Data Management for Business Intelligence

Data Management for Business
Intelligence

March 2008

Data Management for Business Intelligence
Page 2

Executive Summary
As the sheer amount of information available for Business Intelligence (BI) Research Benchmark
applications has grown, and the sophistication of BI capabilities has Aberdeen’s Research
accelerated, organizations are striving to improve time-to-information for Benchmarks provide an in-
business users. Data storage and retrieval decisions have started to be made depth and comprehensive look
based on cost rather than fulfilling a business need, and this has led to into process, procedure,
compromises that impact business performance. In January and February of methodologies, and
2008, Aberdeen Group conducted a research program among 333 technologies with best practice
respondents from several industry sectors, geographies, company sizes, and identification and actionable
job roles. The program included an on-line survey and several one-on-one recommendations
interviews. Thirty six (36) respondents were disqualified due to their status
as either a software vendor or solution provider, leaving a total of 297
respondents that were included in the analysis. This report investigates the
strategies, capabilities, and supporting technologies that Best-in-Class
companies are employing to alleviate the pressure to accelerate the delivery
of actionable information within their enterprises.

Best-in-Class Performance
Aberdeen used four key performance criteria to distinguish the Best-in-
Class from Industry Average and Laggard organizations:
• Time-to-information for users of BI applications. Best-in-Class "Our data is not organized in a
companies achieved an 8% mean average decrease in time-to- way that can provide insight. As
information in the past 12 months, compared to a 2% mean average a Six-Sigma black belt, my job is
to influence management and
decrease among Industry Average companies, and a 1% mean increase
IT to organize information
among Laggards. correctly. The process by
• Number of end-users receiving access to BI applications. Best- which we make decisions based
in-Class companies achieved a 10% mean average increase in the on KPIs is a struggle. Fulfillment
number of users receiving access to BI applications in the past 12 is one of the big KPIs and
involves an end to end process
months, compared to a 5% mean average increase among Industry
from order received to actual
Average companies, and a 3% mean average increase among Laggards. order shipped. We have also
• End-user productivity and efficiency. Best-in-Class companies acquired many companies with
achieved a 12% mean average improvement in end-user productivity and many systems to manage and
efficiency (measured as time spent looking for business information many data sources. It takes a
lot of time to extract the data,
within BI applications), versus a 3% mean average improvement among
and on the business side, to
Industry Average companies, and a 7% mean average decrease in determine what data to
productivity among Laggards. extract. Many things are missing
• Data management infrastructure costs. Best-in-Class companies when we try to aggregate data
achieved a 7% mean average decrease in data management from disparate systems."
infrastructure costs (hardware and software) to support BI applications ~ Six Sigma Process
in the past 12 months versus the prior 12 months, compared to a 2% Improvement Lead, Large US
mean average decrease among Industry Average companies, and a 6% Manufacturer
mean average increase among Laggards.
In combination, these metrics provide a means by which respondents can be
benchmarked based on their ability to achieve improved business
performance.
© 2008 Aberdeen Group. Telephone: 617 723 7890
www.aberdeen.com Fax: 617 723 7897

Page 3

Competitive Maturity Assessment
Survey results show that the firms enjoying Best-in-Class performance
shared several common characteristics. There is no one specific capability
that is more important than another, but rather a mix of capabilities that
must be balanced in order to achieve improved data management
performance for BI deployments. These include:
• The ability to understand end-user needs early in the implementation
process
• The ability to automate the integration and cleansing of data from
multiple sources
• The establishment of query optimization and prioritization rules for data
access
• The ability to measure and plan for data and end-user growth

Required Actions
In addition to the specific recommendations in Chapter Three of this
report, to achieve Best-in-Class performance, companies must:
• Realize that data management for BI initiatives requires a balance of
several factors. This includes gaining a deep understanding of business
user information requirements before a data management roadmap or
strategy can be developed and executed.
• Understand both the human and technological elements of data
management. End user requirements must be understood at the
beginning, and an investment in training should be made to ensure rapid
and deeper adoption of BI applications.
• Be aware that the nature of the data queries and reports should be
investigated to uncover poorly designed or structured queries and
corrective action taken.


Page 4

Table of Contents
Executive Summary....................................................................................................... 2
Best-in-Class Performance..................................................................................... 2
Competitive Maturity Assessment....................................................................... 3
Required Actions...................................................................................................... 3
Chapter One: Benchmarking the Best-in-Class ..................................................... 6
Business Context ..................................................................................................... 6
The Maturity Class Framework............................................................................ 7
The Best-in-Class PACE Model ............................................................................ 8
Best-in-Class Strategies........................................................................................... 9
Chapter Two: Benchmarking Requirements for Success ..................................12
Competitive Assessment......................................................................................13
Capabilities and Enablers......................................................................................14
Chapter Three: Required Actions .........................................................................24
Laggard Steps to Success......................................................................................24
Industry Average Steps to Success ....................................................................24
Best-in-Class Steps to Success ............................................................................25
Appendix A: Research Methodology.....................................................................27
Appendix B: Related Aberdeen Research............................................................29
Featured Underwriters...................................Error! Bookmark not defined.

