FAYETTEVILLE, Ark. — A preliminary analysis of the effect of radio frequency identification on retail-inventory accuracy demonstrated that an automated, RFID-enabled inventory system improved accuracy by about 13 percent in test stores compared to control stores. The investigation, conducted by researchers in the RFID Research Center, a research unit of the Information Technology Research Institute in the Sam M. Walton College of Business, also revealed that manual inventory adjustments by store personnel significantly declined in test stores due to the automated, RFID-based system.

“Inventory accuracy is one of the keys to an efficient and effective supply chain,” said Bill Hardgrave, director of the research center and principal investigator. “Yet, inventory accuracy, which determines important processes such as ordering and replenishment, is often poor, with inaccuracy rates sometimes as high as 65 percent. Our results suggest that RFID technology makes a difference. The 13-percent improvement found in this study can significantly reduce unnecessary inventory, and the value of this reduction for a company like Wal-Mart, with all of its suppliers, can be measured in millions of dollars.”

Inventory accuracy is a chronic problem in the retail industry. Retailers focus on what they call “perpetual inventory,” a name to describe an estimate of inventory, based on various systems and methods of tracking items. As Hardgrave mentioned, previous research has demonstrated huge gaps between perpetual inventory — what managers think is on hand — versus the actual number of items in a store, either on shelves or in a stock room. Studies have found that retailers generally have accurate inventory information on only 35 percent of their items.  

Perpetual inventory can be understated or overstated. Understated, sometimes called hidden inventory, means that perpetual inventory shows fewer items than what are actually in the store. Conversely, overstated, also known as phantom inventory, describes a store in which perpetual inventory shows more inventory than items on hand.

Incorrect manual adjustments by personnel, stolen products, damaged or spoiled products not recorded as such, returned products not properly accounted for, incorrect shipments from distribution centers and cashier error are the six major causes of inventory inaccuracy, which can lead to out-of-stock items or excess inventory. Because of inventory inaccuracy, systems may order unnecessary product or fail to order product that is needed. Hardgrave emphasized that the net result of inventory inaccuracy, as reported in other research, is an estimated 10 percent reduction in profit.

Focusing only on understated inaccuracies, the Arkansas study involved 16 Wal-Mart stores — eight test stores and a matching set of eight control stores. Test stores were selected from the existing set of approximately 1,000 RFID-enabled Wal-Mart stores. Control stores were then chosen based on a set of criteria used to determine a comparable profile, including demographics, size of stores measured by square feet, annual sales and the absence of known impacts such as annual inventories, remodeling or resets, market trials and other known disruptions. The research sample contained a mixture of Supercenter and Neighborhood Market stores.

For 23 weeks — from May through October 2007 — a national inventory auditing group hand-counted all individual items in the air freshener category in all 16 stores. A single category was chosen to provide the opportunity to tag all cases in that category. The daily inventory of a particular store started at approximately the same time each day, and the auditors followed the same counting pattern — starting at bottom left and working to the right and then up. Stores were counted between the hours of 4 and 8 p.m.  

Test stores were equipped with RFID readers/antennas at various backroom locations — receiving doors, sales floor doors and box crusher. Control stores had no RFID technology. Test stores were provided with a perpetual-inventory adjustment system, dubbed “auto PI,” that automatically adjusted understated inventory. Other than the auto-PI system in the test stores, which worked automatically without human intervention, no additional manipulations were introduced, meaning both sets of stores operated business as usual, and store personnel were instructed to carry out their jobs in the same way they would in normal situations.

RFID, via the auto-PI system, served as a supplement to the existing process of adjusting inventory so that results of the study would demonstrate how effective RFID is beyond existing processes. Control store personnel did not modify or stop their manual adjustments. Finally, to establish a baseline for perpetual-inventory accuracy, inventory was counted for 10 weeks before auto-PI system was turned on.

Data revealed that the percentage of understated items off by more than two units fell by 13 percent in the test stores compared to control stores. Furthermore, the RFID-enabled auto-PI system doubled the number of inventory adjustments, suggesting that only half of all manual adjustments are caught in a given retail store. Hardgrave emphasized that increasing the number of manual adjustments to equal those captured automatically by the system would accordingly increase labor dedicated to this task and thus distract workers from stocking shelves or assisting customers.

“Instead, as demonstrated in our study,” he said, “perpetual-inventory accuracy was improved with no additional labor.”

The study is available for download at http://itri.uark.edu/research. Enter “rfid” as the keyword.

Hardgrave, holder of the Edwin and Karlee Bradberry Chair in Information Systems, is also executive director of the Information Technology Research Institute.