May 12, 2003 - As with any emerging technology, there is a lot of confusion about what RFID can and can't do. Many articles in the press either simplify or exaggerate what's possible. While the technology truly has the potential to deliver dramatic benefits to companies in many industries, it's time for a reality check.
We've chosen the Top 10 Myths based on the large number of email messages submitted to RFID Journal
and on questions that are put to vendors and systems integrators almost every day. The aim of this article is not discourage the use of RFID, but rather to debunk some of the nonsense and give companies a more realistic view of what it takes to deploy RFID the technology successfully.
Myth No. 1: RFID is a "talking" bar code.
RFID tags are like bar codes that broadcast a serial number, but using them in this way misses out on many of the potential benefits. Even the simplest RFID tag -- those that only contain a unique serial number -- can carry more data than most bar codes do today. Instead of just identifying the manufacturer and the stock-keeping unit (SKU), they can also identify unique items. That means you can tell which pallets arrived first, which contain products approaching their sell-by dates and so on.
Many tags can store much more than a serial number. Write-once read many (WORM) tags typically store a couple of kilobits of data. Tags on tools can store information about when the tool was purchased, which department it belongs to, where it's supposed to be stored and so on. And read-write tags can be used like a distributed database, where information is updated regularly.
Scottish Courage uses read-write tags to track when its beer kegs need to be washed and when they should be taken out of circulation for servicing. When the keg is serviced, the date is written to the tag and the maintenance people can retrieve it at any time (see TrenStar: RFID with Less Risk
). And Marks & Spencer writes information about specific food shipments -- when the food was picked up, where it needs to be shipped to and so on -- on 3.5 million tagged containers.
Myth No. 2: You can read every RFID tag every time.
It would be wonderful if that were the case, but unfortunately radio waves are subject to certain immutable laws of physics. They bounce off metal and are absorbed by carbon fibers and by water at certain frequencies. Therefore, it's unlikely that you will drive a forklift truck through a dock door and have a portal reader read 100 tags on 100 cases of goods containing metal or water, let alone every single item in those cases.
Readers on forklifts can improve the accuracy of inventory data
The position of the tag vis-a-vis the reader is also an issue. If the tag is positioned perpendicular to the reader, the tag won't be read. If two metal antennas touch one another, both tags will be shorted out. There are nulls, or voids, in UHF fields that can prevent a tag from being read. And sometimes the tag isn't read for unknown reasons.
Good engineering and system design can get around a lot of these problems, but companies may have to alter procedures if reading every tag every time is critical for a particular application. For instance, boxes of items may have to be tagged in places where there are air gaps, and the tags may have to be arranged on a pallet in specific ways to ensure the tags can be read. This can offset some of the labor savings that RFID is meant to deliver.
Myth No. 3: You can take inventory with the push of a button.
Vendors and systems integrators get asked about this all the time. Some companies believe that you can either point a handheld RFID reader around the warehouse and get a complete, accurate inventory count, or they believe you can install shelf readers to do it. The read range of a handheld is going to be limited to a few feet at UHF and a few inches for 13.56 MHz. You could put a reader on every shelf in a warehouse, but at today's prices, that's would be extremely expensive. And mounting a reader on a forklift and driving it around the warehouse isn't going be accurate or cost-effective.
Nevertheless, it is possible to create real-time inventory by tracking everything that comes into a warehouse, where it's placed, and when it leaves. You don't need expensive shelf readers to do this. You can use readers on forklifts and tags in the floor and on shelf racks to track the location of goods (see RFID Speeds P&amp;amp;amp;amp;G Plant Throughput
). The key is to design these systems in such a way that they capture accurate information and to train staff to follow new procedures that ensure the accuracy of the information captured (see Myth No. 4).
Myth No. 4: RFID delivers perfect information.
The goal of any RFID system is to get 100 percent accurate information. But just putting in an RFID system doesn’t guarantee that. For one thing, as we mentioned above, tags are not read every time. Moreover, tags can be read when they shouldn't be read. For instance, a tag might be on the edge of the read field and be picked up when you don’t want it to be. RF waves can bounce off of metal objects in the read field, which can extend the read range beyond what you want.
Human error can also cause problems. In one access control implementation, a smart card kept popping up in the read field. The customer couldn't figure out what was going on. The systems integrator showed up and discovered that the users had created a dummy card to test the system. They hung the card on a key chain holding a key used to turn on the system, so the card was always in the read field.
Companies need to work with a good integrator to design systems that are highly accurate. Photo eyes need to be used to trigger readers to start reading as a pallet is moved or items approach on a conveyor. Indicator lights need to be set up to alert staff when a box, carton or container isn't properly scanned. Software has to be programmed to eliminate double scans of items. Mobile computers have to be designed to give simple, clear instructions to forklift operators and to alert them to mistakes, so the right items are always moved to the right place. And employees need to be trained to follow new procedures. If all these things are done, it's possible to gather near-perfect data.
Myth No. 5: You can buy tags for five cents.
There's been a lot of confusion about the Auto-ID Center's
goal of creating a five-cent tag. That price was always contingent on the manufacturing of billions of tags. You can't purchase a tag for 5 cents today. The cheapest price we've heard of is "under 10 cents," which is what Gillette says it paid for the tags it bought from Alien Technology
. But Gillette had to purchase half a billion tags to get that price. Alien is selling tags in small quantities for about 45 cents.
