (1991). Managing Brand Equity, New York: The Free Press.
Aaker, David A., and Kevin Lane Keller. (1990). “Consumer Evaluations of Brand
Extensions,” Journal of Marketing, V54 (Jan), 27.
Aaker, David W., and James M. Carman. (1982). “Are You Over Advertising?” Journal of
Advertising Research, 22, 57–70.
Abela, Andrew, Bruce H. Clark, and Tim Ambler. “Marketing Performance
Measurement, Performance, and Learning,” working paper, September 1, 2004.
Abraham, Magid H., and Leonard M. Lodish. (1990). “Getting the Most Out of
Advertising and Promotion,” Harvard Business Review, May–June, 50–58.
Ailawadi, Kusum, Donald Lehmann, and Scott Neslin (2003). “Revenue Premium as an
Outcome Measure of Brand Equity,” Journal of Marketing, Vol. 67, No. 4, 1-17.
Ailawadi, Kusum, Paul Farris, and Ervin Shames. (1999). “Trade Promotion: Essential to
Selling through Resellers,” Sloan Management Review, Fall.
Ambler, Tim, and Chris Styles. (1995). “Brand Equity: Toward Measures That Matter,”
working paper No. 95-902, London Business School, Centre for Marketing.
Barwise, Patrick, and John U. Farley. (2003). “Which Marketing Metrics Are Used and
Where?” Marketing Science Institute, (03-111), working paper, Series issues two 03-002.
Berger, Weinberg, and Hanna. (2003). “Customer Lifetime Value Determination and
Strategic Implications for a Cruise-Ship Line,” Database Marketing and Customer
Strategy Management, 11(1).
Blattberg, Robert C., and Stephen J. Hoch. (1990). “Database Models and Managerial
Intuition: 50% Model + 50% Manager,” Management Science, 36, No. 8, 887–899.
Borden, Neil H. (1964). Journal of Advertising Research, 4, June: 2-7.
Brady, Diane, with David Kiley and Bureau Reports. “Making Marketing Measure Up,”
Business Week, December 13, 2004, 112–113.
Bruno, Hernan, Unmish Parthasarathi, and Nisha Singh, Eds. (2005). “The Changing
Face of Measurement Tools Across the Product Lifecycle,” in Does Marketing Measure
Up? Performance Metrics: Practices and Impact, Marketing Science Conference
Summary, No. 05-301.

393

Christen, Markus, Sachin Gupta, John C. Porter, Richard Staelin, and Dick R. Wittink.
(1994). “Using Market-Level Data to Understand Promotion Effects in a Nonlinear
Model,” Journal of Marketing Research, August, Vol. 34, No. 3, 322–334.
Clark, Bruce H., Andrew V. Abela, and Tim Ambler. “Return on Measurement: Relating
Marketing Metrics Practices to Strategic Performance,” working paper, January 12, 2004.
Dekimpe, Marnik G., and Dominique M. Hanssens. (1995). “The Persistence of
Marketing Effects on Sales,” Marketing Science, 14, 1–21.
Dolan, Robert J., and Hermann Simon. Power Pricing: How Managing Price Transforms
the Bottom Line, New York: The Free Press, 4.
Farris, Paul W., David Reibstein, and Ervin Shames. (1998). “Advertising Budgeting: A
Report from the Field,” New York: American Association of Advertising Agencies.
Forrester, Jay W. (1961). “Advertising: A Problem in Industrial Dynamics,” Harvard
Business Review, March–April, 110.
Forrester, Jay W. (1965), “Modeling of Market and Company Interactions,” Peter D.
Bennet, ed. Marketing and Economic Development, American Marketing Association, Fall,
353–364.
Gregg, Eric, Paul W. Farris, and Ervin Shames. (revised, 2004). “Perspective on Brand
Equity,” Darden School Technical Notes, UVA-M-0668.
Greyser, Stephen A. (1980). “Marketing Issues,” Journal of Marketing, 47, Winter, 89-93.
Gupta and Lehman. (2003). “Customers As Assets,” Journal of Interactive Marketing,
17(1), 9–24.
Harvard Business School: Case Nestlé Refrigerated Foods Contadina Pasta & Pizza
(A) 9-595-035. Rev Jan 30 1997.
Hauser, John, and Gerald Katz. (1998). “Metrics: You Are What You Measure,” European
Management Journal, Vo. 16, No. 5, 517–528.
Interactive Advertising Bureau. Interactive Audience Measurement and Advertising
Campaign Reporting and Audit Guidelines, September 2004, United States Version 6.0b.
Kaplan, R. S., and V.G. Narayanan. (2001). “Measuring and Managing Customer
Profitability.” Journal of Cost Management, September/October: 5–15.
Keiningham, Timothy, Bruce Cooil, Tor Wallin Andreassen, and Lerzan Aksoy (2007). “A
Longitudinal Examination of Net Promoter and Firm Revenue Growth.” Journal of
Marketing, Volume 71, July.
Little, John D.C. (1970). “Models and Managers: The Concept of a Decision Calculus,”
Management Science, 16, No. 8, b-466—b-484.
Lodish, Leonard M. (1997). “J.P. Jones and M.H. Blair on Measuring Advertising Effects
“Another Point of View,” Journal of Advertising Research, September–October, 75–79.


