Infomation Week offers insights:
The nucleus of the IT infrastructure, [CIO Linda] Dillman presides over is a single, centralized, 423-terabyte Teradata system that churns data from 1,387 discount stores, 1,615 Supercenters, 542 Sam’s Clubs, and 75 Neighborhood Markets in the United States, plus 1,520 more stores worldwide. “That’s key to how we can leverage what we do into the future,” says Dan Phillips. The VP of operations, data warehousing, databases, large systems, and communications led the IT effort when Wal-Mart opened its first international unit, a Sam’s Club store in Polanco, a suburb of Mexico City, and was a critical player in the company’s decision in 1995 to bring together all its businesses under a common IT system for distribution, replenishment, and so on. “The common system, centrally managed, is our competitive advantage at Wal-Mart,” enabling the same data set for both buyers and suppliers, he says.
Wal-Mart in the 1990s tried having IT executives report to the business, a “well-intentioned” but not well-executed experiment, Phillips says. “We lost touch with what was going on in [the IT group] and with being able to leverage the synergies of being able to do everything for everyone.”
Key to Wal-Mart’s development efforts today is its build-it-once-for-all-systems mentality. That means build it for both domestic and global operations–the retailer’s growing international presence encompasses operations in nine countries and Puerto Rico, including the most recent acquisitions of Brazil’s Bompreco in March . “When you’re writing the code, you automate, enhance, and change processes globally,” says Tony Puckett, VP of international systems. As Wal-Mart learns from its experiences in new countries–about Brazil’s complex tax structure, for instance–“we bring back the structure into our core system and that becomes a tool we have in the future,” he says.
Today, Wal-Mart captures all the day’s sales and product data across its global operations on an hourly basis. Database queries can start running as soon as data is available. That ability comes in handy, particularly on the Friday after Thanksgiving, when Wal-Mart buyers start watching what’s happening in stores at 6 a.m. on the East Coast, then use that data to make decisions in real time that can affect the big day’s sales. Wal-Mart once used its data prowess on a Black Friday to query sales of a PC advertised in a circular; when execs found out it wasn’t selling well, they called stores and discovered the reason was that customers thought they had to pay separately for the system and monitor. So store clerks quickly put the two boxes together and spelled out the pay-one-price deal in a sign. “We’ve done a lot of work for performance and availability, and making sure the data is current,” Phillips says.