Alien tag on Gillette razors -- 5 cents is still the goal
Will there ever be a five-cent tag? We believe that one day you will be able to buy a simple RFID tag carrying an Electronic Product Code for that much. In the weeks ahead we will be working with independent experts and data to project when that might be possible. Low-cost tags are needed if companies are going to be able to track individual items that cost only a few dollars.
But keep in mind that for many applications today, companies don't need a low-cost tag to get a return on investment (see Myth No. 6). And even when low-cost tags are available, many applications will require a read-write tag that will cost more than 5 cents. For instance, the US military won't use low-cost tags because enemies could scan the tags and potentially learn about supplies, troop strength and so on. And if you have reusable containers, you will likely want a tag that can be updated as the container is emptied and refilled. So you may need to use simple tags to track unique items and more expensive tags to track some assets.
Myth No. 6: The tag cost is critical to making any deployment pay off.
The tag cost is only one part of the complex return on investment analysis you need to do. For many projects, the number of tags used will be small, and the cost of the readers will be more significant. For large projects, you may need middleware to manage all the data from the readers, and there will likely be significant integration costs as you create links to back-office systems. We know of one company that spent $5 million on RFID equipment for a major pilot and then spent $15 million to support it. Technical people were constantly being flown in when the system went down, and the system kept going down because the software and integration wasn't handled properly.
Many projects won't be viable unless the cost of the tag is 50 cents or less. Other projects can support a tag cost of more than $10. Railroads, for instance, got a return on their investment when they spent $25 per tag and $40,000 for each reader installed along the tracks. By tracking railcars with this expensive RFID system, Burlington Northern Santa Fe Railroad eliminated 500 positions that previously were required when people walked up and down the tracks and manually recorded numbers painted on the sides of the cars.
Many applications will support a $3 read-write tag with 2 KB of memory and protective housing. For instance, if you have a reusable container that will last 10 years and is turned six times a year, then the cost of the tag is 5 cents for each use. In many cases, the technology will save you far more than that. These kinds of tags can be used to track tools or other assets that are lost or stolen. Companies that are not examining the potential of RFID because the tags don't cost five cents today may be missing out on a potentially large opportunity to cut costs and improve efficiency.
Myth No 7: RFID is primarily a supply chain technology.
Perhaps the biggest myth of them all. There is a great deal of focus on the supply chain because many companies are excited by RFID's potential to enable them to slash inventories while simultaneously reducing out of stocks. The supply chain is certainly a core focus of RFID Journal.
But saying RFID is a supply chain technology is like saying the Internet is just another sales channel. Both technologies can provide many benefits to companies.
RFID is already being used by many manufacturing companies to ensure that parts get to the assembly line on time. Ford, Mercedes Benz, and other auto manufacturers tag chassis as they go up the assembly line, so that workers can identify a specific car and put the right parts on. RFID is also widely used for access control and payment systems, such as smart cards, quick service payments like ExxonMobil's Speedpass and toll collection.
As tags and readers become widely deployed, RFID should be looked at as an opportunity to add value, as opposed to a cost incurred for tracking items. For instance, office and industrial equipment manufacturers might be able to put readers in equipment so that it can be monitored remotely. Retailers might use tags with temperature loggers to guarantee the freshness of their produce, and toy companies might be able to add interactivity to games.
Myth No. 8: RFID spells the end of privacy.
RFID is potentially an invasive technology that could be used to gather information on people. But companies are not going to make the data on their tags open to everyone. Some companies will use proprietary numbering systems to protect critical information. Others will encrypt data on the tag so it can't be read by anyone without authorization. RFID is still an emerging technology, and as it evolves, systems will also evolve to protect privacy.
Wal-Mart isn't waiting to roll out RFID in its distribution centers
will soon run two features looking at how companies should handle the consumer privacy issue and how they can protect themselves from having competitors gather data by reading tags on goods in the supply chain.
Myth No. 9: RFID is easy to deploy.
By now, it should be obvious that this is a myth. It's true that setting up a very simple demo system to read tags in a field is easy. But deploying a system in a distribution center, warehouse, retail store or manufacturing facility can be a challenge. You'll have to deal with some unique characteristics of the environment, such as whether there is a lot of water around, metal shelving or electromagnetic interference from motors or robots. You'll have to come up with strategies for tagging packages containing metal and water, for avoiding interference with other RF devices used in the same facility and for dealing with redundant reads or misreads.
Deploying a system across all of a company's facilities is a major project that we’ve compared to deploying enterprise resource planning software. There is complex integration work that needs to be done. Companies will need to set up systems for managing readers, so they can be maintained and upgraded remotely. And companies will need to change their IT systems and their business processes to take advantage of the data they gather. A full-blown deployment is anything but easy and needs to be done in stages.
Myth No. 10: I can afford to wait.
We're not big believers in "first mover advantage," at least not in the case of RFID. It's better to get a deployment right than to be first. But we strongly believe that those who start early have a better chance of getting it right. That's because they can proceed slowly, learn step by step and not have to try to roll something out all at once to catch up to a competitor.
Some surveys show that retailers are planning to deploy RFID technology in the supply chain more quickly than manufacturers. Wal-Mart, for instance, has said that it will begin tracking pallets and cases by 2005. That means that suppliers who don't start experimenting now will be forced to spend money to tag goods for Wal-Mart's benefit, not their own. But if you start deploying now and figure out how to use RFID internally, then you'll be ready to tag items for your retail customers, and you'll also benefit from the investment.