Aug 06, 2008This article was originally published by RFID Update.
August 6, 2008—Retailers can achieve very good RFID read rates for typical item-level tagging applications, but performance can vary widely depending on which tags are used, according to a new report released by researchers at the RFID Research Center at the University of Arkansas' Information Technology Research Institute (ITRI). The report, RFID Item-Level Tagging for Apparel/Footwear: Feasibility Study, presents the results of tests designed to identify baseline read rates for individually tagged footwear and apparel items in typical retail use cases, including point-of-sale (POS) checkout, inventory monitoring on racks and shelves, theft prevention and identifying items on moving conveyors.
Three types of Gen2 UHF tags, four models of handheld readers, and three stationary readers were used in the tests. Tag and reader vendors were not identified. Researchers attained 100 percent read rates in each test scenario with at least one of the tag-reader combinations, and many combinations produced read rates above 90 percent. There were some major differences in the read rates for different combinations in specific use cases. Researchers also found that read rates went down as the number of items to be identified increased.
"The major differences in read rates came more from different tag types, not different reader types," Dr. Bill Hardgrave, executive director of the research center told RFID Update. "We've always known there's not one tag that's the best fit for all products. If you had the right tag on the right product it performed well, regardless of the use case."
Each tag-reader combination was tested 30 times in each use-case environment. Researchers tried to mimic real-world conditions and did not try to enhance read performance by spacing items on racks and conveyors or otherwise adjusting tag locations, according to Hardgrave.
Overall read rates were lowest for identifying tagged items in boxes on a conveyor belt, which was tested at speeds of 200, 400 and 600 feet per minute. Read rates were worst for the fastest-moving conveyors, ranging from a low of 64.5 percent for one tag-reader combination to a high of 90.1. Read rates for the 200 foot-per-minute conveyor test ranged from 90.9 to 100 percent, and the 400 fpm rates ranged from 68.8 to 94.4 percent.
The point-of-sale environment produced the best read rates, with seven out of the eight tag-reader combinations tested attaining 100 percent successful reads.
"The overall results are very encouraging and indicate a favorable outcome with many types of tags and readers. This project has successfully demonstrated the feasibility of RFID for specific applications and has the potential to satisfy many common-use cases, especially inventory management," Hardgrave said in the center's announcement.
The research was not designed to measure the productivity aspects of using RFID, but for comparison the researchers did record the time it took to inventory a round rack of apparel items by reading product tags by bar code and by RFID. It took about nine minutes to complete the process with bar code scanning, and just 15 seconds with an RFID reader.
The study was presented to sponsors in June and publicly released this week. Since the June release Hardgrave has spoken with several retailers who are running RFID systems or pilots, and he said they report their read rate experiences are mostly consistent with the study findings.
The study provided data that can be used as a guideline for retailers to set expectations and identify potential areas for improvement, but does not answer the question of what read rates are needed to make RFID valuable.
"Everyone wants to talk about 100 percent read rates, but when you sit down with retailers, talk about what they need and really drill down into their processes, we find that 100 percent read rates often aren't necessary," said Hardgrave. "They are not getting 100 percent identification with the processes they have now. The question becomes 'How much better is good enough?' For example, today many retailers may only take full inventory twice a year because of the time involved. With RFID you can take inventory every two weeks, or even constantly with a shelf reader. I think if you can get into the 90s for your accuracy with that frequency of taking inventory, the visibility would be fantastic."
The summer of 2008 may be remembered as the "test season" for the RFID industry. In addition to the new RFID Research Center report, two other organizations also released the results of their studies of RFID tag and reader performance (see Myth Busted' -- Tests Find RFID Works Well on Metals and Report Reveals RFID Performance on Different Surfaces). The RFID Research Center is involved in ongoing RFID research and earlier this year released a report on inventory management (see RFID Yields 13% Reduction in Understated Inventory).