Figures
Figure 1: Percentage of Data Volume Growth during the Past 12 Months by
Percentage of Respondents ........................................................................................ 6
Figure 2: Top Five Pressures Driving Data Management Initiatives within BI
Deployments .................................................................................................................. 7
Figure 3: Top Five Strategic Actions.......................................................................10
Figure 4: ROI Time-Frames for Business Intelligence Projects ........................11
Figure 5: Ability to Optimize Data Queries..........................................................15
Figure 6: Top Organizational Capabilities .............................................................16
Figure 7: Knowledge Management - Data Prioritization and User Demand
Intelligence ....................................................................................................................17
Figure 8: Ability to Measure and Manage Growth of Database and Data
Warehouse Assets......................................................................................................18
Figure 9: Best-in-Class Data Integration and Data Cleansing Capabilities.....18
Figure 10: Best-in-Class versus All Others - Technology Enablers Currently
Deployed .......................................................................................................................19
Figure 11: Best-in-Class versus All Others – IT / Consulting Services
Currently Utilized .......................................................................................................20
Figure 12: Best-in-Class Technology Enablers Current versus Planned
Deployment ..................................................................................................................21


Page 5

Tables
Table 1: Top Performers Earn Best-in-Class Status.............................................. 8
Table 2: The Best-in-Class PACE Framework ....................................................... 9
Table 3: The Competitive Framework...................................................................14
Table 4: The PACE Framework Key ......................................................................28
Table 5: The Competitive Framework Key ..........................................................28
Table 6: The Relationship Between PACE and the Competitive Framework
.........................................................................................................................................28


Page 6

Chapter One:
Benchmarking the Best-in-Class
Business Context
Fast Facts
The volumes and complexity of data within business intelligence applications
are growing, resulting in a new set of challenges for organizations that are √ 88% of all respondents (and
attempting to deliver more relevant information to the right people at the 95% of Best-in-Class
respondents) rate "ease of
right time. "Data management" involves capabilities, methods, and
use for end-users" as the
technologies that companies may employ separately or in combination to most critical criteria when
alleviate this challenge. selecting a data management
Research completed in December 2007 for Aberdeen's Operational BI: vendor for BI initiatives - yet
41% complain that end-users
Getting Real-Time about Performance report revealed that data volumes are
have not provided well-
increasing at alarming rates within many organizations (Figure 1). Seventy- defined information needs
eight percent (78%) of respondents reported growth of data volumes, and
40% reported growth rates of 21% or more over the past 12 months. √ 74% of all respondents (and
84% of Best-in-Class
Figure 1: Percentage of Data Volume Growth during the Past 12 respondents) prioritize
system scalability for the
Months by Percentage of Respondents handling of increased users,
40% 38% data volumes, and
complexity as the most
critical criteria when
selecting a data management
30% 28%
vendor for BI initiatives
% of Respondents

21%
20%

12%

10%

1%
0%
1 - 20% 21 - 40% >40% Stayed Same Don't Know
Data Growth Rate

Source: Aberdeen Group, February 2008

As data volumes have increased, organizations have struggled to maintain
rapid "time-to-information" performance for business users. In January and
February of 2008, Aberdeen Group conducted in-depth research among
over 330 business and IT decision-makers to identify the primary business
pressures driving companies to focus on data management strategies that are
supporting business intelligence initiatives. The predominant pressure driving
companies to adopt new data management strategies for BI applications is
the need to reduce time-to-information for end-users (Figure 2).


Page 7

Figure 2: Top Five Pressures Driving Data Management Initiatives
within BI Deployments
0% 20% 40% 60%

Reduce time-to-
information for 51%
end-users

Deliver BI
capabilities to 27%
more end-users

Simplify data
access for end 25%
user applications

Improve ease-of-
use of BI 23%
tools/applications

Add new sources
of data not
currently 17%
addressed by
existing BI

Source: Aberdeen Group, March 2008

The following research findings describe the strategic and tactical
approaches, as well as the technologies and services that are being adopted
and deployed by companies that are best-equipped to deal with these
pressures.

The Maturity Class Framework
Aberdeen used four key performance criteria to distinguish the Best-in-
Class from Industry Average and Laggard respondent organizations:
• Time-to-information for users of BI applications. Percent
increase / decrease in the time between actual business activity and the
delivery of information to decision-makers
• Number of end-users receiving access to BI applications.
Percent increase / decrease in the number of business users who
receive access to BI applications, but did not in the past 12 months
• End-user productivity and efficiency. Percent improvement in time
spent looking for business information
• Data management infrastructure costs. Percent increase /
decrease in cost of data management infrastructure (hardware and
software) to support BI applications

Page 8

In combination, these metrics provide a means by which respondents can be
benchmarked based on their ability to achieve improved business
performance (Table 1).

Table 1: Top Performers Earn Best-in-Class Status
Definition of
Mean Class Performance
Maturity Class
Achieved 8% mean average decrease in time-to-
information in the past 12 months
Achieved 10% mean average increase in number of
end-users receiving access to BI applications in past
Best-in-Class: 12 months
Top 20% of aggregate
performance scorers Achieved 12% mean average improvement in
employee productivity
Experienced 7% mean average decrease in data
management infrastructure costs associated with
supporting BI applications
Achieved average 2% mean average decrease in
time-to-information in the past 12 months
Industry Average: end-users receiving access to BI applications in past
Middle 50% 12 months
of aggregate Achieved 3% mean average improvement in
performance scorers employee productivity
Experienced 2% mean average decrease in data
Achieved average 1% mean average increase in
time-to-information in the past 12 months
Laggard: end-users receiving access to BI applications in past
Bottom 30% 12 months
of aggregate Suffered 7% mean average decrease in employee
performance scorers productivity
Experienced 6% mean average increase in data

The Best-in-Class PACE Model
Using data management approaches to reduce time-to-information for more
users requires a combination of strategic actions, organizational capabilities,
and enabling technologies that can be summarized in Aberdeen Group's
PACE framework (Table 2).