McGovern, Gail J., David Court, John A. Quelch, and Blair Crawford. (2004). “Bringing
Customers into the Boardroom,” Harvard Business Review, November, 70–80.
Meyer, Christopher (1994), How the Right Measures Help Teams Excel, Harvard
Business Review, May–June, pp. 95–103.
Much, James G., Lee S. Sproull, and Michal Tamuz. (1989). “Learning from Samples of
One or Fewer,” Organizational Science, Vol. 2, No. 1, February, 1–12.
Murphy, Allan H., and Barbara G. Brown. (1984). “A Comparative Evaluation of
Objective and Subjective Weather Forecasts in the United States,” Journal of Forecasting,
Vol. 3, 369–393.
Net Genesis Corp. (2000). E-Metrics: Business Metrics for the New Economy.
Net Genesis & Target Marketing of Santa Barbara.
Peppers, D., and M. Rogers. (1997). Enterprise One to One: Tools for Competing in the
Interactive Age, New York: Currency Doubleday.
Pfeifer, P.E., Haskins, M.E., and Conroy, R.M. (2005). “Customer Lifetime Value,
Customer Profitability, and the Treatment of Acquisition Spending,” Journal of
Managerial Issues, 25 pages.
Poundstone, William. (1993). Prisoner’s Dilemma, New York: Doubleday, 118.
Reichheld, Frederick F., and Earl W. Sasser, Jr. (1990). “Zero Defections: Quality Comes
to Services,” Harvard Business Review, September–October, 105–111.
Reichheld, Fred. (2006). The Ultimate Question: Driving Good Profits and True Growth.
Boston: Harvard Business School Publishing Corporation.
Roegner, E., M. Marn, and C. Zawada. (2005). “Pricing,” Marketing Management,
Jan/Feb, Vol. 14 (1).
Sheth, Jagdish N., and Rajendra S. Sisodia. (2002). “Marketing Productivity Issues and
Analysis,” Journal of Business Research, 55, 349–362.
Tellis, Gerald J., and Doyle L. Weiss. (1995). “Does TV Advertising Really Affect Sales?
The Role of Measures, Models, and Data Aggregation,” Journal of Marketing Research,
Fall, Vol. 24-3.
Wilner, Jack D. (1998). 7 Secrets to Successful Sales Management, Boca Raton, Florida:
CRC Press LLC; 35–36, 42.
Zellner, A., H. Kuezenkamp, M. McAleer, Eds. (2001). “Keep It Sophisticatedly Simple.”
Simplicity, Inference and Econometric Modeling. Cambridge University Press,
Cambridge, 242–262.
Zoltners, Andris A., and Prabhakant Sinha, and Greggor A. Zoltners. (2001). The
Complete Guide to Accelerating Sales Force Performance, New York: AMACON.

Bibliography 395

ENDNOTES

Chapter 1
1. Word Reference, www.wordreference.com. Accessed 22 April 2005.
2. Bartlett, John. (1992). Bartlett’s Familiar Quotations, 16th edition; Justin Kaplan,
general editor.
3. Hauser, John, and Gerald Katz. “Metrics: You are What You Measure,” European
Management Journal, Volume 16 No 5 October 1998.
4. Kaplan, Robert S., and David P. Norton. (1996). Balanced Scorecard, Boston, MA:
Harvard Business School Press.
5. Brady, Diane, with David Kiley and Bureau Reports, “Making Marketing Measure
Up,” Business Week.
6. Strictly speaking, all the numbers can contain some error. Share may be estimated, for
example, from retail sales to consumers. Sales might come from shipment to retailers.
7. Barwise, Patrick, and John U. Farley. (2003). “Which Marketing Metrics Are Used and
Where?” Marketing Science Institute (03-111), working paper, Series issues two 03-002.
8. Ambler, Tim, Flora Kokkinaki, and Stefano Puntoni. (2004). “Assessing Marketing
Performance: Reasons for Metrics Selection,” Journal of Marketing Management, 20,
475–498.

Chapter 2
1. “Wal-Mart Shopper Update,” Retail Forward, February 2005.
2. “Running Out of Gas,” Business Week, March 28th, 2005.
3. American Marketing Association definition. Accessed 06/08/2005. http://www
.marketingpower.com/live/mg-dictionary.php?SearchFor=market+concentration&
Searched=1.
4. Check the Marketing Evaluations, Inc., Web site for more detail: http://www.qscores
.com/. Accessed 03/03/05.
5. Claritas provides the Prizm analysis. For more details, visit the company Web site:
http://www.clusterbigip1.claritas.com/claritas/Default.jsp. Accessed 03/03/05.

397

6. Reichheld, Fred, The Ultimate Question: Driving Good Profits and True Growth
(Boston: Harvard Business School Publishing Corporation, 2006.)
7. http://www.theultimatequestion.com/theultimatequestion/measuring_netpromot-
er.asp?groupCode=2
8. Timothy Keiningham, Bruce Cooil, Tor Wallin Andreassen and Lerzan Aksoy (2007)
“A Longitudinal Examination of Net Promoter and Firm Revenue Growth.” Journal
of Marketing, Volume 71, July 2007.

Chapter 3
1. “Running Out of Gas,” Business Week, March 28th, 2005.
2. This formula should be familiar if we consider that the supplier selling price is
merely the cost to that layer of the chain. So this becomes Selling Price = Cost/(1
Margin %). This is the same as Sale $ = Cost $ + Margin $.
3. Those familiar with basic economics use the term “marginal cost” to refer to the cost
of an additional unit of output. In this linear cost model, marginal cost is the same
for all units and is equal to the variable cost per unit.
4. Both contribution per unit ($) and contribution margin (%) are closely related to unit
margin ($) and margin (%). The difference is that contribution margins (whether unit-
or percentage-based) result from a more careful separation of fixed and variable costs.

Chapter 4
1. Harvard Business School Case: Nestlé Refrigerated Foods Contadina Pasta & Pizza
(A) 9-595-035. Rev Jan 30 1997.
2. Kusum Ailawadi, Donald Lehmann, and Scott Neslin (2003), “Revenue Premium as
an Outcome Measure of Brand Equity,” Journal of Marketing, Vol. 67, No. 4, 1-17.
3. Bruno, Hernan, Unmish Parthasarathi, and Nisha Singh, eds. (2005). “The Changing
Face of Measurement Tools Across the Product Lifecycle,” Does Marketing Measure
4. Young and Rubicam can be found at: http://www.yr.com/yr/. Accessed 03/03/05.
5. Bruno, Hernan, Unmish Parthasarathi, and Nisha Singh, eds. (2005). “The Changing
Face of Measurement Tools Across the Product Lifecycle,” Does Marketing Measure


6. See Darden technical note and original research.
7. The information from Bill Moran comes from personal communications with the
authors.
8. Interbrand can be contacted at: http://www.interbrand.com/. Accessed 03/03/05.

