Sales Floor ROIBy Mel Duvall | Posted 2007-10-03 Email Print
In 2003, Wal-Mart became the great hope of a new technology. But things haven't worked out exactly as planned.
Sales Floor ROI
Wal-Mart's change in strategy to focus more on creating a complete window into its supply chain, from manufacturing facilities to sales floors, is backed by early return on investment studies, says Bill Hardgrave, director of the RFID Research Center at the Sam M. Walton College of Business, University of Arkansas. The center operates a lab at Hanna's warehouse where businesses can bring their products— shoes, razors, apparel—and test how to best tag products. In addition to RFID tag printers, readers and software from various vendors, the lab has a conveyor loop equipped with readers for testing cases, a fork lift with an RFID reader that can record where products are being stored in a warehouse, and loading bays with RFID readers on each side of the door.
In November 2005, Hardgrave and his colleagues also published the most complete study to date on the effects of RFID on out-of-stock situations. The study took place over an eightmonth period and involved 24 Wal-Mart stores in Texas. "It was a huge undertaking, not only because of the number of stores involved, but also because of the number of products that had to be inventoried," Hardgrave says.
The researchers hired a small army of personnel to go through each store and scan the inventory every day. It took 20 people about eight hours to scan the approximately 4,550 unique products stocked at a typical Wal-Mart Supercenter. All items in the stores were scanned, with the exception of "impulse" items that move too quickly to be counted accurately. During the first two months, RFID capabilities were not used, in order to measure typical out-of-stock frequencies. Then, during the next six months, RFID systems were turned on in half the stores for comparison with the other half, which served as a control group. In stores without RFID, department managers make daily rounds to determine which items require restocking. At RFID-enabled stores, Wal-Mart's inventory system compares back-room inventory, obtained by automatically reading RFID tags as pallets come in through the loading bays, with sales data from the point-of-sale system. The software then creates an "autopicklist" of products that need replenishing. The study produced dramatic results. Out-of-stock incidents decreased 26% in the stores using RFID autopicklists. On items with high turnover, the improvement was closer to 60%. Hardgrave says the latter result is particularly significant because it is much more important to ensure that, say, Gillette Mach 3 razors, of which a store might sell 15 to 30 a day, are kept in stock than a lowerturnover item like a toaster, of which a store might sell two or three a week.
The impact of those RFID-enabled picklists can't be ignored. A study published in 2003 by Daniel Corsten, a London Business School professor, and Thomas Gruen, a professor at the University of Colorado, estimated out-of-stock situations shave about 3% to 4% off the average retailer's sales revenue. In other words, a retailer with $100 billion in annual sales could increase sales by as much as $3 billion to $4 billion by eliminating out-of-stock situations. For Wal-Mart, which had $349 billion in sales in 2006, that would mean an additional $12 billion a year.
No technology is going to completely eliminate out-of-stocks, of course. Wal-Mart CIO Ford has estimated that RFID could help the retailer achieve a 30% reduction in out-of-stocks, which could generate an additional $3.6 billion in sales annually.
Getting a handle on costs is more difficult. In 2004, AMR Research estimated suppliers would need to spend $13 million to $23 million to meet Wal-Mart's requirements. But AMR later determined that suppliers had spent the bare minimum, $1 million to $3 million, which was "enough to purchase tags, readers and minimal software." But even those revised figures may be too high. Dallas-based consulting firm Incucomm surveyed 137 initial Wal-Mart suppliers and found that average spending is closer to $500,000, and median spending is about $200,000.
The positive interpretation of those findings, Hardgrave says, is that RFID isn't as expensive to implement as initially thought. The negative interpretation is that companies are not yet investing in systems such as business intelligence and analytics software that would allow them to capitalize on the data being returned by RFID tags and readers.
Still, progress is being made on the technology and standards fronts. EPCglobal, an industry-driven organization working to develop electronic product code (EPC) standards for use with RFID, has succeeded in driving a consensus in several key areas. In 2005, it established a protocol that laid the foundation for a second, more powerful generation of RFID tags, known as Gen2 tags. Earlier this year it created a general protocol for lowlevel RFID readers. And in probably its most important breakthrough, it ratified a global standard for secure real-time sharing of data from RFID systems. The EPCIS standard should simplify many processes, such as container tracking among multiple business partners, promotions management and electronic proof of delivery.
Harold Boeck, a professor at the Universite de Sherbrooke, Quebec, and a founder of Academia RFID, an RFID training and research facility, says one by one, the technology and standards hurdles are being overcome. "There are still some technology challenges, but technology has never really been the main issue," Boeck says. "The biggest challenge has always been around trying to quantify the benefits of RFID and capturing the ROI."