News

Gartner: Data mart promises don't add up

• By Rich Seeley
• April 2, 2003

In a room filled with IT professionals attending a seminar on data warehousing and business intelligence, Gartner's Kevin Strange asked how many had listened to a departmental end user complain: "That's what I asked for, but that's not what I want."

That is why Strange, Gartner vice president and research director for data warehousing, said he recommends caution when developers are asked to build specialized data marts.

The problem with data marts, he said at the recent Gartner ITxpo in San Diego, is that the development time too often exceeds the business need they were supposed to address. "Data marts take too long to implement and opportunities are lost," he said.

The current business climate is so dynamic that departmental users find that the mart they needed three months ago or six months ago, is not the mart they need today.

The alternative that Strange recommended is for IT departments to build sophisticated and flexible data warehouses that can be quickly adapted to meet the immediate needs of departments in the enterprise.

In building that adaptable data warehouse, he suggested that IT standardize on one platform. He said that too often data warehouse managers are faced with a hodgepodge of technologies requiring different management tools. This can lead to a situation where the cost of maintaining the data warehouse outweighs its benefit to the organization.

In ranking database vendors, Gartner lists, in order, IBM DB2, Oracle and Teradata.