Chapter 5
1. “Vodafone Australia Gains Customers,” Sydney Morning Herald, January 26, 2005.
2. “Atlanta Braves Home Attendance.” Wikipedia, the free encyclopedia. http://en.
wikipedia.org/wiki/Major_League_Baseball_attendance_records
3. Thanks to Gerry Allan, President, Anametrica, Inc. (developer of Web-based tools for
managers) for his work on this section.
4. Pfeifer, P.E., Haskins, M.E., and Conroy, R.M. (2005). “Customer Lifetime Value,
Customer Profitability, and the Treatment of Acquisition Spending,” Journal of
Managerial Issues, 25 pages.
5. Kaplan, R.S., and V.G. Narayanan. (2001). “Measuring and Managing Customer
Profitability,” Journal of Cost Management, September/October, 5–15.
6. Peppers, D., and M. Rogers. (1997). Enterprise One to One: Tools for Competing in the
Interactive Age, New York: Currency Doubleday.
7. Berger, P.D., B. Weinberg, and R. Hanna. (2003). “Customer Lifetime Value
Determination and Strategic Implications for a Cruise-Ship Line,” Database
Marketing and Customer Strategy Management, 11(1), 40–52.
8. Gupta and Lehman. (2003). “Customers as Assets,” Journal of Interactive Marketing,
17(1), 9–24.

Chapter 6
1. Material in Sections 7.1–7.5 is based on a Note on Sales Force Metrics, written by
Eric Larson, Darden MBA 2005.
2. Zoltners, Andris A., Prabhakant Sinha, and Greggor A. Zoltners. (2001). The Com-
plete Guide to Accelerating Sales Force Performance, New York: AMACON.
3. Wilner, Jack D. (1998). 7 Secrets to Successful Sales Management, Boca Raton, Florida:
CRC Press LLC; 35–36, 42.

Endnotes 399

4. For more on these total allocations, see Zoltners, Andris A., Prabhakant Sinha, and
Greggor A. Zoltners. (2001). The Complete Guide to Accelerating Sales Force
Performance, New York: AMACON.
5. Zoltners, Andris A., Prabhakant Sinha, and Greggor A. Zoltners. (2001). The
6. Dolan, Robert J., and Benson P. Shapiro. “Milford Industries (A),” Harvard Business
School, Case 584-012.
7. Zoltners, Andris A., Prabhakant Sinha, and Greggor A. Zoltners. (2001). The
8. Jones, Eli, Carl Stevens, and Larry Chonko. (2005). Selling ASAP: Art, Science,
Agility, Performance, Mason, Ohio: South Western, 176.
9. Product category volume is also known as weighted distribution.
10. The authors use the term product category volume (PCV) for this metric. However,
this term is not as widely used in industry as all commodity volume (ACV).

Chapter 7
1. Dolan, Robert J., and Hermann Simon. Power Pricing: How Managing Price
Transforms the Bottom Line, New York: The Free Press, 4.
2. Barwise, Patrick, and John U. Farley, “Which Marketing Metrics Are Used and
Where?” Marketing Science Institute, (03-111) 2003, working paper, Series issues
two 03-002.
3. Constant elasticity functions are also called log linear because they can be expressed
as: log Q = log A + elasticity x log (p).
4. In graphing such relationships, economists often plot price on the vertical axis and
quantity demanded on the horizontal axis. When reviewing a graph, managers are
advised to always check the axis definitions.
5. If price elasticity is expressed in shorthand as a positive number, then we do not
need the negative sign in the formula that follows.
6. Poundstone, William. (1993). Prisoner’s Dilemma, New York: Doubleday, 118.

Chapter 8
1. In this context, we use the term “permanent” with some flexibility, recognizing that
even long-term arrangements must be subject to change in response to market and
industry dynamics.


2. Often, contribution can be used as a proxy for profits.
3. Distribution for coupons is used in the sense of postage and insertion costs, rather
than retail and inventory logistics.
4. For a richer discussion, see Ailawadi, Farris, and Shames, Sloan Management Review,
Fall 1999.
5. Roegner, E., M. Marn, and C. Zawada. (2005). “Pricing,” Marketing Management,
Jan/Feb, Vol. 14 (1).
6. “How to Fix Your Pricing if it is Broken,” by Ron Farmer, CEO, Revenue Technologies
for The Professional Pricing Society: http://www.pricingsociety.com/htmljournal/
4thquarter2003/article1.htm. Accessed 03/03/05.
7. The following are the two main types of injury contemplated by the Act: (a): Price
discrimination might be used as a predatory pricing tactic, setting prices below cost
to certain customers to harm competition at the supplier’s level. Anti-trust authori-
ties use the same standards applied to predatory pricing claims under the Sherman
Act and the FTC Act to evaluate allegations of price discrimination used for this
purpose. (b) Secondary Line competitive injury: A seller charging competing buyers
different prices for the same “commodity” or discriminating in the provision of
“allowances” such as compensation for advertising and other services may be violat-
ing the Robinson-Patman Act. This kind of price discrimination can hurt competi-
tion by giving favored customers an edge in the market that has nothing to do with
their superior efficiency. However, in the U.S., price discrimination is generally
lawful, particularly if it reflects the different costs of dealing with diverse buyers or
results from a seller’s attempts to meet a competitor’s prices or services. Clearly this
is not intended to be a legal opinion, and legal advice should be sought for a compa-
ny’s individual circumstances.

