The Wal-Mart Factor

No question about it: Wal-Mart's RFID initiative is one of the biggest technology news stories of recent years. With so many folks talking about the project, it's difficult to figure out what's really happening and why it's important. Forrester Research got an update from Wal-Mart CIO Linda Dillman and the Wal-Mart RFID team just before the January milestone. Here's how the initiative is progressing, according to Wal-Mart:

• All 137 suppliers are participating. The original 137 suppliers -- Wal-Mart's appointed "Top 100" plus 37 volunteers -- are tagging cases and pallets going to three distribution centers. Wal-Mart calculates that, while tagging plans vary from less than 1 percent to 100 percent, suppliers are tagging an average of 65 percent of these supply chain units.

• January was not meant to be a "Big Bang". No one turned the switch on January 1. Rather, suppliers sent tagged cases and pallets through the first several weeks of 2005. Wal-Mart cited tag availability as a key factor. Carolyn Walton, the vice president in charge of RFID, explains: "Unfortunately, there have been suppliers who waited too long to place orders for tags. That means that they will be shipping in late January or even early February."

• RFID performance is improving. Since April of 2004, Wal-Mart and eight of the Top 100 have been testing tag and reader performance across both distribution centers and Wal-Mart stores for around two dozen products. While initial read rates were between 40 percent and 60 percent, the Wal-Mart team indicated that it's now seeing rates in the high 90-percent range at key read points. One exception: It's currently only reading between 50 percent and 90 percent of cases moving from the back room to the front of the store on a pallet. Still, the team sees this as significant progress, as Wal-Mart couldn't track this movement previously.

• Everyone is still learning. Dillman and the team also pointed to unanticipated benefits learned through the course of the pilot.

For example, Wal-Mart can now "see" if pallets and stocking carts are on the floor by 7:00 a.m. -- which helps Wal-Mart to reduce clutter in the aisles during prime shopping hours. When it comes to RFID's impact on product availability, the team admitted that it and the suppliers are learning how to measure the correlation. Walton shared the fact that product availability "is one number that we [still] can't measure. We won't know the impact until we start fixing and measuring."

The RFID initiative has been controversial. A lot of people remain fixated on whether or not suppliers met the deadline. This is not that interesting. Wal-Mart's efforts -- as well as those of the Department of the Defense and others -- have served their purpose. What was RFID talk is now RFID action. We're closer to a standard tag supported by both EPCglobal and ISO. More retailers now understand the specific improvements that RFID brings to their business. Most importantly, the biggest barriers have become clear.

Forrester has kept in contact with dozens of consumer goods firms and RFID vendors, and their experiences have helped us analyze these barriers. We think that attention will shift from tags and physics to software and data integration (the shift has already begun). Wal-Mart's success -- as well as that of its suppliers and other retailers with ambitious plans will depend on how quickly the RFID community is able to:

• Uncover "killer" use cases for consumer goods manufacturers. For many with low-cost, high-volume items such as food and mainstream apparel the biggest benefits, like demand management, depend on wide-scale deployments. With the RFID costs front-loaded and benefits back-loaded, the economics are still difficult for many suppliers.

• Lower B2B data exchange barriers. Many suppliers are grumbling about the incremental cost of the EPC network, considering their investments in global data synchronization. The Uniform Code Council and EAN International must re-evaluate the fee structure for the EPC and GDS networks to stimulate adoption.

• Bring multi-tiered RFID infrastructure software to market. There's a common myth that companies and their enterprise applications will drown in a sea of RFID data. Early pilots have produced terabytes of raw data, but companies don't need to use and store all of it.

Next-generation RFID middleware must prioritize architecture scalability, flexibility and data management. It also must manage more than just passive RFID data. Many processes, like global logistics, will require passive and active RFID and sensors that detect environmental conditions.

Innovate composite applications that make use of RFID data. RFID opportunities hinge on new processes -- most of which will require users to look beyond their current legacy and packaged applications.

More and more, firms will use composite applications -- a collection of Web services that knit together content and function in a specific sequence to manage RFID activities. This should be good news for enterprise application vendors with strong service-oriented architectures