Last year, Wal-Mart laid down the law. By late 2005, the Bentonville, Ark.-based retailer wants to be able to track all pallets and cases from its top 100 U.S. suppliers using radio frequency identification (RFID) technology and Electronic Product Code (EPC) markings. The rest of its U.S. vendors are expected to fall into line by the end of 2006.

The majority of Wal-Mart's consumer products suppliers seem to have accepted the giant retailer's order to apply EPC tags to their pallets and cases. Still, important questions remain: Is this technology really ready for widespread deployment in the supply chain? Has it been fully tested, and can its cost be justified? In short, are RFID and EPCs ready for use in the business world?

The EPC movement got its legs in the 1990s, when the Auto-ID Center at the Massachusetts Institute of Technology (MIT) began exploring ways to use radio frequency identification technology as a better alternative to bar codes. RFID consists of a tiny microchip, an antenna, and a transmitter tag, which stores a unique identification number. MIT developed standards for EPC tags—in essence, "license plates" that identify products and transmit that information to computers. (EPC, by the way, is a trademarked term.)

Once activated by an RFID scanner, or reader, a tag sends its identification number to the scanner via radio signals, making it possible to track and almost instantly update an item's location. RFID surpasses bar codes because it can read tags simultaneously, and because there's no need for a direct line of sight between the reader and the tags.

The Auto-ID Center handled the research and development stages of the EPC standards. The center has since turned over the commercialization of those standards to the Uniform Code Council (UCC). Sue Hutchinson, project manager at the UCC's EPC Global division in Lawrenceville, N.J., says the UCC has published Version 1.0 of specifications for EPC networks, including the code itself, the tags, readers, and several of the networks' software elements. She says the UCC will work to refine those specifications, and will "hone them on the basis of real-world experience."

Several leading corporations, including Procter & Gamble, Gillette, and Coca-Cola, to name a few, have conducted field tests. Those tests have turned up a host of technological and process challenges. Even Wal-Mart admits to having some difficulties. Spokesman Tom Williams says problems encountered so far include missed signals due to improperly positioned readers, difficulty in determining appropriate antennas for certain settings, and the lack of a universal RFID protocol.

At the PackExpo trade show last October, Wal-Mart's manager of global RFID strategy, Simon Langford, said field tests revealed problems such as reader rates that were less than optimal for efficient production, radio frequency signals interfering with installed wireless local-area networks (LANs), forklifts knocking readers off their mounts, and antenna distortion around metal dock doors. He also discussed the high cost of tags, which fetch 15 cents to 30 cents apiece.

RFID's poor performance around water and metal, including consumer goods such as shampoo, canned food, and home electronics, poses yet another obstacle, says Michael J. Laird, a senior analyst with Venture Development Corp. in Natick, Mass. In addition, Laird says, users must deal with the incompatibility of International Standards Organization (ISO) and EPC standards—an issue that is now being addressed by UCC's EPC Global organization. Finally, he notes that the cost of an RFID system implementation can amount to several million dollars per supplier.

Santa Clara, Calif.-based Sun Microsystems, Inc., which recently opened an RFID test center for Wal-Mart's suppliers, has also encountered problems. Sun took part in some early pilots, including one in 2003 that involved Procter & Gamble, Target, and Gillette. Vijay Sarathy, product line manager for Sun's Auto-ID Infrastructure Solution, says the pilot highlighted two areas of RFID that need more attention: how to better affix the tags, and how to make readers more accurate at the item, case, and pallet levels.

"There were a lot of lessons learned from that pilot," says Sarathy. With cell phones, cordless phones, and bar-code systems all using the same frequencies, for example, interference was common. The pilot also highlighted a need for more rugged tags and more creative placement, such as in between the pallets, where the tags couldn't be damaged as easily. Metal packaging also posed a problem because reflection from the metal makes the tags difficult to read. "The word is that a little sticky tape or other material that separates the tag from the metal will allow the metal objects to be read much better," he notes.

One of the co-founders of MIT's Auto-ID Center dismisses such obstacles as "teething problems." Sanjay Sarma, associate professor of mechanical engineering at MIT, believes those difficulties can be resolved through better engineering. "There were no fundamental problems," he says. Where RFID could pose more significant challenges, he suggests, is on the process side of the implementation, or the actual integration of RFID into companies' and workers' everyday routines. "RFID changes the game fundamentally," he explains. "So while it makes operations more efficient, it's also a whole different ballgame."