Chapter 9
1. Farris, Paul W. (2003). “Getting the Biggest Bang for Your Marketing Buck,”
Measuring and Allocating Marcom Budgets: Seven Expert Points of View, Marketing
Science Institute Monograph.
2. Known as client-side tagging, beacon, and 1 1 clear pixel technology.
3. The Interactive Advertising Bureau gives the following definition of ad impression:
“A measurement of responses from an ad delivery system to an ad request from the
user’s browser, which is filtered from robotic activity and is recorded at a point as late
as possible in the process of delivery of the creative material to the user’s browser—
therefore closest to actual opportunity to see by the user.” Interactive Audience

Endnotes 401

Measurement and Advertising Campaign Reporting and Audit Guidelines.
September 2004, United States Version 6.0b.
4. The spending data is taken from “Internet Weekly,” Credit Suisse First Boston,
14 September 2004, 7–8.
5. http://www.nielsen-netratings.com/. Accessed 06/11/2005.
6. http://www.google.com/support/googleanalytics/bin/answer.py?answer=
81986&cbid=gbo1sdrurcrz&src=cb&lev=answer

Chapter 10
1. Economic value added is a trademark of Stern Stewart Consultants. For their
explanation of EVA, go to http://www.sternstewart.com/evaabout/whatis.php.
Accessed 03/03/05.
2. The weighted average cost of capital, a.k.a. the WACC, is just the percentage return
expected to capital sources. This finance concept is better left to specialist texts, but
to give a simple example, if a third of a firm’s capital comes from the bank at 6%
and two-thirds from shareholders who expect a 9% return, then the WACC is the
weighted average 8%. The WACC will be different for different companies, depend-
ing on their structure and risks.
3. Excel has a function to do this quickly, which we explain at the end of the section.
However, it is important to understand what the calculation is doing.
4. A terminal value in a simple calculation might be assumed to be zero or some sim-
ple figure for the sale of the enterprise. More complex calculations consider estimat-
ing future cashflows; where this is done, ask about assumptions and importance. If
the estimated terminal value is a significant area of the analysis, why have you cur-
tailed the full analyses at this point?
5. Hawkins, Del I., Roger J. Best, and Charles M. Lillis. (1987). “The Nature and
Measurement of Marketing Productivity in Consumer Durables Industries: A Firm
Level Analysis,” Journal of Academy of Marketing Science, Vol. 1, No. 4, 1–8.

Chapter 11
1. Churn = percent of customers lost each year.
2. ACV = all commodity volume, a measure of distribution coverage (refer to
Section 6.6).


Chapter 12
1. An identity is “an equality satisfied by all values of the variables for which the expres-
sion involved in the equality are defined.” American Heritage Dictionary, 2nd Edition,
Houghton Mifflin Company, Boston, 1982.
In finance, economics, and accounting, an identity is “an equality that must be true
regardless of the value of its variables, or a statement that by definition (or construc-
tion) must be true.” Where an accounting identity applies, any deviation from the
identity signifies an error in formulation, calculation, or measurement. http://en.
wikipedia.org/wiki/Accounting_identity#cite_note-0
2. Borden, Neil H., Source: Journal of Advertising Research, 4, June 1964: 2-7.
3. Zellner, A., 2001. “Keep It Sophisticatedly Simple.” Zellner, A., Kuezenkamp, H.,
McAleer, M. (eds.), Simplicity, Inference and Econometric Modeling. Cambridge
University Press, Cambridge, 242–262.
4. Ambler, Tim. (2000). Marketing and the Bottom Line: The New Metrics of Corporate
Wealth, London: Prentice Hall.
5. Meyer, Christopher. (1994). “How the Right Measures Help Teams Excel,” Harvard
Business Review.

Endnotes 403

INDEX

SYMBOLS average acquisition cost, 176-177
average deal depth, 264
# (count), 7 average frequency, 295, 298, 302
$ (Dollar Terms), 6 average margin, 82-84
% (percentage), 6 average price charged, 224
average price displayed, 224
A average price paid, 223
average price per unit, 86-87
A.C. Nielsen, 207 calculating, 87-90
Aaker, David, 137 complications, 90
AAU (Awareness, Attitudes, and Usage), 51 purpose, 86-87
attitude, 53 average prices, 85
awareness and knowledge, 52 average retention cost, 176-177
calculating, 52 averaging estimates, 374-375, 382-383
cautions, 54-55 awareness, 52
data sources, 54 customer awareness, 361
purpose, 51 trial rate, 114
usage, 54 Awareness, Attitudes, and Usage. See AAU
abandoned purchases, 331
abandonment, 328
abandonment rate, 331 B
accepters, 45 balancing sales force territories, 187-188
accountability, 2 banks, counting customers, 160
acquisition versus retention, 176-178 baseline sales, 265, 267
ACV (all commodity volume), 184, 202-205 calculating, 268-273
ad awareness, 53 complications, 273
adjusting for periodic changes, 54 profitability, 273
advertising. See also impressions purpose, 267
as percentage of sales, 101 BCG (Boston Consulting Group) matrix, 36
price versus cost, 314 BDI (Brand Development Index), 40-42
advertising effectiveness, 307, 309 Big Tobacco Company, 364-365
advertising exposure, 307 bonuses. See sales force compensation
Ailawadi, Kusum, 137 Boom
all commodity volume (ACV), 184, 202-205 customer awareness, 361
allowances, slotting, 101 customer profit, 360
apparel retailers, customers, 161 financial statements, 357-358
asset turnover, 369 marketing metrics, 360
assumptions Borden, Neil Sr., 376
infinite horizon assumption (customer Boston Consulting Group (BCG) matrix, 36
lifetime value), 172 bounce rate, 293, 332-333
test markets, 120-121 Brand Asset Valuator, 137, 139-141
attitudes/liking/image, 53 Brand Development Index (BDI), 40-42
attrition, 159 brand equity, 135
availability of data, 3 measuring, 111, 137-141
AVC on display, 209 purpose, 136-137
AVC on promotion, 209

