Where's that one-handed economist when you need him? Market watchers following the latest developments in RFID likely feel some kinship with the frustrated Harry Truman, who famously demanded a one-handed economist, complaining "All my economists say 'On the one hand É. On the other'." Like Truman, RFID observers say they're confused by the mixed signals they're getting. Continuous reports of advances from the front lines keep raising their expectations, but every now and then, someone steps in to burst the bubble (see the accompanying sidebar).

On the positive side, RFID tag prices have begun to drop. Avery Dennison has announced that it will offer Generation 2 inlays at what it claims is the lowest price in the industry—7.9 cents per unit for orders of one million or more. UPM Rafsec, an RFID tag and inlay manufacturer, is making both its Gen 1 and Gen 2 inlays available at "sub-10 U.S. cent pricing" with a minimum delivery of 50,000 pieces. Not to be left out, Alien Technology has cut the price of one of its Gen 1 RFID labels to 12.9 cents.

Another positive sign: the transition from the older Generation 1 technology to the more-powerful Generation 2 appears to be well ahead of projections. By early next year, hundreds of millions of Generation 2 RFID chips are expected to flood the market. Chip maker Impinj, for example, intends to ship 50 million Gen 2 tags by the end of 2005. "We've already started our ramp-up in production, and the demand we're seeing now is quite good," says Bill Colleran, Impinj's chief executive officer. "It's an indication of the transition starting to move very quickly from Gen 1 technology over to Gen 2."

Though Impinj is the only company that has issued a solid volume prediction, other big silicon producers, like Texas Instruments and Philips, are also cranking up production. All in all, it's estimated that up to 700 million Gen 2 tags will be in use by the end of 2006.

At the same time, RFID appears to be making headway in the public mind—or at least that part of the public familiar with supply chain management. In a recent survey conducted by EPCglobal, more than 85 percent of the 400 U.S. supply chain executive respondents said they considered RFID technology "extremely important," "very important," or "somewhat important" to their supply chain management operations. The survey also revealed that 99 percent of those who consider themselves familiar with RFID deemed the technology to be "extremely," "very" or "somewhat" important to the future of supply chain management. Surprisingly, the survey noted that overall, manufacturers are slightly more likely than retailers to see the technology as important.

... on the other hand

Just when it seems the RFID market couldn't get any hotter, along comes an industry analyst with a bucket of cold water. In this case, it's AMR Research, which published a study last month that questions whether retailers will ever benefit from incorporating RFID into their operations.

Although many retailers believe RFID will eventually transform their supply chains, the study says, none has yet been able to develop a solid business case for expanding beyond the pilot stage to an enterprise-wide rollout of the technology.

AMR bases its discouraging projection on a mock case study it conducted on a fast-moving consumer goods retailer with annual sales of $5 billion and an operating margin of 3.25 percent. The retailer took three years to roll out RFID at eight distribution centers and 200 stores, at the end of which time it had tagged 45 percent of its SKUs.

In the final accounting, AMR says, the numbers did not justify expanding the use of RFID across the enterprise. The company spent an estimated $39 million to make its operation RFID-ready—buying readers for the stores and DCs, increasing networking and storage capacity, and upgrading the existing supply chain software. To recoup those costs, the company would have needed to generate an additional $1.2 billion in sales. The retailer, however, fell far short of that goal. Even the considerable labor savings, estimated at $2.5 million, turned out to have little effect on the ROI.