In mid-October, Wal-Mart announced that a University of Arkansas study showed the use of RFID to track cases of products in Wal-Mart stores led to a 16 percent reduction of out-of-stock products and faster shelf replenishment of those items over items tracked via bar codes at the case level (see EPC Reduces Out of Stocks at Wal-Mart).
An 18-page research paper based on the study was published, entitled “Does RFID Reduce Out-of-Stocks? A Preliminary Analysis.” The paper describes the Wal-Mart-commissioned study in detail, outlining how out-of-stock levels were measured and compared between the 12 RFID-enabled stores and 12 control stores. It also explains the methodology researchers used to account for the natural fluctuations of out-of-stock levels within the test and control stores, so that a fair comparison could be made.
Prior to the study, Wal-Mart leveraged its RFID system to make an important business process change in how it monitors and manages shelf stock. Instead of manually inspecting stock levels on shelves or back-room stock to generate pick lists, Wal-Mart now combines point-of-sale data with data generated from RFID readers located at the loading dock, at the doorway between back room and sales floor, and at the box crusher (indicating empty cases). All that data is used to generate these lists automatically.
The RFID-generated lists correlated with reduced out-of-stock levels within the test stores, highlighting the significance of leveraging RFID technology to change and improve business processes—not just in Wal-Mart stores, but in any retail environment.
The report breaks out how out-of-stocks were reduced within the test stores, showing what happened when the staff used non-RFID-generated pick lists, partially RFID-generated pick lists and fully RFID-generated lists. Compared with the weekly out-of-stock levels of test stores using a non-RFID-generated pick list, the out-of-stock levels improved by 15 percent at test stores using a partially RFID-generated list, and by 26 percent at stores using a fully RFID-generated list. During the 29-week test period, however, the control stores also experienced an improvement in the average weekly out-of-stock level. Compared with the control stores, the test stores using a fully RFID-generated list improved by approximately 16 percent. This data is mapped on a graph, as is a comparison of the out-of-stock rates of tagged stock-keeping units (SKUs) versus non-tagged SKUs within the test stores.
The report analyzes the effect of the auto-generated pick lists and notes that when compared with pick lists manually generated by store associates, they contain more items, signifying they identify more low-stock items than associates see. The report also notes that over time, stock levels of SKUs from tagged cases improved and the amount of these SKUs on the daily auto-generated pick list fell, indicating the SKUs were consistently better-stocked.
It also notes that along with the reduction in out-of-stocks, the use of auto-generated pick lists saves time because store associates no longer have to scan empty shelves manually to generate pick lists. Tracking the savings in time, however, was not part of the research study, and the report contains no quantifiable proof of time reductions. However, the researchers who carried out the study postulate that Wal-Mart and other retailers could realize significant operational efficiencies through this and other effects of using RFID technology.
During the 29-week study, which ran from February to September, the researchers conducted daily inventories of 4,554 different (SKUs), representing products from all store departments, looking for out-of-stocks. Cases of these SKUs at the control stores were received and processed using bar codes affixed to the cases—which sometimes contained a single unit of the SKU. RFID tags affixed to the cases of the same SKUs were used to receive and process those cases at RFID-enabled stores. Prior to the 29-week study, the out-of-stock rates at both the test and control stores were measured for a period of time before RFID applications were enabled, in order to establish a baseline.
The report was written by Bill Hardgrave, research lead and director of the university’s RFID Research Center, and executive director of its Information Technology Research Institute (ITRI), along with Matthew Waller and Robert Miller, Walton School of Business professor and doctoral student, respectively.
The University of Arkansas study is one of a number of studies looking into the use of RFID in the retail supply chain. “Current and future efforts are focused on rigorous statistical modeling, which will better isolate the RFID effect so that the magnitude of the change caused by RFID can be determined,” says the report, adding that RFID is one of many factors, such as weather and personnel fluctuations, that affect out-of-stock rates.