What does it take for an organization to be a world-class business intelligence organization? What separates the average company from the really well-run company?

Business intelligence (BI) has many facets. Many technologies and many uses of BI are required in order to get the full value out of the BI infrastructure. What are the characteristics of an organization where BI is truly a competitive weapon? This question was posed to me by a large, well-respected organization. I responded by saying that the characteristics of a world-class business intelligence organization are as follows:

There is a single version of the truth, and that version of the truth is at a detailed level. The data that forms the backbone of BI must be singular and granular. At the most detailed level, the data resides in what can be called the "system of record." Once data resides at the system of record, it can be used in many ways and at the same time support reconciliation. In addition, by having data reside at the system of record, there is a pool of data that stands in readiness for new requirements. Because the data is at a low level of granularity, it can be shaped one way at one moment and reshaped another way at another moment. Without this basic foundation, neither of the objectives of flexibility or reconcilability can be met.

Data is integrated as it enters the data warehouse. There is no question that integrating data is difficult. Integration is imprecise, complex and involves guesswork and sleuthing - tasks which many people feel uncomfortable performing. Integration is infrastructure work, which is thankless. For these reasons, many organizations are scared off by integration. Yet, for all of the difficulties of integration, it is absolutely worth it. By integrating data, the organization puts itself into a position in which it is possible to reap many benefits. The world-class BI organization is not scared of integration and the nasty work that is involved. If there is good news about integration, it is that there are tools designed specifically for integration and that integration needs to be done only once.

World-class BI organizations recognize that there are multiple users with a very diverse set of information needs. Consequently, there are a multitude of analytical tools that serve the organization. The world-class BI organization does not look at the world as a place in which one analytical tool will serve all purposes.

These organizations have a separation of processes. For a variety of reasons, it makes sense to separate processing in a BI environment. There is no single place where all analytical processing is done. Instead, departmental data mart processing is done in one place, on a separate machine for each department. Exploration processing is done elsewhere, where the statistician can get to the data without interfering with the rest of the world. The operational data store (ODS) is in yet another place, separate from the data warehouse, so that ODS processing can be done without consideration from the data warehouse users. There is a distribution of processing in the world-class BI organization. There are several good reasons for this separation - economic, technical and organizational. It is much less expensive to buy hardware so that the workload is separated over many distributed machines. The most expensive processing power you can buy is processing power that is centralized and massive. Distribution of processing is so technically important because each component can be optimized for whatever task is being done. Statistical machines can optimize statistical processing. Departmental machines can optimize multidimensional processing. ODS machines can optimize performance and so forth. By trying to cobble everything together onto a single machine, a large centralized data warehouse machine becomes optimal for nothing. The organizational reason for distribution of processing power is politics. By distributing processing, each organization can "own" its data and processor. Distribution of processing is a hallmark of world-class BI organizations.

These organizations recognize the life cycle of data. Data enters a system, it is integrated, it grows old, then it grows even older. As data ages and as its probability of access changes, it needs to be placed in different locations and on different storage media. Simply taking data and placing it in high-performance storage for the entire life of the data is wasteful of resources and hurts overall system performance. World-class BI organizations recognize the life cycle of data and build systems accordingly; they do not simply gather data into a single place, put it on high-performance disk storage and attempt to use it there forever.

World-class BI organizations build their systems in accordance with business needs. They have end-user involvement. The end users' requirements and the business need for intelligence are factored into the BI environment from the very beginning. It is clear - from a business standpoint - why the BI infrastructure is being built from the outset.

All data in the BI environment is created under a single architecture. The architecture is the master plan for how the organization should proceed. The master plan is based on such important premises as: affordability, ability to be built from existing technology, an evolutionary approach (as opposed to a revolutionary approach), the need for performance throughout the architecture and the need for reconcilability throughout the architecture. Every piece of data fits into an architecture. There are no "loose cannons" where data is separated from other data in an unreconcilable manner, such as in an independent data mart, for example.

The infrastructure is built iteratively, one brick at a time. The world-class BI organization does not attempt to build systems in a big-bang approach. The world-class BI organizations know that end users operate in a mode of discovery. End users cannot tell what the information needs are until they see the possibilities. Therefore, it is understood that a certain part of the infrastructure will need to be reworked. Because rework is an integral part of the data warehouse environment, it is not built in a massive manner. First, one small part of the environment is built. It is reworked at the same time that the second part of the environment is built, and so forth.

There are undoubtedly more characteristics of a world-class BI organization, but these characteristics top the list. 


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