Page 9

Table 2: The Best-in-Class PACE Framework
Pressures Actions Capabilities Enablers
Reduce time-to- Understand end- Integrate data from Data warehouse software
information for user requirements multiple sources BI query and reporting tools
end-users for BI Measure growth of Systems integration tools
Define the business existing database and data Data warehouse appliance
rules and warehouse assets
Data integration tools
calculations Automate the cleansing of
required for reports Data mining algorithms and tools
data
and analysis views End-to-end BI software solution (ETL, cube /
Committee or group for
Create data model, interface for reporting, dashboards,
obtaining end-user and
management etc.)
corporate BI
strategy roadmap requirements Data cleansing tools
Method for prioritizing IT / systems integrator consulting services
high-demand data for end- Training services
user access Management consulting services
Stand-alone ETL tools

Best-in-Class Strategies
In addition to the desire to reduce the time-to-information for end-users of
"I spend too much time
business intelligence applications, one fourth of all respondents also worrying about one-off
identified the need to deliver BI capabilities to more people, simplify the projects that have started
access to data, and improve the ease-of-use of BI applications. within departmental pockets of
the company, and not enough
Best-in-Class companies' use of input from the end-user community is the
time understanding the
top strategy for developing a successful data management program for requirements of the enterprise
supporting business intelligence initiatives. Yet, 44% of all respondents as a whole. This has caused
reported that end users are not doing a good job of expressing their several false starts that waste
information needs. valuable resources and saps the
life out of the IT teams tasked
Conversely, Best-in-Class companies are not focusing as heavily on building with supporting several
data management strategy roadmaps or identifying data sources for BI operational groups."
applications. This is a result of the fact, as revealed in one-on-one
conversations, that Best-in-Class companies have already addressed these ~ Senior System Architect,
challenges, and the realization among this group that without a clear Large Travel Services Company
understanding of end-user information requirements, work on a data
management roadmap or identification of appropriate data sources cannot
begin (Figure 3).


Page 10

Figure 3: Top Five Strategic Actions
0% 20% 40% 60%

57%
Understand end-user requirements 49%
for BI
46%

Define the business rules and 41%
calculations required for reports and 32%
analysis views 39%

30%
Create data management strategy 44%
roadmap
44%

27%
Identify data sources for BI
33%
applications
31%

Establish data access performance 27%
measurement methodology to 22% Best-in-Class
manage growth 17% Average
Laggard

Aberdeen Insights — Strategy

Businesses are feeling the pressure to compress information delivery timeframes to real or near-real-time
(See Aberdeen Group's December, 2007 study Operational BI: Getting Real-Time About Performance). As the
pressure to deliver more relevant, actionable information in less time increases, BI project budgets and
timelines can increase beyond initial expectations. The effects of data management strategies on the
success of BI initiatives are felt most powerfully when it comes to measuring the ROI of BI
implementations. While 62% of Best-in-Class organizations have achieved at least a 100% ROI on their BI
projects, only 43% of Industry Average and 33% of Laggards have found the same success. As companies
struggle with the cleansing and integration of complex and high-volume data sets, the ability to achieve a
100% ROI on BI investments becomes more difficult, and may erode management buy-in and support as
time frames increase well beyond expectations (Figure 4).
continued


Page 11

Aberdeen Insights — Strategy

Figure 4: ROI Time-Frames for Business Intelligence Projects

50%
43%

40%
34%
31% 31%
30%

20% 15%
13%

8% 8% 8%
10% 6% 6%

0%
0%

Within 1st Month 2 - 6 Months 7 - 12 months 1 - 2 years > 2 years Do not know

Best-in-Class All Others


In the next chapter, we will see what the top performers are doing to
improve timeframes for information delivery and achievement of ROI.


Page 12

Chapter Two:
Benchmarking Requirements for Success
The selection of a data management strategy, methodology, and technology Fast Facts
approach can dictate the level of success that can be achieved in meeting √ 92% of all respondents
end-user requirements for information. indicate that they require
business information to be
Case Study — Advantage Sales and Marketing Shares Their delivered within a day of
Experience with a "BI Appliance" actual business activity - yet,
only 64% of respondents
As North America’s largest CPG sales and marketing broker, Advantage currently receive
Sales and Marketing acts as a liaison between manufacturers and retailers information this quickly
across all 50 states and Canada. They offer a full range of value add
√ 100% of Best-in-Class
services to CPG companies including category management and retail
companies have achieved at
merchandising, all delivered as turn key solutions to customers. least 20% ROI on their BI
The company’s product is represented by its people (over 22,000 strong initiatives compared to only
and growing) and the challenge that Advantage Sales and Marketing had 65% of all others
was a growing need to make strategic decisions based on current, factual
data.
“We were making broad strategic plans based on gut feel and without
the level of data needed to substantiate the decisions we were making,”
states John Thamm, Advantage’s National HR Technology and
Communications Director. “We had a lot of ad hoc reporting capability,
but the data that could be accessed by these systems was limited. We
realized that we needed to learn about new solutions, and see if there
was something that could meet our requirement of near real-time
visibility into the business.”
The company needed to gain visibility into its existing data in order to
improve human resource hiring and development plans. This required a
solution to some specific needs:
"This was not the cheapest
• Detailed analysis of headcount and human resource allocation option, but the value for money
• Understanding of turnover rates and their affect on the business is there. One thing I would
recommend to anyone thinking
• Visibility into employee termination reasons / causes about this is to plan for one full
time employee to be dedicated
Thamm and his team investigated several vendors and tools, and
to this project. We did not
narrowed it down to two vendors with similar solutions. They ultimately commit this resource up front,
decided to work with a BI appliance vendor that offered a bundled and had we done it, we would
hardware and software solution that was specialized for meeting their have saved quite a bit of time.”
unique requirements. “The solution we chose offered what appeared to
be the most seamless implementation and deployment for our ~ John Thamm, National HR
Technology and
organization,” said Thamm. “Our journey was not without challenges, but
Communications Director;
overall we are very pleased with results to date and expect to continue
Advantage Sales and Marketing
our development and roll out of this technology.”
continued


