Webhound turning webdataintointelligence

WebHound —
Turning Web Data into Intelligence

A SAS White Paper

Table of Contents

Executive Summary .......................................................................................................1

Introduction ....................................................................................................................2

Boundless Bounty to Realism ......................................................................................2

Enterprise-Wide and Customer-Centric.......................................................................4

The Technology Implications of Web Commerce .......................................................4
Data Collection and Management ................................................................................5
Transforming Information into Meaningful Intelligence.................................................7
Reporting Information ...................................................................................................8

A Total Web Analysis Solution .....................................................................................9
WebHound Web Traffic Warehousing ........................................................................10
WebHound Analytics ..................................................................................................11
WebHound Reporting .................................................................................................11
Unsurpassed Data Warehousing and Analytics .........................................................12
SAS Web Analysis Tools in Action .............................................................................12

Summary .......................................................................................................................13

WebHound — Turning Web Data into Intelligence

Executive Summary
Internet/intranet usage is growing at staggering rates. So quickly that many enterprises have
found it difficult to guarantee quality service for their Web site visitors or to maximize their
investment in e-business.

Enterprises require accurate and up-to-date information regarding their Web site traffic: Who is
visiting? How long do they stay? What are they looking at? How do Web interactions mesh with
other customer contact channels? With supplier and inventory activity?

Answers to these and many other questions are fundamental to extracting maximum value from a
Web site, whether the goal is education, philanthropy or revenue.

The good news is that every visit to a Web site leaves a trail with the potential to reveal how
visitors use the site, and whether the site delivers what visitors came to get. Web traffic analysis
measures and analyzes navigation paths, whether or not the visit concluded in a transaction, and
transforms that information into usable intelligence. Effective application of the intelligence
generated by such analysis yields insights that help get visitors to return more often and conclude
more successful transactions on your Web site.

SAS WebHound is a sophisticated Web traffic analysis and reporting tool that empowers
enterprises to better understand online customers and in turn to maximize the value of Web
channels. With WebHound, e-business managers can learn who their online customers are,
where they come from, what pages they like and didn’t like, how often they visit the Web site, and
much more.

WebHound offers a powerful data-gathering and summarization, flexible and dynamic Web
reporting through any browser, including wireless devices, and complete scalability to handle
floods of Web traffic data. WebHound integrates with SAS industry-leading data warehousing and
analytical technologies that cleanse and manage data for highest quality, that “mine” value out of
volumes and layers of data by applying industry-leading analytics, that gather information from
external operational systems with various database architectures, and more.

WebHound is more than a
tool to collect and analyze
Web data. It integrates
seamlessly with financial,
advertising, sales, supply
chain systems and more to
contribute to a total
enterprise performance
management strategy.

1


Introduction
Successful companies use e-channel data to improve customer relationships, to determine the
How do you gauge and
effectiveness of their investment in the e-channel, to better understand the organization’s
improve the effectiveness of
operational supply chain and to integrate the information of the entire enterprise. According to a
your e-channel in complex
survey conducted by SAS and Hovrath+ Partners in 2001, 72 percent of all companies say they
organization with multiple
cannot accurately measure the effectiveness or value of their e-business investment. That is
customer touch points and
particularly problematic in a tight economy when every dollar and competitive advantage counts.
interrelated functions? Every
day, you amass huge
Looking further, the flexibility, immediacy and ubiquity of the Web make it an ideal proving ground
quantities of online data, but
to test marketing new products, display and pricing strategies, special offers and promotions, and
can you effectively report and
more.
analyze that data? And how
does that information directly
relate to your customers,
suppliers and organization? Boundless Bounty to Realism
A few years ago, enterprises spent vast sums of money on advertising to draw users to their Web
sites. When the dot-com economy was surging, it seemed enough to draw the masses to your
Web site and less critical to look closely at what those visitors did once they got there.

That picture has been reshaped by several market forces:

· For one, sales volume from online channels has steadily increased to the point where it is
worthwhile to take it seriously and award it closer scrutiny. US consumers who research
products on the Web will spend more than $199 billion online and $632 billion offline by 2005.
Internet-equipped shoppers will account for 75 percent of all expected retail revenue from all
sales channels combined. If you want to drive this from current to future trends—8% of retail
transactions are done on the Web, but 50% of those transactions rely on information
researched on the Web.

