The deadline is less than two weeks away. On most days, the head of the supply chain for one of Wal-Mart's top suppliers wishes it was two years away.
His company is one of the largest consumer goods manufacturers in the world, and by January 1, 2005, he is supposed to have a system in place for attaching radio frequency identification (RFID) tags to a portion of products destined for Wal-Mart stores. But this particular IT executive already knows he isn't going to make that deadline.
Sure, he'll stick RFID tags onto just enough pallets to satisfy the folks in Wal-Mart's Arkansas, headquarters, but he's not certain those tags will even be functional upon arrival because of technical problems. And that means the efficiencies that Wal-Mart has been dreaming of achieving - the RFID-enabled transparent supply chain - may not happen anytime soon.
"We don't have a business case for RFID," says the supply chain executive, who insisted on anonymity. "Because the standards are not complete, the equipment isn't developed. And because the equipment isn't developed, I can't fulfil Wal-Mart's demand."
This executive is far from alone. Even though no suppliers will admit publicly that they may not meet the deadline, privately some say that meeting Wal-Mart's expectations is just not possible - at least by the deadline the retail giant has set. And the mission itself has become a moving target. Originally, Wal-Mart insisted that its top suppliers put RFID tags on all of the products shipped to specific distribution centres in Texas. Now, Wal-Mart is saying it expects its suppliers to attach tags to only 65 percent of their products (on average). However, several suppliers have told CIO magazine that the percentage of their products tagged would be much less than 65 percent - somewhere on the order of 10 percent to 15 percent.
"Many of these consumer packaged goods companies are really struggling with the business case," says Christine Overby, an RFID analyst at Forrester Research. "These are very costly projects, and they're hard to do with a technology that's a moving target."
Patrick Sweeney, CEO of ODIN Technologies, an RFID infrastructure software and integration company that works with several of the top 100 suppliers, says Wal-Mart's suppliers are split into two camps: About 30 percent of them are going the whole nine yards and integrating RFID into their infrastructures now. The rest are emulating the supply chain head quoted above and practicing a method known as slap and ship.
Essentially, what these suppliers will be doing on January 1 is sticking an RFID tag on only a certain percentage of cases and pallets in warehouses that are closest to Wal-Mart's Texas distribution centres. Slap and ship involves minimal data integration and leaves the retail supply chain still blind to product movement. And it will apply to only a small percentage of the products shipped to Texas.
Not surprisingly, slap and ship is not a method endorsed by Wal-Mart. "It's something we sort of cringe at," says Simon Langford, Wal-Mart's manager of RFID strategy.
Langford won't say as much, but come January 1, Wal-Mart could use some positive publicity about its RFID program. The reputation of the world's biggest retailer has been tarnished of late with allegations of unfair wage practices, hiring illegal immigrants and discriminating against female employees. And now some industry experts are predicting that suppliers' failure to meet the RFID mandate could be more bad press for the retail chain. Wal-Mart's biggest mistake, they say, was imposing a top-down mandate on its suppliers before the technology and business needs matured to where RFID-tagged inventory made good economic sense for suppliers, customers and Wal-Mart.
"These suppliers are being forced to implement the technology in a way that may not suit their business," says Kara Romanow, an RFID analyst at AMR Research. "The cost model just doesn't work right now."
What Wal-Mart Wants
In 1999, MIT founded the Auto-ID Centre to look more closely at how RFID technology could help businesses track and manage products using embedded sensors. The centre proposed an electronic product code, or EPC, as the latest method for identifying products. The EPC would utilize radio frequencies to identify computer chips placed in tags. This technology would eventually replace bar codes, which require the scanner to "see" the UPC number to read it, with a device that requires no line of sight and little human intervention.
When applied in a controlled environment, RFID tags and the devices that "read" them work amazingly well. Essentially, the system is made up of two components. First, there's the tag, which varies in size and shape. Some look like stickers and labels, and others like thin plastic wristbands. Tags are affixed to cases and pallets. Each tag has an antenna and is embedded with a chip that contains a unique string of numbers that identifies each product. Tags can be passive or active. Active tags have a battery; passive tags, which get their energy from the antennal devices (known as readers), are less expensive and more common.
Readers identify the tags as they pass by. The magnetic field from the antenna wakes up the passive tags as the reader approaches, and the tags transmit their digital information, in the form of the electronic product code, from the reader into a computer system.