Page 13

Case Study — Advantage Sales and Marketing Shares Their
Experience with a "BI Appliance"

Some of the challenges that were experienced along the way included:
• Proper assignment of IT resources and skill sets. The process
of working within SQL to build BI cubes was complex and not within
the skill sets of the IT administrator originally assigned. This
introduced a consulting services requirement from the vendor which
was provided and helped alleviate the skill set requirement.
• The implementation timeframe was longer than expected.
Due to the need for consulting services, and limited internal IT
resources, the implementation went from an expected 60 days to a
year. The solution is not yet fully deployed, but it is implemented
locally. The company is limiting exposure to a larger user base until
wrinkles are ironed out.
Once the first full deployment is completed, management will develop
metrics for measuring ROI, with a goal of achieving 100% ROI within 18
to 24 months.
“The first set of applications and dashboards were developed for the HR
management group as a beta project. This includes about eight people
who are shaking out the wrinkles and engaging in very open dialog about
changes and improvements,” continues Thamm. “Once that is completed
we will deliver an expanded set of deliverables to the executive team,
and then to a larger audience of over 150 business leaders around the
country. We have had a great experience working with the solution
provider. We know that challenges should be expected, and their
approach to working with us to solve them was top notch. This was not
the cheapest option, but the value for the money is there. One thing I
would recommend to anyone thinking about this is to plan for one full
time employee to be dedicated to this project. We did not commit this
resource up front, and had we done it, we would have saved quite a bit
of time.”

Competitive Assessment
Aberdeen Group analyzed the aggregated metrics of surveyed companies to
determine whether their performance ranked as Best-in-Class, Industry
Average, or Laggard. In addition to having common performance levels, each
class also shared characteristics in five key categories: (1) process (the
approaches they take to execute their daily operations); (2) organization
(corporate focus and collaboration among stakeholders); (3) knowledge
management (contextualizing data and exposing it to key stakeholders);
(4) technology (the selection of appropriate tools and effective
deployment of those tools); and (5) performance management (the
ability of the organization to measure their results to improve their
business). These characteristics (identified in Table 3) serve as a guideline


Page 14

for best practices, and correlate directly with Best-in-Class performance
across the key metrics.

Table 3: The Competitive Framework
Best-in-Class Average Laggards
Process Ability to optimize queries
Management 40% 33% 15%
Committee or group for obtaining end-user and corporate
BI requirements
Organizational 53% 52% 21%
Management
Training assets for deploying BI to end-users
37% 27% 9%
Measure growth of existing database and data warehouse
Performance assets
Management
57% 50% 16%
Method for prioritizing high-demand data for end-user
access
Data /
Knowledge 51% 39% 12%
Management Formalized end-user polling or survey
44% 29% 15%
Integrate data from multiple sources
Technology 77% 54% 22%
Management Data cleansing capability
56% 51% 35%

Capabilities and Enablers "Overall, the biggest gain from
Based on the findings of the Competitive Framework and one-on-one our BI efforts has been
interviews with end users, Aberdeen’s analysis of the Best-in-Class speeding up our time-to-action.
demonstrates that no one specific capability or technology investment This has been particularly
provides a silver-bullet solution. Rather, it is a combination of capabilities evident with our loss
prevention initiatives. We can
and prioritizations of technology investments that leads to improved time-
now see discount anomalies
to-information, simplification of data access, and inevitably ease of use of BI faster, and can stop undesirable
applications by end-users. activity before it impacts the
performance of a store or the
Process company. The data is out there
and utilizing it to react more
Reducing the time-to-information for end-users necessitates an
quickly is the ultimate source of
understanding of the information queries that place a burden on the data benefit.”
management system. Best-in-Class companies have identified this as the top
process management capability needed for performance improvement ~ Retail Business Consultant,
(Figure 5). Global Footwear / Apparel
Company


Page 15

Figure 5: Ability to Optimize Data Queries
50%

40%
40%
33%
30%

20%
15%

10%

0%
Best-in-Class Average Laggard

Query optimization entails a two-phase process: Analyst Insight
• The ability to monitor and assess all queries in aggregate, and determine Depending on how the data is
which types of queries place the most strain on the data management stored and accessed, a query
system that is structured one way may
drastically outperform a query
• The ability to alter data structures and query assemblage in order to that is structured differently.
make the most optimal use of the data management resource For example, imagine the
Optimization is achieved when the system's ability to handle query load and difference between searching
deliver desired information falls within the timeframes required by end- for a phone number by area
users. Often, end-user queries are presented in non-optimal ways. One user code, then prefix, and then last
four digits versus searching by
may start by segmenting data by product type, while others may begin by
last name, then first name. Both
filtering data on region or customer type. Query optimization, therefore, methods will eventually yield
involves an understanding of the data structure, but just as importantly, an the phone numbers you are
understanding of the end-user needs and query assembly is also required. seeking, but one will take much
Optimization is achieved when the two are brought together, and end-users less time. The same can be said
are either trained on optimal query assembly, or technology is used to of data queries that are
automate the process of re-assembling poorly constructed queries. structured improperly. Many
organizations interviewed for
this study stated that they have
Organization sought external assistance to
Best-in-Class companies are far more likely to take an enterprise approach solve their query structure
to organizational management of their business intelligence initiatives. When challenges.
it comes to data management for BI, over half of Best-in-Class and Industry
Average companies are currently prioritizing the creation of a formal
corporate committee for obtaining end-user BI requirements.
Unfortunately, the same cannot be said of Laggards (Figure 6).