· Second, the proliferation of Web storefronts has made it harder than ever to differentiate your
site and gain customer loyalty. Greater competition, coupled with distribution and inventory
woes, and the cost and complexity of IT infrastructures, has made profit margins leaner than
ever. It’s no longer enough to bring users to your site; you have to maximize the productivity
of each visit.

· Third, the complexity of many sites makes it difficult or impossible to capture the intricacies of
a user’s experience with simple tracking methods. Nor do traditional tracking methods
correlate users’ experiences online and offline, across all touch points.

2


As a result of these market forces, it is more critical than ever to:

· Improve the acquisition, retention and success of each customer relationship.

· Maximize the productivity of e-business channels.

· Maximize the synergy between the Web and other channels.

· Improve the effectiveness of marketing campaigns.

· Gain more meaningful insights into content delivery, merchandising mix and strategies.

In meeting these business challenges, the good news is that Web activities automatically
generate an enormous amount of quantifiable data to help enterprises know where they stand.
This data is easily gathered through Web traffic logs and activity reports.

The down side is that the data isn’t very meaningful unless it can be transformed into useful
knowledge. That reality brings some familiar scenarios.

Chief marketing officer:
”I have multiple campaigns running and I seem to be throwing money out the window.
We really need to understand where our money is going and how effective it is.”

Vice president of e-business:
“People come to my site because they need to get materials. I would like to send
them emails and push them to my site to inform them of what is going on in the
industry and various new products too, but I am afraid of spamming.”

Chief operating officer:
”We need to integrate our e-business with supplier relationship management systems
so we can get a better view of our operational processes end to end.”

Chief information officer:
“Our Web channel churns out several gigabits of data a day, and it takes well over
10 hours to process the Web logs, much less get reports to run.”

The underlying themes speak to some serious operational issues. How do you collect data from
diverse sources, integrate it into one usable resource, and ensure its integrity and validity for
analysis? How do you distill masses of raw data into actionable intelligence? How do you target
customers more effectively and gain maximum value from marketing dollars? How do you merge
e-business into overall business, rather than treating it as an adjunct activity?

3


Enterprise-Wide and Customer-Centric
Profitability depends on establishing and managing a company-wide strategy. It’s not enough to
manage functional units toward individual successes and expect their merits to trickle up through
the organization.

For years, companies have been breeding competitiveness and independence in organizational
units, punishing or rewarding them for unit-level results, and embracing efficiency gains in one
area even at the cost of inefficiencies in others. Companies now have to foster company-wide
perspective where they measure performance in greater context across the value chain.
Management needs to establish organization-wide vision, align all functional units toward that F
bigger picture of success and reward units for success on organization-wide metrics. i
o
Abandoning the silo mindset also means recognizing that each customer relationship is shaped at c
multiple touch points. Organizations already collect massive amounts of data that reflect day-to- a
day operations in each functional area. Typically, that data is captured in silo transactional k
systems, data warehouses and unstructured documents that lack form and shape for analysis and e
for optimal contribution to enterprise-wide performance management. Organizations must build a a
comprehensive business model that binds information together and defines metrics at the p
process, function and organizational levels. o

The Technology Implications of Web Commerce
Software tools are available to help enterprises meet the business challenges defined by the
Web-enabled economy. A comprehensive Web intelligence software solution provides:

· A single, customer-centric data warehouse that brings together all relevant information about
customers and competitors, including market research data from internal and external
sources.

· Data mining techniques that uncover hidden truths about customers, products and campaign
results.

· Real-time and near-real-time reports that allow quick adaptation to evolving markets and
trends, as well as customized responses to customer inquiries.

4


With these capabilities, Web managers can:

· Optimize the Web channel based on a single customer view derived from all channels.

· Target the most profitable customers, based on an understanding of lifetime value and
changing needs, and identify opportunities to cross-sell, up-sell and retain them.

· Create customized messages, targeted loyalty programs and product offerings that build
customer loyalty, promote the most profitable products and services, and improve the ratio of
browsers to buyers.