This essentially is the system that Wal-Mart and others, including the US Department of Defence and Target Stores, believe will revolutionize the global supply chain. And indeed most supply chain experts predict that RFIDs will eventually take the black holes out of the retail supply chain, providing product shipment and inventory views of unprecedented detail.
"Today's supply chain allows products to be misplaced, misdirected and lost," says Paul Fox, director of external relations for Gillette, which, along with Procter & Gamble, was one of the early backers of RFIDs as a member of the MIT Auto-ID Centre. "There are countless millions of dollars tied up in warehousing because of the inefficiencies in the supply chain." Fox believes RFID technology will tell Gillette "where the product is in our warehouses, what the product is and how much of it we have. No manual counting, no driving around, no question of mispicks, no order number mistakes. Once you get an accurate understanding of inventory position, that information becomes invaluable".
For Wal-Mart, the business case for RFID is about maintaining low prices for its customers. "[RFID] will help us increase customer satisfaction in the near term, and ultimately play an important role in helping us control costs and continue offering low prices," Wal-Mart executive vice president and CIO Linda Dillman said in a statement released on April 30, 2004, the day that Wal-Mart began its RFID pilot. The eight suppliers that started shipping a handful of RFID-enabled pallets were Gillette, Hewlett-Packard, Johnson & Johnson, Kimberly-Clark, Kraft Foods, Nestle Purina PetCare, Procter & Gamble and Unilever.
"You have to applaud Wal-Mart for getting this started," says Simon Ellis, supply chain futurist for Unilever, which makes health and beauty products. "If someone hadn't stepped forward, we would all still be waiting around."
Langford says the RFID payback for Wal-Mart's suppliers will be twofold: It will allow them to reduce their inventory, and because Wal-Mart will always have their product in stock, sales will improve. "We can handle their product much more efficiently in terms of getting it out to the shelf," he says. "It makes a huge difference."
When Wal-Mart Wants It
Wal-Mart's announcement in June 2003 of the January 2005 deadline caught many in the supplier community off guard. Some didn't even know what RFID stood for; others dismissed the announcement as grandstanding.
So when Wal-Mart invited the top 100 suppliers to its Arkansas headquarters in November 2003 for more specifics about its RFID master plan, there was speculation among suppliers and business journalists that Wal-Mart would back off the deadline.
That didn't happen.
During the presentation, Wal-Mart executives defined what the EPC would be, what class of chips they would accept and which three distribution centres would start accepting RFID deliveries. The suppliers, according to one participant, listened in stunned silence. As he recalled, it was like attending a session on nuclear physics.
Yet there were few questions, and "they were just for clarification", says the supply chain executive, who attended the session. "There was no pushback."
Wal-Mart then reaffirmed its deadline and announced it would not give any price breaks to suppliers who complied. "And no one raised their hand and said: Excuse me?" recalls the supply chain executive.
It is now a year later, and his ears are still ringing. He's frustrated by the immature technology, the nonexistent standards, the approaching deadline. "There's just no benefit right now," he says. "So I'm being supportive of something I don't really want to do."
How Is a TV Different From a Razor?
From the beginning, the price of the EPC tag has been the ultimate deal-killer for RFID in the retail world. Before the turn of the millennium, tags cost anywhere from $1 to $2 each (all amounts are in $US). Then they broke the $1 barrier - running suppliers anywhere from 25 to 75 cents, depending, of course, on the volume of the purchase.
But suppliers say the price must roll back even further. RFIDs are still too expensive to make much economic sense, says AMR's Romanow and others.
At this point, RFID-enabled cases can cost a supplier approximately 50 cents more per case. (According to some suppliers, about 20 to 30 cents of that is the cost of the tag, and another 20 to 30 cents is the labour costs to affix the tags.) Forrester's Overby estimates, using a fictional $12 billion supplier that ships 15.6 million cases and pallets to Wal-Mart per year, the supplier would spend an extra $7.6 million in tag costs. (Her example assumes a 40 cent tag price.)
