Any new technology introduces both costs and benefits, and RFID is no exception. Let's begin with the challenges it presents:
- Cost–The most-discussed cost element for RFID systems is the cost of individual tags. However, this is just a part of the overall cost. Successful adoption of RFID will require changes to business processes and information systems, personnel training, and, in some cases, customer education.
- Accuracy–Many of today's RFID systems are far from robust. Most of the projects we surveyed in early 2005 reported a read accuracy rate of between 80 and 99 percent. What this means is that if we move a pallet carrying 100 tagged boxes past an RFID reader, the reader will fail to recognize anywhere from 1 to 20 tags. There are many reasons for this, but most stem from the inherent challenges in moving liquids and metals using RF communications.
- Implementation–Introducing RFID will invariably change your business processes, from how items are labeled to how they are selected, palletized, cycle counted, and so on. For each step in the business process, you will need to incorporate capabilities for exception processing, that is, what to do if the RFID technologies are not working properly. For example, if an RFID reader stops working or a label falls off, how will you continue? Apart from the RF side, there are many challenges to overcome in integrating the RFID data and procedures into your existing systems and the business processes they support.
RFID Adoption Guidelines
With things changing so quickly, it can be difficult for organizations to decide where, when, and how much to commit to any particular RFID product or standard. The following guidelines offer a strategy for approaching RFID. Simple as the steps may seem, ignoring any one of them can lead to lost opportunities at best, and failure at worst.
Determine the business need
The current process exists for a reason. Before you even think about changing it, make sure you know its strengths, weaknesses, and reasons for being. Don't start from the premise, “Where can I use RFID?” Instead, start by asking, “How can I improve this process?”
Evaluate potential changes
Carefully assess the costs and benefits of any potential changes. For example, if you want to automate a manual process, ask yourself a few questions about the change. Would a bar code work better than the manual procedure? Could the process be changed in some way to take more advantage of RFID, or conversely to eliminate it from your process? How will you handle equipment failures or other types of failure?
Note that RFID is sometimes mistakenly referred to as a “replacement for bar codes.” Bar codes are actually better than RFID for some applications, and they're considerably less expensive. Also, with RFID there is a temptation to develop automated systems with minimal human oversight. But what happens if someone puts a box full of RFID tags on a pallet and accidentally runs it through the reader? Will your systems believe you've shipped 100 pallets? With bar codes, someone would probably scan each pallet manually, which would prevent this error.
Develop a long-term roadmap
Instead of implementing RFID systems in an ad hoc manner, develop a long-term business justification for adopting RFID and formulate a vision of how your business processes will look in an RFID world. Follow this up by developing a master plan that shows which systems will need to change and how. Show what your application and infrastructure architecture will look like after deploying RFID. Doing the design up front will give you a clear goal and will also promote the necessary discussions between you and your business units, end users, operations staff, IT staff, and business partners. Don't underestimate how much RFID will impact these stakeholders.
Develop a proof of concept (prototype) to validate your assumptions. It's better to fail small and learn early than to fail large and have to recover. Don't be afraid to revise the roadmap and architecture based on mistakes in your prototype, and be ready to start the cycle over again, learning more each time and managing risk as you progress.
Run in parallel with existing systems
Take a lesson from mountain climbers–don't let go of the last toehold until you are sure the new one won't crumble away beneath you. Not only is it more responsible to run the new system in parallel with the old system for a while, but it can also lead to valuable insights that were invisible when looking at either system in isolation. Don't forget to test the workarounds and recovery strategies you identified earlier. You might be surprised to discover that all of the RFID readers in the building fail when you turn on the exhaust fan for the first time. Only when everything runs smoothly should you begin to depend on the new technology.
Now that you've made these changes, you have a brand new process. The old process took time to develop, and the new one will take time to mature. While this is happening, be ready to take advantage of new capabilities. Readers with new features and smarter controllers will come out; meanwhile, your personnel should be looking for innovative ways to make use of the new equipment. Watching carefully how people use a piece of equipment can provide important clues for streamlining the process.
Share with partners.
You have a great system in place, but your suppliers are still sending signals by carrier pigeon. Work with your less-enlightened trading partners and show them how to improve their own processes. RFID is an evolving technology so taking a leadership role will allow you to define the agenda and the standards for future integration. Wal-Mart is an example of a company that has approached a potentially disruptive technology by choosing to lead in its development. As Dr. Alan Kay said, “The best way to predict the future is to invent it.” If you keep it all to yourself, you'll just have to change your system when one of your partners chooses a completely different approach.
In this chapter, we discussed the following:
- RFID is a technology that allows a small radio device attached to an item to carry an identity for that item
- RFID has been around a while–long enough that we can divide its history into eras and begin to predict future trends.
- In the current era (Compliance), cheap semiconductors and fast Internet connections have encouraged retailers and governmental agencies to require suppliers to place RFID tags on shipping units such as pallets and cartons. However, most suppliers are just tagging pallets and shipping them without using the information internally, and even retailers are simply breaking down the pallets on receipt.
- As the components get cheaper and the information infrastructure becomes more defined and robust, RFID will be used for an increasingly broad array of tasks.
- There are five main categories of RFID applications. Knowing the type of application in question can tell us quite a bit about special considerations and implementation.
- This is a volatile time for RFID, so we must take a disciplined approach to both acquiring knowledge about it and adopting the technology within our organizations. Keep an eye on key players and standards as we move through this era to see which way the technology will shift.
About the Authors
Bill Glover makes his living explaining simple concepts to complex people who tend to work for large companies. Bill believes that anything you do three times should be scripted, and that if a piece of code is 100 lines long it can probably be done better in 10. Bill is a Pisces who enjoys long walks in the hills, full personal state vector uploads and chatting about completely harmless things over strong, public key encryption. He can be reached at: firstname.lastname@example.org.
Himanshu Bhatt heads Sun Microsystem's US RFID & Sensor Solutions Practice. His responsibilities include developing go-to-market strategies and execution plans, developing partnerships with select ISVs, IHVs, and Systems Integrators, people management and all aspects of RFID related sales and delivery for the US market. As the head of the US RFID efforts for Sun, Himanshu developed strong working relationship with product engineering, marketing and industry partners. Prior to leading the charge for the RFID practice, Himanshu has managed the technical sales and consulting practices for Java and related Sun Software products. Himanshu started his career as an enterprise architect and has published papers and spoken at industry conferences on various areas of J2EE and enterprise software architecture.