· Increase the likelihood that a customer’s visit will culminate in a purchase, increase the
number of visits and increase the average purchase amount per customer.

· Increase advertising revenues by showing advertisers hard proof of the popularity and
effectiveness of the Web site and by offering targeted ads to specific customer groups.

· Gauge the effectiveness of the Web site as a sales-starter and as a fulfillment channel for
campaigns initiated via other channels.

Let’s take a look at the key technology components that make these gains possible:
data collection and management, data analysis and reporting.

Data Collection and Management

Web data is growing exponentially, increasing by about 50 percent every six months for typical
companies as e-business becomes more widely accepted. With broadband and mobile access
making it easier and faster to transport data, companies are likely to generate, save and use data
at an even faster growth rate in the future.

Furthermore, Web data is no longer confined to just Web server logs. There are application
servers, content management systems, customer and supplier relationship systems, and many
other systems that all have become channels that a Web visitor touches, on one level or another.
An effective solution captures all the data generated by a site visit, not just the data captured
through transactions, but also from visits that did not culminate in a transaction.

The large increase in Web activity and the multiple touch points the Web has created has placed
a great demand on the technology infrastructure to gather, organize and manage the data. This
information explosion fueled the development of specialized Web traffic data warehouses, whose
sole purpose is to manage and organize the vast information collected by the e-channel. Unlike
relational databases, these specialized data warehouses have been designed to load, store and
access the vast amounts of information obtained from the e-channel.

Data warehouses: customer-centric, enterprise-wide. To generate a cohesive picture of the
enterprise and its Web strategy, operational systems must have access to a single version of the
truth, one that assembles information about the customer and the enterprise from across touch
points and operational systems.

5


The data management strategy must reflect the reality that everything is interdependent. For
example, supply affects inventory affects fulfillment affects customer satisfaction affects return
visits to the Web site. Web site appearance and function affects customer frustration affects
dropped shopping carts affects profitability. A recent statistic showed that 80% of shopping cart
applications are abandoned before a purchase is made.

Reflecting those interdependencies, today’s data warehouses are likely to be distributed across
the enterprise and managed as one resource by means of integration systems, rather than
actually residing in a single warehouse location. For example, a complete enterprise data
management solution:

· Includes elements for extracting data from discrete systems, transforming it to enable
meaningful analysis (no apples with oranges), and transferring it to systems/servers for
analysis, reporting and storage.

· Integrates with cross-functional operational systems, including back-office enterprise-
resource planning systems and front-office systems.

Data warehouses: validated. Information in a single Web traffic warehouse may very likely come
from Web logs, content management systems, customer databases or an enterprise resource
planning (ERP) system. Each of these systems would be loaded into the warehouse for analyzing
the e-channel, but each data source may also introduce unnecessary, duplicate or inaccurate
information.

Web logs, for example, are filled with images and “spider” requests such as GIF image files and
search engine requests that can fill a Web log with extraneous data. There may also be multiple
entries of the same data across different databases or the same database. For instance, a Web
user may have multiple logins for the same site or you may have the same user stored under
different names in different databases. Such data can lead to erroneous results and
misinformation.

That’s where data cleansing comes in. An effective Web analysis system must ensure data
integrity by removing duplicate records (de-duping), identifying multiple logins, validating field
records and such.

Similarly, the system must be adept at managing metadata — the behind-the-scenes data that
defines how real customer data is handled. For example, metadata tells the system how to
propagate the new address information received when a customer moves or where information
resides in the data strategy. Metadata tells the system how to interpret and relate information
among different databases, when to declare a record inactive, how to encode user entries, or
which tables to access when adding new records.

With proper metadata lifecycle management, every change made to the data is immediately
propagated throughout the warehouse, duplicate data is less likely to enter the warehouse, and
analysis will be more accurate.

6


Data warehouses: what kind of data gets collected? Warehouses can gather information
about content and visitors, together yielding intelligence about how users navigated and used that
content:

· From the content point of view: How many pages are being requested? What content is
getting the most requests? What is the shopping cart abandonment rate? What is the most
navigated path within the site? From what point are people leaving the site? Content-focused
knowledge helps you better manage the message-delivery architecture of your e-channel.