Page 16

Figure 6: Top Organizational Capabilities
60%
53% 52%
50%

40% 37%

30% 27%
21%
20%
9%
10%

0%

Committee or group for obtaining end-user and corporate BI requirements
Develop training assets for deploying BI to end-users


This aligns with the findings in Chapter One regarding the strategic
objectives that Best-in-Class companies are emphasizing around end-user
requirements above technical or system-oriented strategies. In addition to
obtaining end-user requirements, Best-in-Class companies are also far more
likely to have implemented training programs around BI deployment to end-
users. Without a thorough knowledge of how to use BI applications, much
of the effort that goes into the construction and maintenance of data
warehouses and operational databases can be wasted. Effective data
management requires that both the back-end database and data warehouse
and the front-end user interface and query engines work together
cohesively. Without a formal training program, the front-end side of the
equation can be thrown into jeopardy.

Knowledge Management
In addition to query optimization, identification of high-demand data is
critical to the successful BI implementations. As BI usage grows within an
organization, it is typical that specific data sets are in higher demand at
particular times during the business day. For example, each morning, the
previous day's sales results and analysis might be of highest importance to
customer-facing sales and marketing staff. At the end of a fiscal business
period (week, month, quarter) financial reports and analysis will be in high
demand among financial / accounting staff and executives. The ability to tune
access to these data sets and provide prioritization for queries related to
them is therefore a Best-in-Class capability (Figure 7).


Page 17

Figure 7: Knowledge Management - Data Prioritization and User
Demand Intelligence
60%
51%
50% 44%
39%
40%
29%
30%

20% 15%
12%
10%

0%

Method for prioritizing high-demand data for end-user access
Formalized end-user polling or survey


The ability to identify and fine-tune systems for high-demand data requests
stems from, once again, gaining an accurate understanding of end-user
requirements. Best-in-Class companies are significantly more likely to
employ end-user polling or survey methods to gain this understanding.
While high-demand data can also be identified through the study of
transaction logs, and certainly should be, the inclusion of end-user polling
adds a dimension that cannot be found in transaction logs - predictability.
Logs only reveal what has happened in the past, and while this is a good
indicator of what might happen in the future, only the human end-users can
tell you this based on their knowledge of information needs that are on the
horizon.

Performance Management
Awareness of data volume growth and the implications of this on system
performance is an important component to successful data management for
business intelligence initiatives. Companies are beginning to wake-up to the
realization that there are untapped data sources in the form of internal
documents, emails, customer service field notes, as well as external data
such as web pages, wikis, blogs, and a plethora of government and
association databases. As integration of these new data sources commences
and grows, new stresses will be placed on data management initiatives by
data volumes that will be and are increasing exponentially. Best-in-Class
companies are realizing this and are more likely to have the ability to
measure growth of existing database and data warehouse assets (Figure 8).


Page 18

Figure 8: Ability to Measure and Manage Growth of Database and
Data Warehouse Assets
60% 57%
50%
50%

40%

30%

20% 16%

10%

0%
Technology Management
As the data volumes and the number of data sources continues to increase,
the ability to integrate new sources with existing data is becoming more
critical to BI implementation success. The speed at which this integration
occurs may be the lynchpin that determines whether or not end-user time-
to-information requirements are met.
In addition, the level of trust that users have in the data dictates the level to
which user uptake and adoption will occur. Integration of data from multiple
sources not only needs to happen quickly, but it also must result in accurate
information. The support of a BI deployment that is tasked with delivering
actionable information can only survive if end-users believe the data they are
consuming. Therefore it is extremely important to incorporate both data
integration and cleansing capabilities into a data management strategy (Figure 9).

Figure 9: Best-in-Class Data Integration and Data Cleansing
Capabilities
100%

77%
75%
56% 54% 51%
50%
35%
22%
25%

0%

Integrate data from multiple sources Data cleansing



Page 19

Enabling Technologies and Services
Best-in-Class companies have made a significantly higher investment in
technology than all other respondents. More than half of Best-in-Class
respondents report that they are making use of all but one of the technology
enablers shown in figure 10 below. This further illustrates the fact that data
management for BI encompasses several capabilities and technical
requirements, and should not be viewed as just one function within the IT
organization (Figure 10).

Figure 10: Best-in-Class versus All Others - Technology Enablers
Currently Deployed
0% 30% 60% 90%

82%
Data warehouse software
56%

80%
BI query and reporting tools
47%

75%
Systems integration tools
45%

68%
Data warehouse appliance
34%

66%
Data integration tools 42%

62%
Data mining algorithms and tools 30%

End-to-end BI Software solution
58%
(ETL, Cube/model, interface for 28%
reporting, dashboards, etc)
57%
Data cleansing tools 35%
Best-in-Class
39%
Stand-alone ETL tools 26% All Others


Data management strategies may also require external assistance in order
to ensure successful implementation and deployment. Forty-eight percent
(48%) of all respondents list "the lack of IT resources" as the top inhibitor
preventing successful BI implementations. For companies that fall into this
category, a lack of IT resources can often be remedied through the
establishment of partner relationships with IT services firms or consulting
divisions of software technology vendors (Figure 11).


