Strong Signals: Tag! You're It!

By John Parkinson  |  Posted 09-12-2003

The value of RFID goes way beyond the supply chain. Imagine knowing exactly how old that carton of milk really is. Or how safe your tires are.

Did you know that the first item ever commercially scanned by bar code was a ten-pack of Wrigley's chewing gum? And that the first scan occurred in an Ohio supermarket in 1974—22 years after bar codes were first patented?

Bar codes have since become a staple of modern life, and they have many virtues: They're easy to read using a low-cost laser scanner and they tell you with great accuracy what you are scanning. But they also have limitations. You have to be able to see a code to scan it, and it's very challenging to perform accurate, fully automated scanning. And bar codes don't identify which carton of eggs or can of tuna you are buying, just who manufactured it and what kind of product it is, so they're not well suited for some tracking and recall activities. Nor do they generally hold price data—in-store systems have to do that—so scanners can be fooled into mispricing items. And finally, the 12-digit UPC code allows for "only" 100,000 manufacturers and 100,000 products per manufacturer, arguably not enough for our modern global economy.

Over the years, there have been several attempts to improve on bar codes. Proprietary schemes, such as those used by FedEx Corp.and UPS Inc., can improve information density and orientation sensitivity, but none has yet combined the bar code's simplicity, low cost, universal readability and the ability to add a printed value humans can read. If we want to improve upon the bar code, we are going to have to fundamentally change technologies.

Why do we want to? First, eliminating human operators and manual scanning of bar codes can reduce costs and speed up the supply and sales chains. Second, bar codes don't help eliminate "shrink," the $30 billion-a-year inventory loss in retail distribution. Third, the ability to track individual items will kick off a new level of supply-chain optimization. And finally, we can improve track-and-trace costs, recall management processes, returned goods handling and even recycling.

Enter the next generation of product identification technologies, which uses radio frequency (RF) communications to link a coded chip (ID) attached to an item with information about that item—hence RFID. To control costs, most of the tags—a chip plus an antenna—are unpowered; they pick up enough energy from a low-powered RF field emitted by the tag reader to trigger a short broadcast of a 96-bit string of data that forms the electronic product code, or EPC. The code consists of four fields, which provide unique identifiers for 268 million companies, each of which can have 16 million "object classes" (equivalent to SKUs) with about 68 billion serial numbers in each class. Thus, the EPC lets us not only identify a pallet or carton or pack, but also tag, track and trace unique items. And because RFID tags don't have to be visible to be read, a fully automated reading and ID system is feasible with some attention to engineering issues.

The simplicity of the EPC means the infrastructure around RFID has to be able to do a lot of work really fast. Most RFID readers can handle more than 50 reads a second from different tags, but up to 200 a second will be standard soon, and much faster readers may be needed later. The readers generally have to be within about four feet of the tag, so it's often necessary to put multiple readers in a shipping area or receiving dock. And the same tag will probably show up on multiple readers, so duplicates will have to be eliminated. Special software agents (called Savant by the Auto-ID consortium that is setting standards for all this) will take care of the buffering of tag data, error correction, duplicate removal and tasks such as associating a tag with a location. The tag does not "know" where it is—that info resides with the reader—or where it's been or where it's going; the reader has to interact with data from previous reads or shipping schedules to establish the course of the item's journey through the supply chain.

RFID's initial ambitions—tagging trucks and containers, pallets and cases, rather than individual items—may seem modest, but the incremental approach (shipping container to pallet to carton to item) looks right, because early field trials are turning up a lot of practical difficulties. An incremental approach also allows time for the tag manufacturers to get their production volumes up and costs down. And finally, it allows product manufacturers to redesign or modify their production systems so that tags can be automatically and reliably attached.

All these issues will take some time to sort out, but it looks as though we are past the point of supply-led innovation (where technology companies push a new capability into the market to see if it works and if anyone wants it) and into the demand-led innovation phase (where companies start to design business processes that can work only if effective enabling technologies are widely available). And that's when things start to snowball.

As great as the potential for RFID is in the supply chain, however, its impact will be felt beyond inventory management. First, consider asset tracking. In many organizations, such as hospitals, high-value assets often just sit around underused or always seem to be in the wrong place at the wrong time. People waste time searching for them, and there is good evidence that, because it's so difficult to track usage, too many assets are deployed. Add RFID, however, and you get a whole new class of management tools that let you optimize the number and location of assets of even moderate value by tracking and analyzing actual usage patterns—and a way to combat "shrink" as well. RFID asset-tracking systems have quick paybacks and generally don't require very high volume or complex technology, so they are likely to generate lots of interest in "asset-rich" businesses in manufacturing and distribution as well as in high-end retailing.

Second, deploying RFID adds the concept of "presence" to existing business processes, which can bring people into the RFID-enabled mix. With reliable presence information (via an RFID-equipped smart card, for example) you can add personalization to many processes in retailing and customer service: personalized shopping, dynamic pricing, targeted cross-selling. But that also raises serious questions about privacy and the potential for covert and overt intrusion into people's lives; indeed, this is already generating concern.

Finally, beyond these initial opportunities comes a whole new application area: item-lifecycle tracking. Eventually, we will be able to track the manufacture, location, ownership and usage of anything, forever. No more "sell by" issues. No more "used" items passed off as "new." No more fraudulent returns of items that were never purchased in the first place. No more doubts about provenance or gray markets. Applications in this area may arrive sooner than you think. The TREAD Act, which mandates the ability to monitor the condition of tires on automobiles, will probably result in the use of an RFID tag in each new tire manufactured as of the 2004 model year. It is only a short step from there to a vehicle "black box"-type recorder.

As—and if—we decide to deploy these capabilities, we will require much more data storage than anything we have today and will likely spawn a lot of new businesses selling track-and-trace services for every kind of item imaginable. It's interesting to speculate on just where in the value chain these services might fit: eBay, where you go to find what you want (and where item history is already described but not guaranteed)? FedEx or UPS, who deliver it to you and may want to guarantee that you get what you bought? Or the issuer of the credit card you used to pay for it, who wants to eliminate fraud?

Such capabilities aren't completely here yet—and may never arrive if we don't plan intelligently and systematically for them. Tags will have to get cheaper and we will have to look at new technologies (such as printing antenna patterns out of metallic ink) before we will completely replace bar codes. But RFID, even in its milder forms, is going to bring about a lot of changes in business and society. Better get ready—before you get tagged.

John Parkinson is chief technologist for the Americas at Cap Gemini Ernst & Young. Please send comments on this column to editors@cioinsight-ziffdavis.com.