· From the visitor point of view: How many unique visitors came to the Web site? How many
sessions did the Web site support? What categories of visitors came to the Web site most?
Who are the most loyal users? Visitor-focused knowledge helps you better target the Web
site to your most profitable Web customers.

Combining content and visitor information produces rich insights into overall e-channel
effectiveness. For example, information on:

· The number of page views and unique visitors to your site becomes information on which
visitors selected which pages.

· The most loyal users of the e-channel and most-requested content becomes information on
what your best repeat visitors are seeking from the site.

· The type of visitors that come to the Web site, combined with the most-navigated path,
becomes information on how you can design the site for the specific needs of your users.

These are a few of the many permutations that can emerge when you combine content
information with visitor information, thereby, creating a knowledge base for which to analyze Web
content and user behavior.

Transforming Information into Meaningful Intelligence

Advanced analytics enable Web managers to “mine” the data warehouse to transform masses of
data into meaningful Web channel intelligence, on a formal or ad hoc basis. To be effective, the
analysis system must not only offer industry-standard algorithms and report structures, but also a
way to apply unique company metrics and logic to any information in the data warehouse.

How many page views are expected for the coming month? How effective is the e-newsletter?
Why is the shopping cart abandonment rate up during this marketing campaign? How did Widget
sales respond when we moved it to the high-level page? Which pages tended to be viewed
together? What are the cost and benefits of the e-channel to the entire organization?

Armed with this information, e-business managers can create highly tailored Web-based
campaigns, fine-tune the site for optimum performance, and identify the best ways to appeal to
high-value visitors.

7


There’s a hitch though. To support truly informed decisions, an effective Web analytics system
must provide context across operational functions. Thus, the system should have access to
shared data warehousing with deep data mining capabilities, not just limited pass-through of
information from application to application.

It’s not enough for discrete systems to be willing to inject data into neighboring systems.
Companies must fully integrate systems from functional units into an enterprise framework that
supports business and information value chains. Business value chains provide efficiencies
across business processes without regard for organizational boundaries. Information value chains
provide paths for information to flow along those business value chains.

An effective strategic system seamlessly integrates applications serving all the functional areas,
including finance, customer relationship management, marketing, human resources, IT, supplier
relationship management and e-business.

SAS leverages an enterprise data warehouse offering as the platform for building a complete, fully
integrated suite of applications to manage enterprise performance across all channels. The
system can read data from virtually any source in virtually any form, consolidating operational
data from disparate systems into a powerful information resource, with full metadata management
and data cleansing functions.

Reporting Information

Naturally, a Web traffic data warehouse is only as useful as its information is accessible. The
results of data analysis must be delivered to the right person at the right time, in an
easy-to-understand format that distills actionable intelligence from the masses of data. No wading
through pages of unproductive clutter to decipher meaningful content.

Why not exploit the power of the Web for Web reporting? Web reporting gives e-business
managers point-and-click access to static reports that address everyday business requirements,
drill-down capability to look deeper into multidimensional databases to reveal what’s behind those
numbers and hyperbolic tree views that visually depict user behavior on the Web.

8


A Total Web Analysis Solution
SAS WebHound supports an effective Web traffic warehouse by collecting e-channel data,
transforming that data into cohesive and usable information, and holding the data for later
analysis and reporting. As such, WebHound:

· Extends the value of the SAS information delivery architecture and enterprise data
management umbrella solutions to include e-channels as well.

· Integrates seamlessly with SAS solutions for data cleansing and validation, data analysis,
static and ad hoc reporting, and Web reporting to create a total Web management solution.

· Integrates with front- and back-office systems from SAS and other vendors, such as
PeopleSoft, Oracle, Siebel and BroadVision.

· Is a chief component of the SAS e-Discovery solution for customer relationship management
(CRM), which combines Web traffic behavior with purchasing, customer service,
demographic and psychographic data for a comprehensive view of customer relationships.

· Emerges as a key value component of the SAS or multivendor enterprise performance
management solution with balanced scorecard.

Figure 1: WebHound Report Viewer

Let’s take a snapshot look at how WebHound measures up in the functional areas mentioned
earlier: data collection and management, data analysis, and reporting.