Page 20

Figure 11: Best-in-Class versus All Others – IT / Consulting
Services Currently Utilized
0% 10% 20% 30% 40% 50%

IT/Systems
Integrator 49%
consulting 34%
services

44%
Training services
24%

Management 43%
consulting
services 28%
Best-in-Class
All Others


As Best-in-Class companies look forward, their planned investments in
technology are focused on data integration, cleansing, and the incorporation
of these capabilities within and end-to-end business intelligence application
platform (Figure 12).


Page 21

Figure 12: Best-in-Class Technology Enablers Current versus
Planned Deployment
0% 25% 50% 75% 100%

Data warehouse software 82% 10%

BI query and reporting tools 80% 18%

Systems integration tools 75% 23%

Data warehouse appliance 68% 21%

Data integration tools 66% 34%

Data mining algorithms and tools 62% 26%

End-to-end BI Software solution (ETL,
Cube/model, interface for reporting, 58% 35%
dashboards, etc)

Data cleansing tools 57% 34%

Stand-alone ETL tools 39% 37%

Current Planned next 24 months

Aberdeen Insights — Technology

Data management for BI is not a single technology, nor is it confined to
being solely an IT function. Rather, it is an accumulation of organizational,
operational, knowledge, process, performance, and technology
enablement. Best-in-Class companies have taken a multi-faceted approach
to data management in order to support their BI implementations. This
entails a mix of both internal management actions as well as investment
in technologies and services.


Page 22

Case Study — Sendio Improves Competitive Positioning With
“Competitive positioning was
Innovative Approach to Data Management another business pressure
driving our search for a better
Sendio is an Irvine, CA based company that provides “email integrity” data management solution.”
technology products and services to customers who want to eliminate
viruses, "trojans," "bots," and spam while not risking the inadvertent loss ~ Tim Lee-Thorp, Vice
of legitimate email messages. As part of the process to guarantee that President of Marketing, Sendio,
Inc.
customers receive all of the email they want and none of the junk, Sendio
required a bullet-proof data aggregation and analysis tool that they could
provide to customers in a packaged implementation which could be
quickly and easily deployed.
In just four years, the company has grown its customer ranks to over 800
companies. While the increase in customer activity was a welcome boon
to Sendio’s growth, the demands that customers started placing on
Sendio for more sophisticated reporting and analysis of their email
activity was causing significant business pressure.
“Competitive positioning was another business pressure driving our
search for a better data management solution. Our sales team was
identifying a threat from competitors who were delivering analytic
capability with their email anti-spam solutions, and management came
under pressure to leap-frog the competition,” said Tim Lee-Thorp,
Sendio’s vice president of Marketing. “While reporting capabilities are a
‘given,’ the sophistication of analytical capabilities represents a real
differentiator.”
Sendio’s management team assessed building a centralized data collection,
analysis, reporting, and alerting facility with a combination of traditional
relational database and data warehousing software, but the resource
requirements for this approach using either in-house developers or third-
party contractors was significant. The option to simply provide local
reporting for individual customers was also rejected, since this also did
not allow for real-time transaction analysis across the entire Sendio user
community, something customers wanted to have in order to identify
emerging threats as they occurred and actively address them before they
could harm the organization. Local reporting is also CPU-intensive,
which could impact the basic email processing workflow.
continued


Page 23

Case Study — Sendio Improves Competitive Positioning With
Innovative Approach to Data Management

Instead, Sendio chose to invest in a data warehouse appliance approach
that is deployed in a “split-architecture” environment. While the email
processing / transactional system runs locally to insure speed and
performance within the customer’s email server environment, a data
warehouse appliance is installed and hosted remotely by Sendio to
maintain and deliver real-time data for customer ad-hoc and custom
analytical reporting and analysis purposes. As the system usage scales
upward (data volumes and number of user sessions), the appliance
approach enables Sendio to simply add additional server capacity
centrally.
“Our core email integrity services are delivered to customers via a
distributed appliance architecture, so we were very comfortable with the
concept of a data warehousing appliance. Simple deployment, rapid
applications development, and predictable scaling lets us easily deliver
aggregate reporting and global trending in a ‘virtual data warehouse’
model that customers seem to like. They can now see their email
performance dynamics versus 800+ other companies. This provides a
higher level of service and valuable peer analysis that customers can use
to benchmark their performance and see progress over time.”


Page 24

Chapter Three:
Required Actions
Whether a company is trying to move its data management and related BI Fast Facts
performance from Laggard to Industry Average, or Industry Average to √ 68% of Best-in-Class
Best-in-Class, the following actions will help spur the necessary performance companies have
improvements: implemented their data
warehouse in less than one
year, compared to 45% of
Laggard Steps to Success Industry Average and 19% of
• Focus on understanding end-user requirements. Best-in-Class Laggard companies
companies have shown that a successful data management strategy √ Respondents are managing
starts with an understanding of end-user information requirements. Too an average of 10.2 unique
often, companies launch data management projects to support the data sources to support
technical requirements of a BI implementation without spending the current BI initiatives
necessary time up-front with a representative group of business users.
Taking this step can prevent wasted effort later when the potential need
for multiple iterations of data constructs and modeling arise due to lack
of BI deliverable alignment with business requirements.
• Establish implementation performance metrics for data
management projects. In order to reduce the time-to-information
for end-users, data management projects must also be completed in a "Our business end-users are all
compressed timeframe. This requires that initiatives include both non-technical people not used
business and project metrics that can be tracked and measured from the to writing queries and reports.
outset. By adopting this method as a standard practice, a deeper We needed a combination of
understanding of project roadblocks and inhibitors will be gained, data access that was simple
allowing decision makers to be more aware of potential problems and enough for non-technical
proper solutions before project budgets and timelines are affected. people to use, and an internal
training program focused on
• Invest in data integration enablement. Best-in-Class companies how to create information
are 30% more likely than all respondents, and nearly three-times as views, reports, and analysis
likely as Laggard companies to adopt and utilize data integration tools to applications. We have found
enable automation and acceleration of data access for BI applications. that a half-day workshop with a
Respondents are dealing with an average of over 10 unique data sources hands-on tutorial containing
real company data is the key to
within their data management systems underlying BI implementations.
success."
The degree to which each of these sources differs from the others adds
a level of complexity and therefore potential added time to data ~ Director of Financial Systems,
management projects. Large Transportation and
Logistics Company

