Inventory Auditing and Replenishment Using Point-of-Sales Data
41 Pages Posted: 28 Oct 2014 Last revised: 23 Sep 2018
Date Written: September 14, 2018
Events such as spoilage, expiration, damage due to employee and/or customer handling, employee theft, and customer shoplifting reduce inventories in retail stores without these reductions being reflected in inventory records. As a result, inventory records often include phantom inventories (i.e., units of goods present in inventory records that are not actually available for sale). These phantom inventories cause replenishment delays and stockouts, which ultimately hurt service levels. The optimal auditing and replenishment policy in the presence of phantom inventories is complex. We analyze the structure of the problem using tools from discrete mathematics and dynamic programming to derive a polynomial algorithm for computing the optimal policy in our setting. Moreover, we propose a simple auditing and replenishment policy based on a threshold on the estimated fraction of demand to be met on a given day conditional on the POS data up to that day, a statistic that we refer to as the daily expected service level. We compare that policy with the optimal policy in our setting, which we compute algorithmically, and show it performs comparably in most cases. Our simple policy is easy to compute and interpret, and thus offers an attractive solution for inventory managers facing the challenge of phantom inventories.
Keywords: shelf availability, inventory-record inaccuracy, POMDP, optimal replenishment, retail analytics
JEL Classification: C11, D81
Suggested Citation: Suggested Citation