9


WebHound Web Traffic Warehousing

Data extraction. WebHound collects the data from all possible Web information sources (such as
Web logs, application server logs and other Web data collectors) — from flat files, ERP systems,
hierarchical files and relational databases. Using multi-threaded processing to take advantage of
the entire hardware configuration of the WebHound server, the software then processes the data.

WebHound cleanses the data by selectively filtering out all images, spider requests and any other
custom filters, such as internal Internet Provider (IP) addresses. During this phase, WebHound
adds information from other databases, and compares and updates any metadata.

Data transformation. After data extraction, WebHound transforms and prepares the data to be
loaded into the WebHound warehouse. During this process, WebHound identifies sessions and
user navigation paths, compresses data without any loss of information, and changes from a
transactional database schema to the warehouse schema, tuned for reporting, ad-hoc analysis
and data mining.

Data transfer. WebHound loads the transformed data into three types of databases in the
warehouse:

· Summary databases store data for long time periods yet enable quick access to information.

· Detailed databases store less historical information, but support a more in-depth view.

· Multi-dimensional databases (MDDBs) are the backbone of hybrid online analytical
processing (HOLAP) and allow for drill-down and viewing of multiple data elements within
one view.

Physically, data can be stored in SAS tables, Scalable Performance Data Server (SPDS), and
Relational and Multidimensional databases. The WebHound data warehouse solution offers an
extensive array of products to manage enterprise data, such as managing metadata; integrating
with front- and back-office systems such as SAP, Oracle Financial, PeopleSoft, BroadVision,
BlueMartini and others; and importing and exporting data to Oracle, MS-SQL server and others.

Through this cross-functional integration, WebHound and its companion modules link e-business
functions to customer-profile and other databases to track each customer's transactions,
personalize Web pages based on preferences and history, and automate the process of
responding to customers.

10


WebHound Analytics

WebHound integrates with SAS Enterprise Miner to “mine” valuable insights from the warehouse.
Enterprise Miner distills useful information from utilization data as well as textual information, such
as from chat rooms and user surveys. Enterprise Miner enables scoring, clustering and
segmentation of customers based on user-defined metrics. This product also lets you analyze the
impact of promotions, evaluate trends and even forecast page views and visitors, completed
transactions and any other metrics relevant to your business.

Combined with WebHound data warehousing and online analysis, Enterprise Miner creates a
synergistic, end-to-end solution that addresses the full spectrum of knowledge discovery.

WebHound Reporting

WebHound integrates with SAS Enterprise Guide to create reports from more than 200 standard
templates that report on the basic metrics necessary to run an e-channel. However, SAS
recognizes that each business is unique and has it’s own organizational needs and requirements.
Thus, WebHound’s reporting infrastructure lets you customize the standard templates, or place
fully custom reports directly into WebHound’s metadata layer, making your report a standard
WebHound report.

WebHound goes beyond standard and custom batch reports, and offers a user-friendly, ad hoc
reporting interface that lets you query the information stored in the data warehouse on demand.
WebHound’s use of advanced data structures in MDDBs allows a business user to drill down into
the data and gain even more depth and clarity, putting ad-hoc analysis in the hands of the people
who really need it and removing the IT overhead.

WebHound also offers intuitive visual analysis to examine the complex architecture of a Web site,
presenting useful information that is normally unobtainable in a single view. A tree view, a
reporting structure designed specifically for hierarchical representation of data, shows various
levels of activity across different types of data. So, not only do you see the most active paths, but
you also see the most visited paths (and most abandoned paths).

To enable team members to view and collaborate on e-business intelligence, WebHound supports
publication and subscription of information through an enterprise collaboration network with
Microsoft SharePoint Portal Server and Digital Dashboard. With Microsoft Digital Dashboard,
WebHound intelligence can be rapidly and effectively shared, to improve decision-making and
competitive advantage.

11


Unsurpassed Data Warehousing and Analytics

Globally recognized as the industry leader in analytics and data warehousing, SAS ensures
that enterprises have the most accurate, consistent and reliable analyses of business
information.

