Leaving aside the easy predictions (Wal-Mart and the Department of Defense will continue to deploy, tags and equipment will fall in price), I thought it might be interesting to translate a few shifts in customer buying habits into trends we can expect in the coming year.
Item-level RFID Concedes Some Ground to 2D Barcoding
We have seen three compelling situations in which 2D barcoding is being adopted in favor of item-level RFID. The first is for situations in which the product in question is too inexpensive to bear the burden of a 10 tag, yet can still have high consequences. These are typically consumer items (such as some drugs) or maintenance items for critical equipment that benefits from a high degree of traceability. The second situation occurs when the item  such as a very small medical device  is in a package that is too small to provide enough free surface area for an RFID tag.
The third situation can occur when items in cases need to be individually numbered but where it is sufficient to use inference (the assumption that you can tell what is in a sealed case based on prior knowledge of its contents and its security). RFID is still used to track each case, but 2D barcodes are used on each item. Of course, companies should be very careful to document when the observations that they log in the tracking and pedigree of such goods are from direct observation versus when they have just been inferred.
The Best Deployments Will Be About Profit, Not Compliance
Certainly complying with customer (read: Wal-Mart, DoD) or regulatory (FDA) mandates qualifies as a compelling reason to become involved with RFID technology, but it doesn't generally translate into profits. These situations haven't produced the kinds of ROIs that inspire non-mandated companies  or even non-mandated divisions of mandated companies  to embrace RFID as a core technology.
We have seen some great business cases for RFID deployment, however. These projects share two characteristics:
>> They are closed loop, meaning that the company doing the tagging is doing so for its own internal benefit
>> The project was undertaken to solve a very clearly defined operational issue or enhanced service offering
For instance, we have seen both distributors and retailers applying their own RFID tags and solving their own problems without waiting for help or investment from a manufacturer or trading partner. Two examples illustrate this nicely. The first is ASD Healthcare, the specialty drug distribution business of $61 billion giant AmerisourceBergen. Within the last year, ASD has begun offering a service that uses RFID to monitor inventory levels within its customers hospitals and to replenish them automatically. Offered to each customer at no up-front charge, the simple drop-in refrigeration unit sends immediate consumption signals back to ASD fulfillment, allowing ASD to offer a level of service that its competitors have yet to match. The ability to promise zero stock outs, pay upon use and an elimination of expired drug disposal write-offs is a compelling customer offer. The service isalready operational in over 50 hospitals and is expanding rapidly.
Another example involves a major retailer using RFID at its distribution centers  and soon its stores as well. Although it is known as one of the most efficient operators in the business (as well as one of the largest), it has begun investing in RFID as a core part of its operational systems in the distribution centers. The application? Higher shipment accuracy to stores. Although the pick and pack part of the distribution centers has become highly automated and accurate, manual shipping and handling between the DC and the stores have now been highlighted as the point of improvement in their supply chain. Arguments still regularly ensue between the stores and the DCs about missing or misplaced shipments.
Far, far less than one percent of products sold by this chain are RFID-tagged. What is being tagged, however, are the plastic totes that carry items to the stores. The totes and their tags re-circulate through the system, so the tag cost per item is essentially zero. The company has begun by RFID-enabling every dock door at a given distribution center so that workers can be warned within two seconds (a service level guarantee) if any totes are being placed on the wrong truck.
The next phase of the project will be to begin equipping stores with RFID sensors as well to detect unloading mistakes. Quite soon, the company will be detecting and correcting delivery mistakes as they occur and have an end-to-end record of what actually happened  versus what was supposed to have happened. It will also eliminate all manual shipping and receiving documentation steps at each end. It's a better, cheaper, more accurate process.
Relying on, Not Trying, RFID in 2008
Pay attention when you hear or read about RFID deployments to see if you can figure out whether or not RFID has truly become a core part of the company's business processes. In the two examples listed above, there is no other system also collecting the same information as the RFID system. In other words, if the RFID system fails, the operations team becomes blind to product movement.
This has some sturdy implications. The entire RFID system (tags, readers, event processing, sensor coordination, networking, systems integration) will have to meet stringent service level guarantees. Such guarantees will become more commonplace in 2008 as customers demand uptime, accuracy and response time guarantees from their technology partners. Systems will be expected to monitor their own health and to be easily administered. These are no different than the kinds of requirements that other enterprise technology providers have had to provide for years, and RFID vendors will need to evolve to become more service-oriented and just as collaborative and responsive as more established automation providers.
Another implication is that RFID needs to be well-integrated with operational procedures and existing core software systems, such as ERP and WMS. When RFID is used as a core system, its responsibilities should not just include the generation of a standalone report or a set of isolated database records, but instead should include:
>> Real-time operational alerts and instructions to workers and machines
>> Transactions to other official software systems of record, such as SAP
This will require that companies that purchase RFID systems think of their RFID vendors as true partners, looking beyond simple price comparisons to issues of architecture, reliability, system compatibility, track record and even cultural fit. A lot is at stake, much more than when companies began setting up labs and running trials in 2005 and 2006.
The Hardest Part Is Business Process Change
Finally, an important litmus test of the success (or failure) of RFID adoption in any business will be the extent to which business processes have changed along with the introduction of RFID. It's easier to install equipment than it is to plan, train and execute business process change, so the rate at which companies can execute these changes will be the dominant dynamic of RFID adoption in 2008.
Although it's a bit of a gross simplification, I believe that the positive impact of RFID must always happen simultaneously with business process change and the streamlining or elimination of activities and costs  or an improvement in a company's service offerings. If the operations stay the same, and the customer can't tell the difference, where's the benefit?
John Beans is the Vice President of Marketing for Blue Vector Systems, a company that has developed a highly distributed automation platform that integrates a variety of RFID, barcode, temperature and other industrial sensors with back end enterprise applications. John's experience spans manufacturing systems at Datasweep and Camstar, networking at SBC Communications and strategy consulting at Bain & Company. He can be contacted via email at john@bluevector.com.