Over the last decade, many large information technology (IT) shops have undergone a dizzying array of change including outsourcing, insourcing, organizational realignments, new CIOs and new technologies. This turmoil has left many large IT shops without the fundamentals for providing excellent business intelligence (BI) and data warehousing services to their customers. As if that wasn't enough, many stovepipe warehouses developed during the last 10 years will need to be integrated (thus redesigned and combined) to meet customer demands at reasonable expense over the next 10 years. What's going to insure that IT is far more successful at this than it has been? The answer is an enterprise business intelligence center of excellence (CoE).

Today's smart service vendors in the business intelligence, data warehousing and data mining arena are establishing centers of excellence in an effort to provide businesses what many internal IT groups are not – expert one-stop shopping. A number of big name vendors and many smaller ones are staking a claim to the business intelligence services market through formation of centers of excellence. It could well be that big blue's recent acquisition of PwC Consulting was partly founded in acquiring just such expertise.

How do we know if we are excellent at the basics? Excellence can be defined and measured by applying a framework that clearly identifies the differentiators. At the highest level, businesses that have truly knowledgeable business intelligence and data warehousing people implement technology that works, provides clean usable data and uses proven processes to consistently maintain value (see Figure 1).

Figure 1: Pillars of Success in BI and Data Warehousing

Measuring the degree of excellence is accomplished through use of the capability maturity model (CMM) against each of the pillars of success. Many organizations will find their CMM rating to be one or less in one or more of these pillars, while excellence requires a rating of two to three. A center of excellence plan would then define the time frame and detailed objectives for achieving enterprise-wide excellence (see Figure 2).

Figure 2: CMM Excellence Timetable

The Pillars of Success

In order to clearly understand what deficiencies need to be addressed, we need to appreciate the value that excellence brings to each pillar.

Knowledgeable and Well-Trained People

Knowledgeable people have by far the greatest impact on total cost and degree of success because they also have a cumulative effect on the other three pillars. While there are no exact calculations to draw on, people who are not knowledgeable in BI – either IT, clients or both – can easily add 10 to 50 percent to the total cost of a BI application. In a worst-case scenario, inadequate knowledge can add several hundred percent to the total cost of a delivered application. This happens when a seriously flawed development continues to try to produce results, taking the entire team two to five times longer than what should have been required.

Proven Processes

As in most endeavors, having proven processes to guide work into the most efficient and high-quality paths has clear cost advantages. The building of data warehouses and BI applications needs the discipline such a process provides. The problem may be that the enterprise does not understand that there are significant differences in the steps necessary to design and build a BI application than those required to design and build a transaction or reporting system.

Lacking well-defined and proven "iterative" processes to build and maintain large complex BI applications can be costly. In larger enterprises, weakness in this pillar can add 5 to 20 percent to the overall cost of a project. In the worst case, it can also lead to a project's failure to ever obtain the required results.

Technology That Works

Providing technology that works is not as simple as one might think. This pillar stands on the client's definition of what works and what doesn't, and is often the most difficult to assess for weaknesses. Typical weaknesses here include performance, usability and integration issues. These weaknesses add cost in either correction or lack of functionality. In general, technology woes can add 5 to 25 percent to the total cost. Again, the worst case is that expected benefits are not derived because the technology issues don't get resolved.

Clean Data

Having clean usable data throughout enterprise warehouses is like having a well-organized, accessible and dirt-free garage. Even if you get it cleaned and organized once, it quickly returns to a degree of chaos if not continuously managed. Measuring the effect of unclean or unusable information can be quite difficult, except in rare instances of obvious disconnects. More often, the cost of inconsistent or unreliable information is measured in wasted time using it, an inability to derive value and future requests for new applications with better data.

Weakness in this pillar can also be expensive. For newly launched applications, serious weakness can cause significant rework in the data architecture and ETL components, costing from 5 to 15 percent of the total spent to get it launched the first time. Additional costs will be incurred if data becomes less reliable over time, and these costs will continue to rise until the issues are resolved.

Business Intelligence Excellence

By breaking down the four pillars of success in business intelligence into their critical success factors, we can begin to define principles for excellence that guide the enterprise, as well as identify more concrete CoE objectives.

Value Knowledgeable and Well-Trained People

Building and retaining knowledgeable BI and DW resources needs to become a priority within IT, particularly at the senior-architect level. These architects will need to face off with senior customers and external solution providers, which is not an easy task in any discipline.