Our warehousing solution has been updated to include data quality tools and an improved
application programming interface (API) that lets you build a metadata-driven or business
rule-driven process for transforming data into data warehouse tables.

Once data is properly warehoused, you can use our cutting-edge statistical capabilities to
create usable market intelligence from that data. Enhanced analytics include new capabilities
for multiple time series forecasting and demand planning, new procedures for dealing with
missing data, and tools for non-parametric modeling.

Data mining in an enterprise performance management (EPM) system returns tremendous
bottom-line impact as it turns data into information, information into knowledge, and
knowledge into greater business value.

SAS Web Analysis Tools in Action

· A state-chartered social work group at a major university supports a Web site that presents
dynamic statistics and analyses about state welfare reform programs. With an out-of-the-box
WebHound solution from SAS, the group’s project director and IT managers can immediately
see the results of announcements and site changes, generate documentation to support
project funding, and plan future development in ways that will best satisfy the site’s critical
public mission.

· A highway safety research center at a major state university develops project-specific Web
sites for contracting agencies that translate accident research into practical interventions at
the local, state, and national level. The center relies on WebHound to exploit maximum
advantage from the Web for communicating the center’s life-saving research and outreach
projects.

· A leading catalog marketer with $575 million in sales in 2000 uses SAS e-intelligence to
analyze its Web traffic and buying patterns and match that knowledge against corporate
strategy, to “generate actionable data that helps us make decisions in real time,” said the
retailer’s vice president and general manager for e-business. “We’ve noticed that certain
product categories have much higher conversion rates from Web traffic to buying… [which]
helps us present merchandise in a logical format that leads to more buying.”

· A leading bricks-and-clicks retailer turned to SAS Web analysis to provide “a dashboard to
determine what all this Web activity really means, and more importantly, the direction we
should be moving as a result,” said the retailer’s director of e-business. “It’s a total strategy…
one that goes beyond ‘here’s what’s happening’ to help us understand ‘and here’s what you
can do about it.’”

12


Summary
Web traffic analysis is crucial for any organization that has a web site. Commercial enterprises
want to increase revenues and profitability. Public agencies want to increase the visibility and
utilization of their public-service Web sites. Not-for-profits want to enhance their missions and use
limited funds for greatest charitable gain. Government and educational institutions want to
effectively provide services and information to their publics.

Even though these sectors have different missions — profit, charity or information-sharing — their
Web goal is shared: provide the best possible service to their audiences, by:

· Providing the most relevant information to Web site visitors.

· Improving communication via the intranet and extranet.

· Optimizing the acquisition process to maximize revenue.

· Promoting the most popular information.

· Creating a convenient experience by more effectively organizing information on the site.

· Improving user satisfaction by reducing the number of clicks it takes to access information.

Web traffic analysis enables an organization to determine where their users come from, what path
they take before finding the information or product they want, what path they take before leaving,
which pages are most/least popular, how many clicks did it take for users to reach their goal, and
so on.

SAS WebHound is a Web analysis tool that captures, stores, manages, analyzes and reports
Web usage data and supports informed decisions about Web channels. SAS WebHound
processes large volumes of Web-generated data, integrates it with other online and off-line
sources, validates and prepares the data for analysis, passes the data to award-winning SAS
analytics software, and dynamically reports the results through any Web browser.

SAS WebHound lets Web analysts understand clearly how their Web sites are performing and
enables them to support informed decisions that optimize e-channels for premium results.

13

World Headquarters SAS International
and SAS Americas PO Box 10 53 40
SAS Campus Drive Neuenheimer Landstr. 28-30
Cary, NC 27513 USA D-69043 Heidelberg, Germany
Tel: (919) 677 8000 Tel: (49) 6221 4160
Fax: (919) 677 4444 Fax: (49) 6221 474850
U.S. & Canada sales:
(800) 727 0025 www.sas.com

SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA
and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies.
Copyright © 2002, SAS Institute Inc. All rights reserved. 47646US.0102

Webhound turning webdataintointelligence

More Related Content

What's hot (20)

Viewers also liked (7)

Similar to Webhound turning webdataintointelligence (20)

More from Automotive Digital Marketing Professional Community (20)

Webhound turning webdataintointelligence