405

Brand Equity Index, 138-139 competitor price elasticity, 254
Brand Equity Ten, 137 competitor reaction elasticity, 252
brand identity, 141 complications
brand image, 141 average price per unit, 90
brand penetration, 42-43 channel margins, 81
brand position, 141 Compound Annual Growth Rates (CAGR),
brand strategy, 141 109, 111, 129
Brand Valuation Model, 141 compounding growth, 126, 128-129
brand/product knowledge, 53 Concentration Ratio, 38
brands, number purchased, 48 conjoint analysis, 137, 141-144, 228
breadth of distribution, 208 conjoint utilities, 112, 142, 147-151
break-even analysis, 101-102 constant elasticity, 236-238
break-even on incremental investment, 105 constructing frequency response functions,
classifying costs, 105 307-308
purpose, 102 consumer off-take, 214
break-even number of employees, 195, 197 consumer preference, 142-146
break-even on incremental investment, 105 consumer ratings, 53
break-even point, calculating, 102-104 contractual situations, 156-157
break-even sales level, 68 contribution analysis, 101
breakage, 277 contribution margin, 66, 68, 104
Brita water filters, 86 contribution per unit, 68, 101-103
budgeting risk, assessing, 97-98 converting markups to margins, 80
budgets, 2 cookies, 331
buying power, 188 cost effectiveness of Internet marketing, 323
cost of incremental sales, 267
cost per click, 323-326
C cost per customer acquired, 327
CAGR (Compound Annual Growth Rates), cost per friend, 335
109, 111, 129 cost per impression, 323-325
cannibalization rate, 111, 130-135 cost per order, 323-325
cash flows, internal rate of return, 349 cost per point (CPP), 300
category development index (CDI), 40-42 cost per thousand impressions rates. See CPM
category performance ratio, 202, 207 cost-plus pricing, 248
cautions (AAU), 54-55 costs
CDI (category development index), 40-42 assigning to customers, 165
chaining margins, 75 average acquisition cost, 176-177
channel margins, 75, 81 average retention cost, 176-177
channel metrics, Prestige Luggage, 363 classifying for break-even analysis, 105
choosing metrics, 3 commissioned sales costs, 99
churn, 159 fixed costs. See fixed costs
classification of variable costs, 96 overhead costs, 341
clickstream, 329-330 total cost, 92, 95
clickthrough rates, 320-322 total cost per unit, 94
cluster analysis, 148 total selling costs, 98
CLV. See customer lifetime value total variable selling costs, 98
cohort and incubate (customer lifetime value), variable costs. See variable costs
168-169 count (#), 7
cold leads, 200 counting customers, 156-161
commissioned sales costs, 99 contractual situations, 157
commissions. See sales force compensation non-contractual situations, 157-158
company profit from new products, 125 recency, 156-158
comparing sales force territories, 188 retention, 158
compensation. See sales force compensation
compensatory decisions versus noncompen-
satory consumer decisions, 144-146


coupons, 275 deciding who to serve, 166
evaulating, 278 defining, 159-160
percentage sales with coupons, 275 ever-tried customers, 45
profitability, 276 impressions. See impressions
redemption rate, 275-277 market penetration, 42-43
CP. See customer profit purpose, 156
CPM (cost per thousand impressions), 289, second tier customers, 162
299-300 surveys. See surveys
CPP (cost per point), 300 third tier customers, 162
cross elasticity, 251 top tier customers, 162
cross price elasticity, 252, 254 total number of active customers, 45
Cruise unprofitable customers, 166
customer awareness, 361
customer profit, 360
financial statements, 357-358 D
marketing metrics, 360 dashboards. See marketing dashboards
customer awareness, Boom and Cruise, 361 data, availability of, 3
customer lifetime value (CLV), 153, data parameters, market share, 34
166-167, 174 data sources
calculating, 169-170 AAU, 54
cohort and incubate, 168-169 heavy usage index, 50
discount rate, 171 decline (life cycle), 129
finite-horizon, 172 decomposing
infinite horizon assumption, 172 for diagnostic purposes, 373-374
purpose, 167-168 indirect metric estimates, 376
retention rate, 170 law of large numbers, 374-375, 382-383
versus prospect lifetime value, 174-176 market share, 44
customer lifetime value with initial reasons for using, 372-373
margin, 171 sales, 371-372
customer profit, 153, 161-162, 165 deductions, 214, 284
Boom, 360 demand
calculating, 162-164 linear demand
Cruise, 360 optimal price, 240-246
purpose, 161-162 price elasticity, 233-236
quantifying, 167 reservation prices, 228-231
whale curve, 167 price tailoring, 285
customer responses, separating from demand curves, constant elasticity, 236-238
non-customer responses, 54 diagnostic purposes, decomposing for, 373-374
customer satisfaction, 56-57 differentiation
measuring, 57-58 brand equity, 139
purpose, 56-57 product differentiation, 142
sample selection, 59 direct product costs, 216
surveys, 59 direct product profitability. See DPP
customer selling price, 75-78 discount rate, 171, 350
customer service, 194 discounted trial, 124
customer survey data, triangulating, 55 discounts, 283
customer time, 159 distribution, trial rates, 115
customers, 156, 159 distribution chains, 75
abandoning, 166 distribution channels, calculating selling prices
accepters, 45 at each level, 76
acquisition versus retention, 176-178 distribution metrics, 202
assessing value of, 167-168 ACV, calculating, 204-205
assigning cost to, 165 data sources, 207-208
brand penetration, 42-43 numeric distribution, calculating, 203-204
counting, 156-161

