Continue in 2 seconds

How BI Evolved

  • July 01 2002, 1:00am EDT

Recent discussions with friends, colleagues and business intelligence (BI) professionals regarding my article "New Directions for Business Intelligence" (DM Review, April 2002) have caused me to reflect on the source of our knowledge about BI and what it means for those of us who believe that information is undervalued and underutilized in business today. I heard from people who experienced the situations discussed in the article on their way to best practice implementations, others who felt that issues such as conforming dimensions and data synchronization needed addressing, and yet others who recognized similarities with data practices from the past. The path to BI today is one of innovation and practical application.

That Was Then ...

The drive to enterprise-wide information, executive dashboards and data analytics began two decades ago. The technology was referred to as decision support systems and executive information systems, terms that are still used today. There was software designed to integrate data from several mainframe legacy systems. The need was there, as was the associated marketing hype for early products, but the technology was lacking. In particular, using separate data stores provided information that differed with source systems reports. The result was that during the 1980s, report writers (especially those that could be used without the help of IT) were favored by business people while executives looked to the finance department to satisfy their information needs.

These early attempts established the importance of data accuracy (data that is correct and stands alone without needing to be decoded), synchronized data (so that comparisons of information become apples to apples) and enterprise-wide data. BI best practices developed out of these important perspectives.

Data that stands alone is a defining characteristic of the relational model for data management. Originating with the work of Edgar F. Codd in the 1970s, the relational model is still the only data management approach founded on mathematical principles. The process of data normalization is nothing more than a set of rules to ensure that data stands on its own. Commercial availability of relational database management systems (RDBMSs) in the 1980s and their implementation in the business world can be attributed to the work of Codd and Chris Date to promote the relational model.

A push to data dictionaries and data administration began in the late 1970s and grew during the 1980s. This was the first corporate attempt to define and understand data. Data administrators identified and defined enterprise data, focusing primarily on application data. Client/server systems and, more importantly, the move to decentralized acquisition of them by business departments in the 1990s, brought proprietary and poorly documented data into the enterprise.

The recent trend to massive ERP, CRM and SCM applications and their large, complicated data stores continues to introduce poorly documented data into the corporation. Cynics among us say that software vendors document their proprietary data stores poorly to increase reliance on the software vendor's products and services. The more charitable say they simply don't know any better. The ongoing implementation of purchased software has ended centralized data administration in most companies.

Synchronizing data makes information usable across the enterprise. Synchronization aligns a "slice" of financial data, for example, with "slices" of customer and product data. The concept of conformed dimensions concerns synchronizing data around the dimensions used by the business for reporting and analysis. Ralph Kimball recognized the importance of using dimensions to make information easily accessible for business people. The use of dimensions is an integral part of BI suites and platforms today, but not yet integral to software applications.

Data with conformed dimensions in a central data warehouse is essential. It consolidates accurate and synchronized enterprise-wide data in one place. The Corporate Information Factory espoused by Bill Inmon and Claudia Imhoff and the hub-and-spoke architecture espoused by Kimball ensure that data, to the degree necessary, is accurate and synchronized across all enterprise data marts.

... And This Is Now

The goal of information integrated across the enterprise remains the holy grail (mythical and unattainable) for most organizations, even those committed to BI. Companies still buy application software packages and BI suites and applications. Companies increasingly share business data with suppliers, partners and customers (this just in ­ XML is not a data standardization strategy!). Companies, therefore, continue to underutilize and undervalue their data.

The news, however, is not all bad. Consider the following BI trends:

  • BI suites, platforms and applications continue to make it easier to tailor data presentation and delivery for each individual using BI.
  • BI tools continue to lower the total cost of ownership for developing and maintaining BI solutions.
  • BI products continue to improve their sharing of meta data (although most BI vendors do not make this as easy as they should), improving the ability to ensure consistency of data from operational source systems to delivered intelligence.

If the technology is improving, what keeps BI from being phenomenally successful and universally used? In most companies, executive management does not understand the value of BI and the role it can play in improving business results. Executive understanding and commitment, as in so many company initiatives, is essential for success. BI professionals must work to educate and persuade business executives to make BI an essential component of achieving their strategic business goals.

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