Pillar Principles for Excellence:

  • Place best-of-the-best resources into defined CoE roles – supplement with vendors as needed.
  • Have resources balance work on critical issues across the enterprise.
  • Lead key resource/vendor relationships.
  • Build and maintain CoE business management sponsorship.

Pillar Objectives:

  • Educate and involve business clients.
  • Plan to harvest business value quickly.
  • Preach and value enterprise information integration.
  • Build and retain internal expertise for all crucial success roles.

Use Proven Processes

In established IT organizations, it is easy to take this pillar for granted. Most have had refined development methodologies for years. The problem is that if development of business intelligence applications follows the traditional waterfall process, the applications are put at serious risk. Traditional development methodologies do not allow for iterative analysis, design and build techniques that are the cornerstone of predictably successful developments.

Pillar Principles for Excellence:

  • Be responsible for unique content and integration with existing development methodology.
  • Be responsible for unique content in IT project management templates and methodology.
  • Be responsible for methodology and standards for assessing BI/DW projects or portfolios.
  • Be responsible for BI/DW in central architecture review board project reviews.
  • Insure BI/DW total cost of ownership is defined, calculated and communicated.

Pillar Objectives:

  • Buy or develop an iterative process model.
  • Integrate the iterative model into existing organizations and processes.
  • Empower the center of excellence through an architecture review board.
  • Build and maintain annual ROI and metrics.

Use Technology That Works

Managing business intelligence technology wisely involves managing several specific technology life cycles within a larger enterprise technology life cycle. While IT may be good at this, we often forget that senior clients want and need to be involved in the process, as they ultimately define whether or not the technology works.

Pillar Principles for Excellence:

  • Manage key technology life cycles with clients.
  • Define and maintain key architecture frameworks.
  • Provide support for quick proof-of-concept developments.
  • Continuously support improvements in customer usability.
  • Be the product owner for key tools.

Pillar Objectives:

  • Integrate and simplify technical architectures.
  • Maintain technical leadership.
  • Actively manage the technology life cycle.
  • Insure information delivery is highly usable.

Strive for Clean Data

Clean data is another area that can easily be taken for granted in large enterprises. Making data client-usable can be a daunting task if not well managed on a continuing basis. It is critical that information across the enterprise gives the same result to the same question. The average client would say, "If we query for the cost of subassembly X from the production bill of materials, and then query for the actual cost at manufacturing plant Y, we should get the same answer." This type of hypothesis assumes two levels of data quality, specifically, the cost value itself and the definition (calculations) associated with each cost. Understanding both is critical to client success. There may be perfectly valid reasons why they are different; however, if the client does not know that, they are at risk of making faulty decisions and questioning data credibility.

Pillar Principles for Excellence:

  • Have analytical data managed within the context of an enterprise warehouse road map, coordinated with an enterprise information architecture road map.
  • Integrate meta data with applications where it significantly enhances user knowledge.
  • Develop excellent meta data that reduces costs associated with developing and managing applications.
  • Develop metrics and tools to continuously improve strategic analytic data quality.

Pillar Objectives:

  • Actively manage data cleanliness.
  • Manage and use meta data throughout the enterprise.
  • Insure lowest granularity.

A Center of Excellence is not Free

We have now covered the four pillars of business intelligence. However, another primary success factor for building the business intelligence and data warehousing center of excellence is a budget. Seed money will be required to establish the CoE, and budget allocation and cost rules will need to be agreed upon to continue its operation. In large enterprises, there is usually more than one IT group focused on BI and data warehousing. Trying to consolidate these groups and individuals at the onset may not be politically achievable and may not even be necessary. If the CoE really leads and provides what the organization needs, these other groups will join CoE efforts.

Establish a Workable Budget Strategy

  • Insure your budget agreements do not require use of the proverbial tin cup.
  • Obtain at least 80 percent seed budget for the first year, with a plan to be 80 percent self-funded over the next three to five years.
  • Acquire executive business sponsorship of budget plans.
  • Set internal consultant rebilling targets for the next three to five years that also fund tool research and a proof-of-concept lab.

Insuring we have the basic skills, tools and processes in place to be an excellent and efficient provider of business intelligence and data warehousing service is of paramount importance. Applications involving these technologies are increasing in scope, cost and criticality to the business. It is time to get serious about providing real value to the business.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access