A Penny Saved

The return from RFID would be much higher, however, if the cost of the technology were lower. To a large degree, the inflated expenses associated with setting up an RFID system are responsible for the technology's lukewarm reception—and a prime reason why companies are moving gingerly into adoption of the technology. "The challenge for companies is how to offset the expense," says Mark Roberti, editor of RFID Journal. "And unless a company has a very targeted, carefully drawn up and highly discrete application with a built-in ROI, there's no easy answer for this yet."

A large supplier could pay as much as $9 million in the first year to comply with Wal-Mart's narrow regional mandate, with about 80 percent of that going to the electronic tags for upwards of 16 million cases and pallets, according to Forrester Research Inc. The next highest expense involves the readers for scanning the EPCs as the products leave the manufacturer's distribution centers, and, surprisingly, the additional labor costs for manual tagging—automated tagging and embedding RFID in packaging material is far from perfected.

Currently, tags cost anywhere from 20 cents to 40 cents each, depending on volume purchased and the quality of the tags. "If you're shipping a high-margin product, that's a negligible amount, but if you're shipping tomato paste, where the margin on a case may be 20 cents, adding a 40-cent tag turns your product into a losing venture," says Dave Douglas, executive vice president for products and strategy at ConnecTerra Inc., a maker of RFID middleware.

That's why most large consumer-goods companies are minimizing their financial exposure by not going beyond Wal-Mart's requirement that they use RFID tags in the Dallas area. For even the biggest Wal-Mart suppliers, such as P&G, this will limit tagging to just 5 percent or so of their cases and pallets. VF Corp., maker of Lee and Wrangler jeans, among other consumer products, illustrates the wait-and-see response of suppliers to Wal-Mart's initiative. Most of VF's shipments to Wal-Mart are so-called multi-SKU boxes, which means that each box contains a variety of products in different sizes, colors and styles. Tagging these boxes is like slapping the label "building" on a house; it sheds no light on what's inside. Consequently, unable to justify RFID tagging of pallets and cases, "VF knows that it will just have to accept Wal-Mart compliance as a cost of doing business, with no payback," says Roberti.

VF's gains from RFID will come only when the price of the tags drops low enough to let the company tag every item it makes as the product is being manufactured, says Jim Jackson, VF's RFID project manager. When that happens, VF will be able, theoretically, to track products from plants in Asia through the "black hole" that items enter when they leave overseas factories and are placed on ships bound for the U.S., and then follow them as they're packed in cases and sent to distribution points, warehouses and stores. As retailers install shelf readers, VF could eavesdrop on sales activities at stores, thereby gaining a much richer idea of the specific color and size of items that are running low and need replenishment. That type of visible supply-chain application will be cost-effective and viable only when EPC tag prices slip to about 5 cents each, RFID experts say. And that's not expected until 2006, at the earliest.

But that also assumes that all the other critical RFID constraints will have been addressed by then. The accuracy of RFID systems is still well below the 99.9 percent rate that is required before they can be trusted to replace bar codes. The key problem is that with current technology, the tags have to be placed in just the right locations on boxes and individual items in order for readers to detect them. On top of that, RFID signals can be disrupted by metal, which reflects radio waves, and by water, which absorbs them. If 48 cases of Evian are stacked on a pallet, it's still unlikely that a case in the middle will be identified. And since most readers are programmed to scan their environments constantly—often as frequently as every five minutes, or less—to report on current inventory or the presence of new assets, inconsistent results would paint such a confusing picture of a warehouse that it could turn a supply chain into a Tower of Babel.