Beyond the price of the technology, another problem is Wal-Mart's one-size-fits-all plan, say many suppliers and industry analysts. In Wal-Mart's world, there's no difference between a Gillette Venus razor, an HP printer and a 12-pack of Coke. All of the cases on each pallet have to have an RFID tag when shipped. Unfortunately, like the different sizes of clothing on Wal-Mart's sales racks, there are as many small and medium suppliers with low-cost items as there are large and XLs with high-value merchandise. Romanow calls it the "toilet paper and toothpaste" problem. "If you look at TVs, DVDs and video games, the price of the tags doesn't matter," she says. "But when you're talking about TP and toothpaste, there is no tag cost at the case and pallet level that makes the numbers work."
What's the Frequency?
There's another problem with RFIDs: No one standard for the technology currently exists (see "In Search of a Standard" below). And nowhere is the lack of a global standard more vexing than with the most basic elements of RFID. Not all tags and readers can communicate with each other. So Wal-Mart, for example, might need to have two (or more) types of readers in its warehouses to read different tags from hardware vendors such as Matrics or Alien Technologies, to name just two. Kevin Ashton, cofounder of the MIT Auto-ID Centre and vice president of marketing for reader-maker ThingMagic, explains it by looking at televisions.
"Whether you buy a Sony or Panasonic or whatever, you know that TV will work because of the National Television System Committee broadcast standard," he says. But that's not the case with RFID equipment; the standard just doesn't exist, and therefore, there's no universal interoperability among hardware equipment.
The radio waves that underlie the technology have also caused problems in several pilots. ODIN Technologies' Sweeney, who's the author of the upcoming [ital]RFID for Dummies[end], recalls one "bake off" between competing RFID technology providers (whoever got the best read-rates would get the supplier's business). One provider wasn't getting a good rate so their engineer kept cranking up the power, and adding more antennas and an additional reader to the loading dock door. The read-rates never got above 50 percent.
"People's inclination is to add more antennas and more power. In most cases, you just make it worse," he says. "It's no wonder they couldn't get it to work, with one reader drowning out the other reader." And this was a company with experience in the RFID space.
One important property of radio frequency is that it tends to act abnormally when it's near certain elements - liquids, metals, porous objects. For Mike O'Shea, director of RFID strategies at Kimberly-Clark, this has been a problem with the company's baby wipes. "The baby wipes absorb RF signals," he says. "Our packaging engineers have worked on antenna designs, looking at where you place the tags and chips on products to get the best read-rates."
Unilever's Ellis has had similar problems. "The liquids and moisture issue is more difficult because the liquid is inherent in our product," he explains.
Ian Robertson, director of the RFID program office at Hewlett-Packard, and his team have done a lot of experimenting with tags to determine how to best tag (among other items) HP's ink-jet cartridges, which contain ink and metal, and are wrapped in metallized paper. Robertson's work on the ink-jet packaging has been a showcase of success for the Wal-Mart pilot, but that's probably due to the fact that HP is also a solutions provider to other suppliers struggling with RFID and has been working on the area since 2002.
Suppliers in Pain
For the anonymous executive in charge of supply chain operations, it's been a nightmare trying to get to compliance. For one, the "squiggle" RFID tags, which are the mostly widely available and cheapest, are failing because some of his products skew the radio frequency signal. So he's having to use a more expensive tag.
The next generation of tags, which he's hoping will help solve some of these problems, most likely won't be available until the middle of 2005. And it's anyone's guess how much those new tags will cost.
As to the status of his ongoing warehouse RFID pilot, which should be cranking right now: "Frankly, we've not even opened up the warehouse door yet," he says.
The supply chain executive thinks the technology just isn't there. "I don't think RFID is a mature application at this point and time," he says. "I think it's probably two years premature."
The supply chain manager at another top 100 supplier, who also insisted on anonymity, agrees. "Our experiences in our pilot have shown that this is not ready for prime time," he says. "The tag and reader performance problems are a long way from being solved."
Suppliers must consider some unpleasant alternatives. If they hang back and wait for RFID to develop, thereby lowering some of the costs of entry and the expense of retrofitting their operations, they fear Wal-Mart might get mad, and they may also lose ground to competitors. So their approach has been to comply with the mandate via slap and ship. Many suppliers are applying the tag on pallets to be shipped to warehouses in the region where the Wal-Mart distribution centre is located. Using the product that is destined to be shipped the next day (which is already palletized and stretch-wrapped), they are taking off the stretch-wrap, depalletizing, applying the tag, writing the tag, repalletizing, rewrapping and sending. And even then, suppliers are afraid the tags might not function on arrival because of mechanical problems.
