RFID Implementation Is an Art

By Jonathan Collins

June 14, 2006—RFID technology may be ready for deployment, but according to Neco Can, it also requires a good deal of ingenuity and support to implement. Can spearheaded item-level RFID trials at U.S. clothing retailers Gap and Abercrombie & Fitch (A&F).

"RFID is not a science, but an art," Can told attendees at the RFID Networking Forum in London last month. "You won't believe what we learned [working with RFID]. No one knows everything about RFID, and every day [working with the technology], we would learn something new."

In one RFID trial, Can recalled, the test portal was unable to deliver a near 100 percent read rate on UHF-tagged clothing items, as required to make the testing of any RFID application viable. The systems integrators on the project recommended adding two additional interrogator (reader) antennas to the four already installed. However, this step would require adding a second reader to every portal, further driving up any deployment costs dramatically.

According to Can, his team found that if they placed an inch-tall metal strip on the floor across the portal threshold, they could achieve the required level of read accuracy. "It worked because it shook up everything that had to be read in the load as it was wheeled over the strip, and it cost just a few dollars," said Can. A similar process was also used on the rollers of a conveyor. "By putting little curves in the rollers," he added, "it would shake the box and improve read rates dramatically."

In 2001, Can was the director of Gap's project-management office, heading an RFID pilot that provided an inventory accuracy of up to 99.6 percent. Despite the success of the pilot in both improving in-store inventory and increasing sales, a host of factors caused the company to chose not to invest immediately in RFID tagging. The experience has led Can to believe that any RFID project within a company needs full support at the highest levels of management.