Evan Schuman's StorefrontBacktalk

An item-level RFID pilot done with Bloomingdale’s and the University of Arkansas delivered more statistics supporting what is already widely believed, that item-level RFID should improve inventory accuracy. But it also found a 96 percent improvement in inventory count speed, which suggests the potential for even more dramatic accuracy improvements.

In the trial—conducted over 13 weeks and published Monday (Aug. 24)—looked at a non-item-level store (the control store) and one item-level store and the same limited product area (denim jeans) were examined for both.

“The test store was equipped with static readers at all employee and customer entrance/exit doors. Cycle counting (i.e., physical inventory counting) was conducted with handheld RFID readers (in the test store) and with barcode scanners (in both the test and control stores),” the report said. “There were between 9800 and 10,500 items included in this study. Items were RFID-tagged upon arrival at the store; thus, they could be read when moved from the receiving area to either a stock room or the sales floor. Tags were removed at point of sale and discarded. Returned merchandise was re-tagged by a store associate using a printer/encoder near the department.”

The study found that the adjusted overall inventory accuracy improved by more than 27 percent, with a corresponding decrease in understock of 21 percent. The overstock figure accounts for the remainder of the adjustment, decreasing by a little more than 6 percent, the report said.

Beyond inventory greater inventory accuracy, the Bloomingdale’s study also quantified how much faster inventory counts were with item-level RFID: 96 percent. The item-level scanning of more than 10,000 items “took two hours, far less than the 53 hours required with barcode scanning. On average, 209 items could be counted per hour via barcode whereas 4,767 items could be counted per hour via RFID.”

The study said that this suggests that actual field deployment could have even more dramatic inventory accuracy improvements, as the faster counts would likely enable far more frequent counts.

“Given the substantial reduction in time to cycle count, retailers such as Bloomingdale’s could create cycle counting strategies for taking and updating inventory counts more frequently than once or twice per year. With the above example, Bloomingdale’s could take inventory counts 26 times with an RFID handheld reader in the amount of time it takes to do one inventory count with a barcode scanner,” the report said. “Thus, they could take inventory counts every other week for an entire year (for a total of 26 cycle counts) in the same amount of time it takes them to do an enterprise-wide annual inventory count. Certainly, inventory accuracy is higher when taking and updating inventory counts bi-weekly than it is when taking inventory counts annually.”

The report also pointed out that given the limited nature of the trial, true battlefield condition results could deliver better efficiencies in other ways. For example, “Bloomingdale’s did not use RFID to directly affect store-level out of stocks; that is, changes were not made to supply chain replenishment practices and RFID was not used to generate replenishment orders.”

Editor’s Note: I don’t typically do with this comments, but there is a superb set of comments at over at RetailWire on this story.

The most intriguing comes from Nikki Baird from Retail Systems Research (formerly Forrester analyst): “I note that Bloomie’s carefully removes the tags at checkout, in a nod to consumer privacy. But if you’re going to go to all that trouble, why is there nothing designed into that test or that application of RFID that benefits the consumer? It would be so easy: there will be a day when someone makes it possible to find the right size of jeans in the right color without having to paw through every ever-lasting pair of jeans in the store.”

Or there’s this gem from QuantiSense’s Bill Robinson: “Frankly, it comes as no surprise that RFID brings more accuracy into inventory management in comparison to barcode. And it doesn’t seem headline worthy that counts would be quicker. The notion of counting RFID tagged merchandise sounds a little ridiculous. It reminds me of the early days of POS when retailers scanned the merchandise tags a second time to make sure the POS scanned them.” All in all, some time-worthy comments.