Index 407

PCV, calculating, 206-207 FIFO (First In, First Out), 213
purpose, 203 financial statements, Boom and Cruise, 357-358
districts, 190 finite-horizon (customer lifetime value), 172
diverted goods, 214 first channel member’s selling price, 78-79
diverted merchandise, 214 First In, First Out (FIFO), 213
Dollar Terms ($), 6 first-time triers in period, 113
double jeopardy, 47 fixed costs, 91, 100
downloads, 335-336 calculating, 91-95
DPP (direct product profitability), 182, 186, classification of, 96
215-218 purpose, 91
Drucker, Peter, 65 followers, 293, 333
DuPont Model, 369-370, 372. See also identities calculating, 334
durability, 138 cautions, 334
cost per friend, 335
outcomes per friend, 335
E purpose, 334
eBay, active users, 158 forced trial, 124
EBITDA (earning before interest taxes, forecasting
depreciation, and amortization), 341 marketing spending, 97-98
Economic Profit, 339, 343-345 trial volume, 116
Economic value added (EVA), 337, 343 upcoming sales, 198
EDLP (everyday low prices), 284 Fortune, 159
effective frequency, 290, 310-312 frequency, 301
effective market share, 138 average frequency, 302
effective reach, 310-312 effective frequency, 290, 310-312
effectiveness. See sales force effectiveness frequency response functions, 289, 305, 309-310
elasticity. See price elasticity construction, 307-308
empirical relationships, 373-374 learning curve response model, 305-306
marketing mix models, 378, 380-381 linear response model, 305-306
esteem, brand equity, 139 purpose, 306-307
estimates threshold response model, 306
averaging, 374-375, 382-383 friends, 293, 333
for indirect metrics, 376 calculating, 334
EVA (economic value added), 337, 343 cautions, 334
evaluating cost per friend, 335
coupon programs, 278 outcomes per friend, 335
inventories, 213 purpose, 334
multi-period investments, 345-346
sales goals, 191
temporary price promotions, 264 G
workload distribution, 198 geo-clustering, 55
ever tried customers, 45, 124 globalization, 3
everyday low prices (EDLP), 284 GM, retail sales, 65
evoked set, 125 GMROII (gross margin return on inventory
expenses, sales force effectiveness, 194 investment), 182, 186, 215-217
exposures, 293 goals, sales, 189-191
goodwill, 136
gross margin, 75, 239
F gross rating points (GRPs), 288, 294-297, 302
facings, 208 growth, 125
fair share draw, 111, 130-134 CAGR, calculating, 129
features in store, 208 compounding growth, 126-129
Federal Trade Commission, 285 life cycle, 129


percentage growth, 126, 129 inflation, estimating, 90
same stores growth, 126-128 intangibles, goodwill, 136
value of future period, 128-129 intention to purchase, 53
year-on-year growth, 125 intentions, 53
GRPs (gross rating points), 288, 294-297, 302 interactive media. See rich media
Interbrand, 137, 141
interest creation, 200
H–I Internal Rate of Return (IRR), 338-339, 345-349
heavy usage index, 44, 49-50 Internet, 288. See also web pages
Herfindahl index, 38-39 assessing cost effectiveness, 323
HI-LO (high-low), 284 effective reach, 312
hierarchy of effects, 55 search engines, 325-327
hits, 314-315 introductory life cycle, 129
hybrid channel margins, 81 inventory, 208
evaluating, 213
I (Index) notation, 7 inventory days, 211-212
identifying profitability of individual inventory tracking, 211
customers, 161-162 inventory turns, 209, 211
identities investments, multi-period, evaluating, 345-346
decomposing sales, 371 invoice price, 281-282
defined, 369 IRR (Internal Rate of Return), 338-339, 345-349
for diagnostic purposes, 373-374
for estimates of indirect metrics, 376
marketing mix models, 376-381 J–K–L
reasons for using, 372-373 Kaplan, Robert, 163
impressions, advertising, 293 Kelvin, Lord, 2
calculating, 295 knowledge
clickthrough rates, 320-322 brand equity, 139
complications, 298 brand/product knowledge, 53
cost per click, 323-326
cost per impression, 323-325 Last In, First Out (LIFO), 213
cost per order, 323-325 law of large numbers, 374-375
CPM. See CPM numerical example, 382-383
data sources, 297-298 leading national advertisers (LNA), 313
frequency response functions. See frequency learning curve, 289
response functions learning curve response model, frequency
GRPs, 294-297 response functions, 305-306
net reach. See net reach life cycle, 129
pageviews, 314-316 LIFO (Last In, First Out), 213
purpose, 294 likeability, 55
share of voice. See share of voice linear cost model, 96
incentive plans, 197-198 linear demand
income statement, Prestige Luggage, 362 optimal price, 240-246
incremental sales, 267-268 price elasticity, 233-236
indexes reservation prices, 228-231
Brand Development Index, 40-41 linear response model, frequency response
CDI (category development index), 40-42 functions, 305-306
heavy usage index. See heavy usage index list price, 281
Herfindahl index, 38-39 LNA (leading national advertisers), 313
indicators, separating leading from lagging, 55 loyalty, 122, 359
indirect metrics, estimates for, 376 double jeopardy, 47
infinite horizon assumption (customer lifetime number of brands purchased, 48
value), 172 willingness to search, 62-63

Index 409

M Marlboro Friday, 365
mastering metrics, 4
mail-in rebates, 277 mature life cycle, 129
make-goods on promotions, 214 maximum reservation price (MRP), 229,
margin on new products, 125 240, 246
margins, 65, 69 maximum willing to buy (MWB), 229-230, 246
average margin, 82-84 measuring
chaining, 75 brand equity, 137-141
channel margins. See channel margins customer satisfaction, 57-58
contribution margins, 66-68 market share over time, 35
converting from markups, 80 media exposure return on marketing
costs, including or excluding, 75 investment, 354-355
customer lifetime value with initial media plans, net reach, 302
margin, 171 metric usage survey, 385-390
gross margin, 75, 239 metrics
percentage margins, 69-71, 82 defined, 1
as percentage of costs, 72 reasons for having, 2
purpose, 69 survey
reported margins, 72, 74 cautions about, 10-11
selling prices, defining, 72 rankings, 21-24
unit margin, 69-71 results, 13, 385
versus markup, 73-75 sampling size, 11-12
weighted contribution margins, middlemen, 278
cannibalization, 132 misshipments, 214
markdowns, 214-216 Moran, Bill, 137-138
market concentration, 35, 38 MRP (maximum reservation price), 229,
market penetration, 42-43 240, 246
market share, 28, 32 multi-period investments, evaluating, 345-346
bias in reported shares, 35 MWB (maximum willing to buy), 229-230, 246
data parameters, 34
decomposing, 44
measuring over time, 35 N
purpose of, 33 net operating profit after tax (NOPAT), 342
quantifying, 34-35 Net Present Value (NPV), 338-339, 345-350
relative market share. See relative net price, 281-282
market share Net Profit, 338, 340-341
revenue market share, calculating, 33 net promoter, 60-62
served market, 34 Net Promoter Score (NPS), 60-62
unit market share, 33 net reach, 297, 301, 303
market share rank, 39 complications, 305
marketing as a percentage of sales, 101 overlap effects, 304-305
marketing budgets, developing, 100 purpose, 301-304
marketing dashboards, 365-367 noise, 54
marketing metrics, 359-363, 367, 383 non-compensatory consumer decisions versus
marketing mix models, 376, 378, 380-381 compensatory decisions, 144-146
marketing spending, 97 non-contractual situations, 156-158
calculating, 99-100 NOPAT (net operating profit after tax), 342
fixed costs, 100 NPS (Net Promoter Score), 60-62
purpose, 97-98 NPV (Net Present Value), 338-339, 345-350
slotting allowances, 101 number of complaints, 59
markups number of new products, 125
converting to margins, 80 numeric distribution, 184, 202-204
versus margins, 73-75