Industry Average Steps to Success
• Institute a formal process for polling end-users about
information requirements. To move from Industry Average to Best-
in-Class status, companies should consider institutionalizing a regular
method for collecting and updating end-user information requirements.
While it is important to conduct this activity at the start of a project, it
is just as critical that polling of end-users' needs becomes a regular
activity. The dynamics of today's business climate necessitates that


Page 25

frequent review of information requirements be incorporated into any
data management strategy.
• Develop training assets for non-technical business end-users.
While the top pressure driving data management projects relates to
compression of time, the next two business drivers deal with ease-of-
use and simplification of data access. Best-in-Class companies have
illustrated that investment in data management projects that support BI
deployments must encompass the education of users. Otherwise, the
initial investments in back-end technology and services may go to waste
as users find the level of difficulty a barrier to adoption.
• Investigate methods for optimizing data queries. Business
intelligence reporting, ad-hoc query, and automated analytic applications
can place demands on a data management system that outstrip its ability
to respond to all users in a timely manner. In order to understand the
system-level strains, a method for gaining visibility to query performance
is necessary. Discussions with respondents revealed that often, there
are only a handful of queries that cause problems, but it only takes one
"runaway query" to bring down a system and cause information delivery
delays enterprise-wide. The ability to identify and correct these
anomalies, and even determine a new format or structure for the
problem query can provide immediate time-to-information performance
improvement.

Best-in-Class Steps to Success
• Identify and prioritize data sources for BI. Conversations with
respondents have revealed that there is often a temptation to try to
incorporate all data sources into one centralized data management
system for the support of multiple BI deployments and users. While
technology enablers may make this possible, it might not be the best
approach to take. Consider a prioritization process for determining
which data sources are necessary to answer which business questions
and information requirements. This may result in the creation of
multiple data "marts" or "domain warehouses" that address specific
needs with a more efficient use of resources and resulting in greater
speed of delivery.
• Plan for future growth of data volumes, complexity, and
addressable end-users. There is no question that data volumes are
growing, business requirements are becoming more sophisticated, and
the number of addressable end users is increasing. To deal with this,
companies must establish a planning process that incorporates several of
the capabilities outlined earlier. An integrated approach to continually
updating intelligence about end-user requirements, combined with the
identification and prioritization of data sources, and powered by the
automation of data integration and cleansing will lead to increased ability
to deal with the non-stop growth of data management requirements for
business intelligence initiatives.


Page 26

Aberdeen Insights — Summary

Data warehousing has historically been the leading method for supporting
BI implementations, but within the next 12 to 24 months, a shift is
occurring. Integrated technological approaches are gaining the attention
of performance-minded Best-in-Class companies that are seeking to
further improve their ability to compress the time-to-information for
their business end-users.
Success of business intelligence implementations depends upon the ease
and speed by which relevant and actionable information can be delivered
to business users. Best-in-Class performance toward this goal is
determined by several factors that must be taken into account before,
during, and on an ongoing basis with data management projects. This
starts with gaining a deep understanding of business user information
requirements before a data management roadmap or strategy can be
developed and executed. Several data management capabilities and
technology enablers are required to ensure success, and no one "silver
bullet" exists that can provide a shortcut. Best-in-Class companies have
illustrated that performance improvement comes from a holistic
approach that includes both organizational and technological elements.


Page 27

Appendix A:
Research Methodology
Between January and February 2008, Aberdeen examined the use, the Study Focus
experiences, and the intentions of 247 enterprises using data management
technologies and methods for the support of Business Intelligence (BI) Responding executives
completed an online survey
initiatives.
that included questions
Aberdeen supplemented this online survey effort with interviews with select designed to determine the
survey respondents, gathering additional information about strategies, following:
experiences, and results. • The degree to which
Responding enterprises included the following: companies are defining a
data management strategy
• Job title / function: The research sample included respondents with the for BI initiatives
following job functions: procurement, supply chain, or logistics manager • The level of capability
(15%); operations manager (13%); IT manager or staff (37%); sales, respondents have to
customer service and marketing staff (19%); finance (4%); and consulting provide access to relevant
/ other (13%). information for all types of
end users
• Industry: The research sample included respondents from several
industries: manufacturing (17%); heavy industry / engineering (14%); high • The extent to which
tech / software (13%); retail / wholesale / CPG (13%); finance / banking / companies can accelerate
the “time-to-information”
insurance (9%); public sector / education (9%); supply chain / logistics
through a variety of data
(8%); and telecommunications (6%) were the largest responding sectors. management approaches
• Geography: The majority of respondents (55%) were from North • Respondent performance
America. Remaining respondents were from Europe (22%), the Asia- toward decreasing the cost
Pacific region (13%) and Rest-of-world (10%). and resource requirements
associated with supporting
• Company size: Thirty-four percent (34%) of respondents were from
data access for BI users
small enterprises (annual revenues below US $50 million); 38% were
from midsize enterprises (annual revenues between $50 million and $1 • The degree to which end-
billion); and 28% of respondents were from large businesses (annual users obtain information
revenues of greater than $1 billion). autonomously without
requiring IT intervention
• Headcount: Twenty-two percent (22%) of respondents were from small The study aimed to identify
enterprises (headcount between 1 and 100 employees); 34% were from emerging best practices for
midsize enterprises (headcount between 101 and 1,000 employees); and data management within BI
44% of respondents were from large businesses (headcount greater implementation environments,
than 1,000 employees). and to provide a framework by
which readers could assess
Solution providers recognized as sponsors were solicited after the fact and
their own management
had no substantive influence on the direction of this report. Their capabilities.
sponsorship has made it possible for Aberdeen Group to make these
findings available to readers at no charge.