Despite the problems encountered in field tests, Wal-Mart is sticking to its mandate, confident that the bulk of its suppliers will comply with its deadline. Williams says the company will start using RFID in three of its Texas distribution centers in early 2005, and will roll out more systems throughout the year.

For companies like Wal-Mart, the incentives to switch all suppliers over to a state-of-the-art RFID system are significant. Calling RFID a "huge leap forward," Williams said he expects implementing that technology will reduce out-of-stock situations, which occur when a consumer comes in for a certain product and finds an empty shelf, either because the store doesn't have the product or because the person working the floor doesn't know that the desired item is just a few feet away in the stockroom. "At that point, the tendency is to order more," says Williams, who estimates Wal-Mart's in-stock rate at higher than 99 percent. "That gap between 99 and 100 percent in-stock represents $1 billion in sales for us," he says.

While Wal-Mart touts the benefits of reducing out-of-stock levels, saying that it will generate savings for its suppliers, some industry analysts suspect that manufacturers won't see a return on their investments in RFID anytime soon. Laird expects Wal-Mart's top suppliers to spend a "significant" amount of money on RFID systems, especially on infrastructure, through 2005 as they scramble to meet the retailer's deadline. And in its recent report, Meeting the Retail RFID Mandate, Chicago-based consulting firm A.T. Kearney predicts that retailers will "reap extensive inventory and labor cost savings from the adoption of RFID," but adds that consumer products manufacturers will face higher costs and delayed benefits from adopting the technology.

Those costs include, but aren't limited to, the purchase of RFID readers, the building of the technology infrastructure, and the labor involved with tagging every case and pallet. A.T. Kearney estimates that a mid-tier grocery products manufacturer with annual sales of $5 billion, shipping 3.2 million pallets and 218 million cases annually, would require approximately 221 million tags to cover all of its pallets and cases. At a price of 15 cents per tag, that company theoretically could shell out more than $33 million annually just for tags. Even if the tag price were reduced to 5 cents, the manufacturer's annual cost would still be a hefty $11 million.

"Initially, manufacturers will only get cost, and no value," says Steve Banker, service director, supply chain management at ARC in Dedham, Mass. "Over the course of many years, as RFID is used across the extended supply chain, there may be ROI for manufacturers, but don't expect to see that for four or five years."

For now, says John Hill, principal with consulting and systems integration firm Esync in Toledo, Ohio, the RFID trend is more about keeping Wal-Mart happy than about improving shipping and receiving processes at the manufacturer level. RFID technology and the labor required to implement it represent new costs for manufacturers that won't be recouped, he says. "During a Wal-Mart announcement in Bentonville recently, I asked whether the retailer would consider sharing the wealth of the many benefits that it will derive from RFID," recalls Hill. "Everyone laughed at me."

As technology vendors work through the glitches spotlighted by recent RFID trials, and as companies like Wal-Mart get ready to enjoy the benefits that RFID adoption will bring, logistics managers are probably wondering just what these changes will mean for them.

The answer is pretty clear: With organizations such as the U.S. Department of Defense, Target, U.K.-based retailer Tesco, and German retailer Metro AG all conducting their own RFID tests and issuing usage mandates, it's only a matter of time until virtually all manufacturers will be expected to integrate RFID into their operations.

Does that mean it's time to chuck your bar-code systems and replace them with RFID? Not yet, says Stephen Lambright, vice president of marketing at Sunnyvale, Calif.-based Savi Technology, a member of the Auto-ID Center. He likens the RFID/EPC trend to the development of a new airplane. "We're at the point right now where we've designed the aircraft, but we're not even near the point of being able to taxi down the runway with passengers yet," he says. "We've taken test flights, but there are still a lot of final design issues that need to be addressed before the plane is ready for a commercial flight."

Because demand for RFID and EPC tags won't start and stop with Wal-Mart, ARC's Banker advises manufacturers to start small by educating themselves about what's required to comply with these early mandates. At the same time, they can try to squeeze some funding out of their budgets to cover initial RFID equipment costs. Then they need to ask themselves questions such as, Can we start out doing this manually? At what point should we automate the process? Do we implement it in the factory? Do we do it just in our distribution centers?

"Eventually, it's going to be much bigger than Wal-Mart, so now is the time to start thinking about it," says Banker. "There are a lot of decisions to be made over the next six months."