O pipeline analysis, 198
construction, 199-201
obsolescence, 214 purpose, 198-199
opportunities-to-see (OTS), 293 sales funnel, 201-202
optimal price, 239 pipeline sales, 214
calculating, 246-248 PLV. See prospect lifetime value
complications, 248 post-purchases, 200
purpose, 240-246 pre-purchase, 200
relative to gross margin, 247 Prestige Luggage, 362-363
slope, 244 price discrimination, 248, 250-251, 284-285
optimality condition, 247 price elasticity, 220, 232-233, 239. See also
OTS (opportunities-to-see), 293 residual price elasticity
out-of-stocks, 185, 209-210 calculating, 233-236
outcomes per friend, 335 constant elasticity, 236-238
over-servicing, 187 cross elasticity, 251
overhead costs, 341 linear demand, 233-236
overlap, assessing, 305 purpose, 233
overlap effects, 304-305 price increases, evaluating, 90
own price elasticity, 252-254 price of a specified competitor, 222
price per statistical unit, 67, 86, 88-89
P price premiums, 222-225
price promotions. See promotions
pageviews, 314-316, 328 price tailoring, 248, 250-251, 284-285
pass-through, 266, 278-280 price waterfalls, 264, 266, 280-283
payback, 346 prices
payback period, 106 average price charged, 224
PCV (product category volume), 184, 202 average price displayed, 224
calculating, 206-207 average price paid, 223
net out-of-stocks, 210 average price per unit, 86-87
penetration, 42, 112 calculating, 87-90
brand penetration, 42-43 complications, 90
calculating, 43, 113-114 purpose, 86-87
cautions, 45 average prices, 85
market penetration, 42-43 competitor price elasticity, 254
share, 42 cost-plus pricing, 248
penetration rate, 43 cross elasticity, 251
penetration share, 43-44 cross price elasticity, 254
Peppers, Don, 167 customer selling price, 75, 77-78
perceived quality/esteem, 53 first channel member’s selling price, 78-79
perceived value for money, 53 invoice prices, 281-282
percent good value, 226 list price, 281
percentage (%), 6 net price, 281-282
percentage growth, 126, 129 optimal price. See optimal price
percentage margins, 69-71, 82 own price elasticity, 254
percentage of unit sales, 82 percent good value, 226
percentage sales on deal, 278-279 price discrimination, 284
percentage sales with coupons, 275 price elasticity. See price elasticity
performance, 2, 156 price of a specified competitor, 222
performance reviews. See sales force price per statistical unit, 86, 88-89
effectiveness price premiums, 222-225
periodic changes, adjusting for, 54 price tailoring, 248, 250-251, 284-285
price waterfalls, 264, 266, 280-283
prisoner’s dilemma pricing, 256-262
reservation prices. See reservation prices

Index 411

residual price elasticity. See residual price R (Rating), 7
elasticity rain checks, 214
selling price, 72, 76 Ramsellar, Leon, 140
supplier selling price, 75-77, 85 rankings in marketing metrics survey, 21-24
theoretical price premiums, 226 Rating (R), 7
primary line competitive injury, 251 rating point, 293
prisoner’s dilemma pricing, 256-262 reach, 301-303. See also net reach
Prizm, geo-clustering, 55 rebates, 275-277
product category volume. See PCV recency, 156, 158
product differentiation, 142 redemption rates, 275-277
Professional Pricing Society, 283 regulations, price discrimination, 251, 285
profit margin, 369 relationships, 160, 373-374
profit-based sales targets, 106-107 relative market share, 35-37
profitability relative perceived quality, 53
baseline sales, 273 relative price, 138. See also price premiums
coupons, 276 relevance, brand equity, 139
price tailoring, 284 repeat, 124
of promotions, 271 repeat rates, 48, 121
redemption rates, 276 repeat volume, 117-118
profitability metrics, 214 reporting margins, 72, 74
complications, 217-218 repurchase rate, 48
DPP, 215-217 resellers, 279
GMROII, 215-216 reservation prices, 226
markdowns, 215-216 calculating, 226, 228
purpose, 215 finding, 228
projected volume, repeat volume, 117-118 linear demand, 228, 230-231
promotional discount, 279 purpose, 226
promotions, 263 residual price elasticity, 251
baseline sales. See baseline sales calculating, 254-255
complications, 279-280 complications, 255-256
coupons. See coupons purpose, 252-254
evaluating temporary price promotions, 264 response bias, 59
long-term effects of, 274-275 responses, customer survey, 116
profitability, 271 results of marketing metrics survey, 13
rebates, 275-277 retail margins, 362
redemption rates. See redemption rates retail profit, Prestige Luggage, 363
short-term promotional objectives, 263 retailers, apparel, 161
prospect lifetime value (PLV), 173 retention, 48, 158-159
calculating, 173-174 versus acquisition, 176-178
complications, 174-176 retention rate, 156, 159, 170
purpose, 173 return, 337
versus customer lifetime value, 174-176 return on assets (ROA), 342, 369-370, 372.
prospects, 200 See also DuPont Model
pull marketing, 203 return on capital (ROC), 342
purchase intentions, 53 return on incremental marketing investment
purchases, 200 (ROIMI), 352
push marketing, 203 return on invested capital (ROIC), 342
Return on Investment (ROI), 338, 342-343, 357
return on marketing investment (ROMI),
Q–R 338-339, 350-351
quantifying budgeting, 354
customer profit, 167 calculating, 351-352
market share, 34-35 complications, 354


