MMH Data Warehousing, Corporate Portal & e-business Applications.

Data Warehousing, Corporate Portal & e-Business Intelligence Applications

  

Mimno, Myers & Holum

"Is Your Data Warehouse Business-Driven?"

FlashPoint Column - November 11, 2001
"Is Your Data Warehouse Business-Driven?"


The first step in every development methodology is to identify the business purpose of the application. In data warehousing, the recommendation is to survey business managers, identify the most painful business problems, and use the data warehouse to resolve these problems.
However, in discussions with organizations that are embarking on the development of a data warehouse, I have found that many IS organizations are technology-driven, not business-driven, in planning for their data warehouse. The most obvious symptom of technology-driven planning is to select a decision-support tool, such as BusinessObjects, Cognos PowerPlay, or MicroStrategy first, and then select other components of the data warehousing architecture. Data warehousing components, such as DSS tools, ETL tools, data modeling tools, data cleansing tools, etc., can be bought off-the-shelf and used by IS personnel to construct a data warehousing application, without talking to business users. The resulting data warehouse is delivered to end users, but the IS department may be dismayed to find that the end users don't come. The end users do not understand the purpose of the data warehouse and they were not consulted adequately in the planning of the data warehousing application.
A business-driven approach is the exact opposite of a technology-driven approach to planning a data warehouse. The recommended strategy is to define the business problem first, then select an ETL tool, and finally select a DSS tool that integrates at the metadata level with the ETL tool that has been selected. Which approach are you using - are you business-driven or technology-driven?
The first step in a business-driven methodology for data warehousing is to survey your top-level business managers to identify the most painful and specific problems the organization is facing. Typical business problems that I have encountered in my consulting practice include the following:

  • Inability to share information across the organization
  • Multiple, inconsistent sources of data
  • Lack of ability to generate reconcilable financial reports
  • Inefficient, paper-driven processes
  • Analysts must contact multiple business units for data
  • Lack of effective utilization of Internet access
  • Proliferation of non-integrated, stovepipe applications
One client, a pharmaceutical company, was struggling with a severe business problem: it took 21 days at the end of each quarter to produce a consolidated financial report across all divisions of the business. Each division produced reports that were meaningful to the local business units, but could not be reconciled across business units due to a lack of consistent entity definitions and business rules. Another client could not provide accurate predictions of sales revenue and profit for the next quarter. At the end of each quarter, their financial projections for the quarter were in error by 20 to 30%, which caused their stock price to be severely beaten down. Some clients find it difficult or impossible to share data across divisions of their organization. Other clients have numerous old, stovepipe applications that do not meet current business requirements and produce source data that cannot be reconciled across applications.
These problems can be extremely painful to the organization and may be at the top of the list of problems that must be solved. As we know, data warehousing applications are ideally suited to solving these problems. ETL tools can be used to access data from numerous disparate data sources, resolve the inconsistencies in source data, and produce a single, clean target database that can be used for decision support. The metadata generated by an ETL tool can be used to synchronize central metadata definitions with local metadata definitions. Using a metadata bridge, supported by both the ETL tool and the DSS tool, reports generated by local business units can be reconciled automatically with reports produced by corporate managers. With this architecture, the time required to produce reconciled, consolidated, quarterly financial reports can be reduced from weeks to minutes. The same mechanism can be used to share clean, consistent data across all divisions of an organization.
The recommended approach to strategic business analysis is to ask high-level business managers a series of questions that lead to a definition of the strategic objectives of the data warehouse. These questions may include:
  • What are the strategic success factors of the organization?
  • What is the Enterprise Business-Intelligence strategy for the organization?
  • What information does the organization require to analyze its current and future business problems to gain a competitive advantage?
  • What decision-support processes are in use at the organization and what are the deficiencies of these processes?
  • Is there a clean, consistent source of data that can be used for decision-making?
  • Is there a single, consistent source of business entity definitions and business rules?
  • Do individual business units define business entities, such as profit, sales, depletion, etc., differently from other business units? How are these differences currently resolved to produce consolidated financial reports?
  • What problems does the organization have in sharing data across divisions? Are the owners of source data willing to share their data transparently across the organization?
  • Are there serious data integrity and security issues involved in access of data across the organization?
  • Do the owners of source data proactively clean-up their data to resolve data integrity issues, such as logical inconsistencies, missing data, data range errors, and referential integrity problems?
  • Is it possible to analyze operational problems by drilling down from high-level summary information to reconciled, lower-level, detailed information?
  • Are many business processes paper-driven?
  • How much effort is currently spent by business analysts resolving data integrity issues?
  • How much effort is currently spent by IS analysts in producing custom data extracts and reports?
All of these issues can be addressed through the use of data warehousing solutions. The objective of the strategic business analysis phase is to define a short list of business drivers for the data warehouse based on interviews with top-level managers.
In my experience, it is important to select painful, high-visibility business drivers for the data warehouse, not trivial ones. Stick your neck out and tackle one of the most painful business problems in the organization. If you succeed, you may be invited to make a presentation to the Board of Directors of the organization to describe how you solved a top-level concern of the Directors. This actually happened to one of my clients. Conversely, if you fail, in spite of using the best data warehousing architecture, components, and practices, you will have learned valuable lessons. You are likely to be given a second chance to use your experience to solve the strategically important issue.
I have found that it is important to define the business purpose of the data warehouse in a single sentence, which is often called an "elevator test". Let's say that you have just been appointed manager of a data warehousing project. You get on an elevator and the head of your division enters the elevator, punches the button for the floor above you, and says "Congratulations on being appointed manager of the data warehousing project. What is the primary benefit of the project to the division?" You have one floor to answer the question.
Try to quantify the answer; don't just say that the objective is to provide better quality information for decision-making. Examples of quantified one-liners include:
  • Reduce headcount by 6% and increase workforce productivity by 20% in 12 months
  • Reduce the cost of pharmaceuticals across a group of hospitals by 50% in 18 months
  • 20% of insurance claims take 80% of the effort to resolve claims. Reduce the average claim effort and equalize the effort across all claims
  • Decrease the time to process small procurements from 16 hours to 4 hours via a Web-based process
  • Improve productivity by 30% by leveraging technology to automate workflow and minimize flow of paper
  • Eliminate all hand-coded extraction and transformation procedures through the use of an ETL tool in 18 months
  • Reduce the time spent by business analysts in resolving inconsistencies in source data from 3 ½ days per week to zero in 6 months
  • Reduce the time spent by IS analysts in generating custom reports to zero in 6 months

The one-liner is a concise definition of the business purpose of the data warehouse. It can be used to decide whether a proposed new function should be implemented now or deferred to later development. If the proposed functionality directly supports the one-line definition, then the data warehouse administrator will try to include it in the ongoing development effort. Conversely, if the proposed functionality does not meet the goal of the one-liner, the proposed function will be deferred.
The deliverable of the strategic business analysis phase is a short document that describes the primary sources of pain in the business that will be addressed by the data warehouse.
Note: This series of articles describe steps in a bottom-up methodology that I have found to be successful for the implementation of data warehousing.applications. The primary goals of the methodology are to reduce the up-front effort required to specify the functionality of a data warehousing application and deliver data marts in 90 days or less, at low cost and low development risk. The overall methodology is summarized in a previous FlashPoint article by Pieter Mimno entitled "Fast Payoff for Data Warehouse Investment".
For further information about the issues discussed in this report, please contact Pieter Mimno, Independent Consultant, at pmimno@mimno.com, or visit his Web site at www.mimno.com. Mr. Mimno specializes in the selection of system components and support for all phases of development for data warehousing, corporate portals, and eBusiness-intelligence applications.

Reprinted with permission from The Data Warehousing Institute. 
Copyright 2002. The Data Warehousing Institute.

All rights reserved. 

[ TOP ]

 
 
 

©2001 Mimno, Myers & Holum