The results of Wal-Mart's RFID pilot are no secret—the world has known about the 16-percent reduction in stock-outs and two-thirds drop in replenishment times since October. Now we're learning a little more about how they did it. In a white paper released late last year, the mega-retailer identified specific process improvements that helped it achieve those dazzling results.

Bentonville attributes most of the savings to its newfound ability to generate automatic pick lists of the specific items needed to restock shelves. As customers go through checkout, the RFID system swiftly combines point-of-sale data on their purchases with RFID-generated data on what's available in the back room to produce the pick lists. Basically,Wal-Mart changed its pick-list generation process from a reactive one that required associates to take the initiative to determine restocking needs, to a proactive one, in which lists are automatically created in real time, based on sales.

Though the study didn't attempt to quantify the savings, it appears that the RFID-enabled process also saves time and labor. Associates no longer need to scan store shelves to determine what's out of stock. Nor do they have to scan cartons and cases arriving at the stock room; RFID readers mounted at the doors collect data on the incoming goods. And there are no more wasted trips to the back room. Associates now have assurances that a box will be there when they go to retrieve it.

"The message here is that RFID is making a difference," says Bill Hardgrave, director of the RFID Research Center at the University of Arkansas' Sam M. Walton College of Business, which prepared the report for Wal-Mart. "The message is that the data that RFID collects can allow for process improvements."

The tests themselves date back to last February. For the study,Wal-Mart tracked 4,554 unique items, representing merchandise from nearly all of the stores' departments, as they moved through 24 stores. Twelve of those stores were outfitted with RFID readers/antennae at various locations, including receiving doors, sales floor doors and box crushers. The other 12, the control stores, were not.

Among other results, the tests demonstrated the depth of tracing detail RFID is able to provide. In one test, for example, Wal-Mart was able to determine that a particular product arrived at its distribution center on Aug. 4, that it was put on the conveyor system five days later and that it departed shortly thereafter. Upon arrival at the store (12 hours after it left the DC), the product was whisked to the store's back room and moved to the sales floor the following day.

Hardgrave says that his team of researchers is working on another white paper that will provide further details on Wal-Mart's RFID pilot. This one will describe the impact of RFID on specific product categories and store departments within Wal-Mart.