media exposure return on marketing same stores growth, 126-128
investment, 354-355 sample selection, customer satisfaction, 59
purpose, 351 sampling size of marketing metrics survey, 11-12
return on net assets (RONA), 342 search engine marketers, 327
return on Sales (ROS), 338, 340-342, 357, 369 search engines, 325-327
returns and target, 108 seasonal variations (ROI), 343
revenue attributable to marketing, 352 second-price auctions, 228
revenue from new products, 125 secondary line competitive injury, 251
revenue market share, calculating, 33 segment utilities, 112
revenue return to incremental marketing, 352 segmentation by geography, 55
revenue return to total marketing, 352 segments
revenue share of requirements, 46 BDI, 42
reward structures, supply chain metrics, 213 CDI, 42
rich media display time, 291, 317-318 conjoint utilities, 147-149
rich media interaction rate, 291, 318-319 selling price, 72, 76
ROA (return on assets), 342, 369-370, 372. See separating customer responses from
also DuPont Model non-customer responses, 54
Robinson-Patman Act, 251, 285 served market, 34-35
ROC (return on capital), 342 service levels, 209-210
Rogers, Martha, 167 Shames, Erv, 364
ROI (return on investment), 338, 342-343, 357 share of category, 39
ROIC (return on invested capital), 342 share of requirements, 45-47
ROIMI (return on incremental marketing share of shelf, 208
investment), 352 share of voice, 313
ROMI. See return on marketing investment share of wallet, 44-47
RONA (return on net assets), 342 shopping basket margin, 218
ROS (return on Sales), 338, 340-342, 357, 369 shrinkage, 214
signals, 54
SKU (stock keeping unit), 86, 215
S slope, optimal price, 244
salaries. See sales force compensation slotting allowances, 101
sales, decomposing, 371-372 social networking, friends/followers/
sales force compensation, 195 supporters, 333-335
calculating, 196-197 sole usage, 47
incentive plans, 197-198 spreadsheets, calculating NPV, 350
purpose, 196 State Farm, 157
sales force effectiveness, 192 statistical units, 88, 90
calculating, 192-195 stepped payments, 100
customer service, 194 store versus brand measures, 208
expenses, 194 supplier selling price, 75
purpose, 192 calculating, 77
sales force funnel, 199 calculating average, 85
sales force objectives, 189-191 supply chain metrics, 209
sales force territories, 186 complications, 212-213
balancing, 187-188 inventories, evaluating, 213
comparing, 188 inventory days, 211-212
estimating size of, 189 inventory tracking, 211
purpose, 187 inventory turns, 211
redefining, 189 out-of-stocks, 210
sales force tracking. See pipeline analysis purpose, 209
sales funnel, 184, 201-202 reward structures, 213
sales goals, 190-191 service levels, 210
sales pipeline, 184 supporters, 293, 333-335
sales potential, 182, 186-191

Index 413

surveys, 114 U–V
customer satisfaction, 59
customer survey responses, 116 under-servicing, 187
marketing metrics survey unit margin, 69-71
cautions about, 10-11 unit market share, 33
rankings, 21-24 unit share of requirements, 46-47
results, 13 units, 69
sampling size, 11-12 USAA, 157
metric usage survey, 385-390 usage, 54
user behavior, web sites, 328-331
T value of future period, 128-129
variable cost per unit versus total cost
target market fit, 125 per unit, 96
target rating points (TRPs), 288, 296-297 variable costs, 91
target revenue, 106-107 calculating, 91-95
target volume, 68, 106 classification of, 96
target volumes not based on target profit, 108 purpose, 91
targets, profit-based sales, 106-107 Venn diagram, 304
terminal values, 349 video interactions, 320
territories. See sales force territories visitors, 327-328, 331-333
test markets. See also trials visits, 292, 327-328, 331-333
assumptions, 120-121 volume projection, 112-113
awareness, 114 conjoint utilities, 150-151
distribution, 115 spreadsheet, 119
simulated results and volume projections,
trial volume, 114
theoretical price premiums, 226 W–Z
three (four) firm concentration ratio, 38 Wal-Mart, 27, 345
threshold, 289 warm leads, 200
threshold response model, frequency response wear-in, 310
functions, 306 wear-out, 310
time, measuring market share over, 35 web pages. See also Internet
tolerable discrimination, 285 hits, 314-315
top of mind, 53 pageviews. See pageviews
total cost, 92, 95 visitors, 327-328, 331
total cost per unit, 94-96 visits, 327-328, 331
total coupon cost, 276 web sites
total distribution, 184, 207 traffic, assessing, 314-315
total number of active customers, 45 user behavior, 328-331
total outlet sales, 208 weighted contribution margin,
total selling costs, 98 cannibalization, 132
total variable selling costs, 98 weighted share of sales allotment, 190
total volume, 118-119 whale curve, customer profit, 167
“the trade,” 278 willingness to recommend, 56-57
trade satisfaction, 59 willingness to search, 62-63
trial rate, 113-115 workload, 186-187, 198
trial volume, 116-117
trial-repeat model, 124 year-on-year growth, 111, 125
trials, 112, 121, 124. See also test markets Young & Rubicam, 137, 139
discounted trial, 124
forced trial, 124 Zellner, Arnold, 377
purpose, 113
repeat volume, 117
total volume, 118-119
TRPs (target rating points), 288, 296-297


Marketing Metrics

More Related Content

What's hot (20)

Similar to Marketing Metrics (20)

More from Fernando Holanda (20)

Recently uploaded (20)

Marketing Metrics