February 27, 2006

The RFID Hype Effect

By Evan Schuman, Ziff Davis Internet

News Analysis: Deployment has slowed and even the players that gave RFID its first boost—including Wal-Mart—are starting to cool.

RFID was supposed to revolutionize the supply chain and—by mid-2006—dominate most aspects of product handling within retail and manufacturing. Today, even the most ardent RFID advocates are conceding that hasn't happened and it's quite frankly not even close.

There are many reasons why RFID's progression has been disappointing so far, but the most likely culprit is that the initial expectations were not realistically set in the first place.

The fact that many vendors, retailers and manufacturers overhyped RFID's likely timetable is not news, but the extent to which the hype was believed may be.

"My expectations were that things were going to progress further," said ABI Research RFID analyst Sarah Shah. "People have been less eager to actually jump into the technology" than they were to talk about it. "Obviously, there is initial hesitation."

Two events have signaled RFID's troubles. The first was Wal-Mart.

It's fair to say that Wal-Mart—along with Gillette and Germany's Metro Group—truly created the momentum for the RFID market.

After all, the theory went, the suppliers of those huge companies would have to go along. If Wal-Mart mandated it, what supplier would dare say no?

Initially, that is precisely what happened. But after error problems, conflicts, and extensive and expensive investments, even many of Wal-Mart's suppliers suddenly developed into vertebrates.

And Wal-Mart, which has a well-earned reputation among suppliers of being one of the few companies that can make dealing with Microsoft seem comparatively pleasant, also backed off.

One of the key factors that buoyed those Wal-Mart suppliers is that they were also suppliers for many other major retailers, and those other retailers were starting to ease off the RFID gas peddle. It wasn't as though retailers started issuing statements saying they no longer cared about RFID, but around the summer of 2005, the "urgency left. It dissipated. Suddenly, the retailers stopped asking about RFID," said a consultant to one of the larger suppliers.

The suppliers initially suffered from "unnecessary fear about what big, bag Wal-Mart would do" if they didn't enthusiastically deliver on Wal-Mart's RFID mandate, said ABI Research's Shah. But suppliers soon "saw hesitation on Wal-Mart's part, too."

One of the earliest proponents of RFID, Wal-Mart had an internal secret program to create its own proprietary version of RFID, said Randy Sweeney, one of the founding members of MIT's AutoID Center, which itself was replaced by EPC Global. Sweeney today works as a senior IT executive with a $90 billion global manufacturer.

Wal-Mart ultimately gave up on its proprietary efforts and joined other retailers and manufacturers forming MIT's AutoID Center, Sweeney said.

Wal-Mart executives "always believed that the kickoff date for RFID should have been 2003. They had no basis to believe that it was that close, but their pushing for it led to the earlier adoption of large-scale testing," Sweeney said. "It really led to expectations within the industry that things were farther along than they were."

Sweeney argues that Wal-Mart's attitude smacked of arrogance, that the company could make things happen by ordering them to happen. "There's a fundamental reality that we all live in, and that reality can't be changed by fiat," he said. "If you believe that sheer force of will can make anything happen quickly, that in and of itself is a misunderstanding of the world. It's not like with nine women and one month, you can get a baby. You still need the time."

Carol Rozwell, a vice president at industry analyst firm Gartner who tracks RFID from the manufacturing and pharmaceutical sides, said Gartner expects RFID to start becoming solid in the back-end applications space by next year and then experience "a slow, gradual adoption due to its complexity, lack of standards, the need for cooperation and the volume of data and rules for dealing with that data."

Rozwell's point is that, unlike the vast majority of technologies today, RFID has the potential to not merely accelerate, simplify and improve the handling of data, but it's such a significant advance that it will also likely force new ways of using and retaining data. Although that's ultimately a good thing—in theory—it is one of the reasons that RFID is going to take so long to fully deploy.

"We're going to need to think differently about data. You don't need to know where [a product] is every 5 seconds in your supply chain," Rozwell said. "You only need to know the aberrations."

AutoID's Sweeney agreed, stressing that the last few years of the RFID struggle will seem like playtime compared with the next couple of stages of true integration.

"Once we learn that there will have to be a new way of doing things, which is coming up by the end of the decade, that's when the real pain starts," Sweeney said. "As painful as this is, this isn't the bad one. The bad one's coming up."

"When you have end-to-end product transparency, people will realize, 'Maybe the way the company is organized isn't quite right,'" Sweeney said. "The organization of businesses today is based on the 19th century business model" where mail was how communication with customers and suppliers happened and freight trains and boats were how products were received and sent.

But before we even get into the second-level agony that awaits the next stage, RFID's first-stage isn't quite done with us yet, Sweeney said.

"The fact today is that we still do not know how to tag a great number of SKUs," he said. "In order to achieve the savings, the benefits, many RFID business plans are based on 100 percent accuracy" and every item being tracked.

Sweeney also said that many retailers did not fully engage their IT departments early enough, which was especially troublesome given that "the XML software standards were not really gelled," forcing more than typical integration headaches.

"These learnings come with engagement, and they carry a price. You really need to be deeply in the game," he said. "Most IT organizations, certainly in the early RFID days, they set aside special instances so they test [RFID] isolated from the rest of the network rather than integrate them into their business systems. The idea that you can do a small stand-alone simulation of the world and then take that and plug it into the [true enterprise] is ridiculous."

Still, Sweeney—in expressing a view that meshes with most RFID analysts and end users—doesn't truly criticize RFID development. Isolating RFID test systems was the only responsible way to do that testing, and the years of pain were going to be necessary at some point, so nothing truly has been lost, he said. "The only net effect is that Wal-Mart's push got the industry to where it would end up anyway, but did so somewhat earlier.

"We are where we originally expected to be, not where some people desperately wanted us to be," Sweeney said. "Wal-Mart had desperate desires. They truly believed that they could will it into being."

Paula Rosenblum, who tracks retail technology issues for the Aberdeen Group, said she isn't at all puzzled about why RFID has taken longer to deploy than most vendors and retailers promised.

"The more interesting question is, 'What the heck were they thinking of, anyway?' The technology was immature, the standards weren't set, and I don't believe there is any ROI for the supplier without a stunning amount of collaboration with the retailer once the other parts are solved."