Wal-Mart declined to comment on why it pulled the smart-shelf test. But deploying RFID in stores has never been a top priority for the retailer. In fact, Wal-Mart had delayed the trial numerous times since January, to the frustration of Gillette executives.

The trial was conceived as an experiment to see what kind of real-time information could be gathered so that Wal-Mart and Gillette could begin to figure out how to use the data. Both companies knew it would not be economically viable to deploy the technology widely in stores for several more years. And Wal-Mart didn't want to disrupt its in-store operations. Even a small test of the smart shelf would require resources to support the technology and personnel to make sure the test boxes and regular boxes, which look alike, didn't get intermingled.

Did negative press reports about the potential of using RFID to track consumers' actions play a role? Again, Wal-Mart would not comment on this. But it's quite possible that the conservative company didn't want to risk the ire of privacy advocates over a trial that wasn't critically important.

But the cancellation of the trial in no way undermined Wal-Mart's commitment to RFID. To stem confusion in the industry, Wal-Mart hastily sent a letter to its suppliers letting them know that the retailer remained committed to tracking pallets and cases with RFID technology beginning in 2005. Wal-Mart spokesman Tom Williams said the company wants to devote its attention to its ambitious plan. "By 2006, we will roll it out with all suppliers," says Williams.

Industry experts believe that, given the huge commitment of IT and operational resources necessary to fulfill its mandate, Wal-Mart could not afford to be distracted by a smart-shelf test that wouldn't reap any immediate benefits. Edward Rerisi, director of research at Allied Business Intelligence Inc., a market research company that focuses on wireless technologies, says he doesn't believe "anyone should read anything into" Wal-Mart's decision to back out of the smart-shelf pilot. "It was two separate applications, two separate projects," says Rerisi. "You can't evaluate them in the same light."

Next page: Inside Wal-Mart's RFID planning.