What this kind of wasteful effort reflects, the supply chain executive concludes, is a "premature, underdeveloped industry trying to support a major launch".
The other supply chain manager interviewed for this article also has concerns with the tags. His company is trying all kinds of tags to work with his products. He's found that many tags are poor quality, and when they are delivered (usually in rolls), it's not uncommon to find that up to 30 percent are unusable. "We don't pay for them, but in terms of process efficiency . . . ," he pauses and sighs. He also has concerns about the manual tagging of cases and pallets, citing the high cost and inaccuracy.
These suppliers haven't even begun to grapple with the issue of how to integrate the new RFID data with the legacy data. "RFID demands the destruction of silo-based corporate procedures and processes," says John Greaves, head of the global RFID technology integration group at Deloitte. "You are divorcing yourself from historical legacy systems."
Wal-Mart's Langford says he's aware of supplier pain. "We obviously realize that for some suppliers, the current cost of the technology and the tag will be very difficult and cost-prohibitive to tag 100 percent," he says. Wal-Mart, he says, will work with the suppliers to get them up to speed.
Langford concedes that not every supplier will make the January deadline. "We've already worked with a couple of suppliers in the top 100 that won't be live in January, [to create] some alternative plans for them," Langford says. "We will work with each supplier individually and look at the reasons why [they can't make the deadline]."
Mission Impossible?
Though Wal-Mart's January deadline has never wavered, many analysts and some suppliers whisper that Bentonville's expectations for compliance have softened. Wal-Mart itself acknowledges that it will settle for less than 100 percent compliance on products shipped to its Texas distribution centres. Langford says he expects that on average, 65 percent of cases and pallets to Texas will be tagged by January 1. However, suppliers claim that Wal-Mart is privately negotiating even lower shipment percentages with suppliers that say they won't be able to meet the original mandate. And several suppliers reported that those negotiated percentages are far below 65 percent.
But what if many suppliers still can't meet Wal-Mart's lowered expectations? Will their failure turn into more bad press for Wal-Mart? A lot of that will depend on how the retailer deals with those suppliers that are struggling to comply. Will there be monetary penalties for delinquent suppliers, as many industry observers have hinted? Langford will not even speak of such penalties. "We will take each case on merit and discuss it with that supplier," he says.
Regardless of how Wal-Mart plays it, it's clear that many suppliers will do just enough to stay in Bentonville's good graces - and no more. "I can imagine a supply chain that is enormously efficient with RFID," says the supply chain manager. "My issue is that the capabilities required to have a successful RFID initiative don't exist yet."
In Search Of A StandardWal-Mart suppliers are still waiting for the next generation of standardized tags Mike O'Shea, Kimberly-Clark's director of RFID strategies, and others with worldwide supply chain operations are desperate for a global RFID standard. Due to various radio frequency spectrum issues and bandwidth constraints, the wait is still on for a commercially available RFID tag that will work across continents. "We're all looking forward to a global tag," O'Shea says. "There's one for the United States, but it's different in Europe, and it's different in Asia too." The next generation of tags, called Gen 2, is supposedly going to unite worldwide RFID operations. Right now, Wal-Mart has said it will accept EPC Version 1, Class 0 and Class 1 tags from its suppliers on the January 1, 2005, deadline for the first phase of RFID implementation. EPCglobal is the organization entrusted with making sure that electronic product code (EPC) eventually becomes the global RFID standard. It is a joint venture of the Uniform Code Council and EAN International, which are the standards administrators for bar codes. EPCglobal has more than 200 member companies, including most of Wal-Mart's top 100 suppliers. There are also industry-specific action groups, made up of representatives from those 200 or so members, that work on making sure the technical specifications jibe with the retail industry's real-world needs. Gen 2, which everyone is buzzing about, isn't scheduled to be ratified by the EPC board of governors until the fourth quarter of 2004. "It's incredibly difficult to predict," says Kevin Ashton, who was cofounder of MIT's Auto-ID Centre and is now vice president of marketing for ThingMagic, a maker of RFID readers. "It's a gruesome political game," largely because each vendor has its own standard and wants that standard to be the global rule of thumb, he says. If and when one standard is adopted, it's anyone's guess when commercial products will be available, says Ashton. "Some say commercially available in the first quarter of 2005. Others claim in the first quarter of 2006. We'll all just have to wait and see," he adds. Thomas Wailgum |
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.