Page 28

Table 4: The PACE Framework Key
Overview
Aberdeen applies a methodology to benchmark research that evaluates the business pressures, actions, capabilities,
and enablers (PACE) that indicate corporate behavior in specific business processes. These terms are defined as
follows:
Pressures — external forces that impact an organization’s market position, competitiveness, or business
operations (e.g., economic, political and regulatory, technology, changing customer preferences, competitive)
Actions — the strategic approaches that an organization takes in response to industry pressures (e.g., align the
corporate business model to leverage industry opportunities, such as product / service strategy, target markets,
financial strategy, go-to-market, and sales strategy)
Capabilities — the business process competencies required to execute corporate strategy (e.g., skilled people,
brand, market positioning, viable products / services, ecosystem partners, financing)
Enablers — the key functionality of technology solutions required to support the organization’s enabling business
practices (e.g., development platform, applications, network connectivity, user interface, training and support,
partner interfaces, data cleansing, and management)

Table 5: The Competitive Framework Key
Overview

The Aberdeen Competitive Framework defines enterprises In the following categories:
as falling into one of the following three levels of practices Process — What is the scope of process
and performance: standardization? What is the efficiency and
Best-in-Class (20%) — Practices that are the best effectiveness of this process?
currently being employed and are significantly superior to Organization — How is your company currently
the Industry Average, and result in the top industry organized to manage and optimize this particular
performance. process?
Industry Average (50%) — Practices that represent the Knowledge — What visibility do you have into key
average or norm, and result in average industry data and intelligence required to manage this process?
performance. Technology — What level of automation have you
Laggards (30%) — Practices that are significantly behind used to support this process? How is this automation
the average of the industry, and result in below average integrated and aligned?
performance. Performance — What do you measure? How
frequently? What’s your actual performance?


Table 6: The Relationship Between PACE and the Competitive Framework
PACE and the Competitive Framework – How They Interact
Aberdeen research indicates that companies that identify the most influential pressures and take the most
transformational and effective actions are most likely to achieve superior performance. The level of competitive
performance that a company achieves is strongly determined by the PACE choices that they make and how well they
execute those decisions.


Page 29

Appendix B:
Related Aberdeen Research
Related Aberdeen research that forms a companion or reference to this
report include:
• “On-Demand” Is Not Far Behind BI on the Technology Wish List June, 2007
• Data Management 2.0: Making Sense of Unstructured Data July 2007
• Delivering Actionable Information to the Enterprise: Does On-Demand Solve
the Skill Set Shortage? July 2007
• On-Demand BI: Not Just for SMB August 2007
• Serving the Underserved: Is On-Demand BI the Answer? August 2007
• Enterprise BI: Comparing the BI Giants September 2007
• Smart Decisions: The Role of Key Performance Indicators September, 2007
• Measuring Marketing Performance: The BI Roadmap to Information Nirvana
October 2007
• Operational BI: Getting Real-Time About Performance December 2007
• The Expansion and Contraction of Business Intelligence January 2008
• Managing the TCO of Business Intelligence February 2008
Information on these and any other Aberdeen publications can be found at
www.Aberdeen.com.

Author: David Hatch, Research Director, Business Intelligence,
david.hatch@aberdeen.com
Since 1988, Aberdeen's research has been helping corporations worldwide become Best-in-Class. Having
benchmarked the performance of more than 644,000 companies, Aberdeen is uniquely positioned to provide
organizations with the facts that matter — the facts that enable companies to get ahead and drive results. That's why
our research is relied on by more than 2.2 million readers in over 40 countries, 90% of the Fortune 1,000, and 93% of
the Technology 500.

As a Harte-Hanks Company, Aberdeen plays a key role of putting content in context for the global direct and targeted
marketing company. Aberdeen's analytical and independent view of the "customer optimization" process of Harte-
Hanks (Information – Opportunity – Insight – Engagement – Interaction) extends the client value and accentuates the
strategic role Harte-Hanks brings to the market. For additional information, visit Aberdeen http://www.aberdeen.com
or call (617) 723-7890, or to learn more about Harte-Hanks, call (800) 456-9748 or go to http://www.harte-hanks.com

This document is the result of primary research performed by Aberdeen Group. Aberdeen Group's methodologies
provide for objective fact-based research and represent the best analysis available at the time of publication. Unless
otherwise noted, the entire contents of this publication are copyrighted by Aberdeen Group, Inc. and may not be
reproduced, distributed, archived, or transmitted in any form or by any means without prior written consent by
Aberdeen Group, Inc. 010908a


Data Management for Business Intelligence

More Related Content

What's hot (20)

Similar to Data Management for Business Intelligence (20)

More from FindWhitePapers (20)

Recently uploaded (20)

Data Management for Business Intelligence