Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

May 31, 2003

The Quest for the Universal Data Model

Like the Holy Grail, the universal data model may be unattainable

by Joshua Greenbaum

Search the Web for "enterprise software" and "Holy Grail" and you'll discover one of the great industry truths: The hundreds of times we use those terms in the same sentence proves that we're always searching, and rarely finding, that elusive software solution that will somehow make all the wrongs in the world right again. The quest isn't just quixotic; it's confusing as well. One reference site describes "an obvious obscurity surrounding the details" of the Holy Grail. Unfortunately, anyone who has tried to get to the bottom of some of the latest software offerings knows all too well what this site means.

An "obvious obscurity" just about sums up the latest enterprise software quest, the universal data model. Here's the theory: If only users, or vendors, could agree on a single universal definition of a customer, an order, or a bill of materials and then use it to design an industrywide data model, all would be right with the world. Business-to-business commerce could now be easily automated, and everyone could freely exchange information without the headache and expense of translating data from one system to another.

The Trouble With Paradise

It's a highly laudable position that's flawed only by the utopian magnitude of its vision. It turns out that corporate data models express an important component of individual business culture that many companies are loath to cede to some outside third party or standard. Most companies shouldn't be very interested in giving up that flexibility in order to create a universal data model, even if it lets them greatly simplify their business-to-business interactions. And creating that universal data model also means creating a level playing field for their competitors. Unfortunately, for these reasons the concept of a universal data model is going to be seriously difficult to promote in the market.

It's not just that users probably can't agree on a universal data model for a given industry; no one can. Rick Bergquist, chief technology officer of PeopleSoft, was once asked by an industry analyst why he doesn't sit down with other vendor CTOs to thrash out a universal data model. Rick's well-tempered response was that he would gladly oblige just as soon as the analyst met with all his competitors to agree on a framework for how the vendors should proceed towards such a goal. Touché, Rick.