Real time out of shelf detection using embedded sensor network

Out-of-shelf problem is important to solve for retail store since the absence of products on the shelf can lead to a significant reduction of shoppers and a consequent drop on sales. For this purpose, it is necessary to study and to introduce approaches able to establish the lack of products on the shelves and thereby promptly ensuring their repositioning. In this context, the paper investigates the use of artificial intelligence techniques in detecting the out-of-shelf products. Particularly, having sales data, ordering info and product assortment of the store available, we study the development of low cost shelf detector that is based on wireless sensor network, and that can automatically discover out-of-shelf situations on a daily basis for all the stores of a retail chain. The use of an automatic method for detecting products that are not available on the shelf based on sales data would offer an accurate view of the shelf availability, both to retailers and to product suppliers. The tool presented is the first being installed for a long time in a high number of stores and products demonstrating the ability to gather data from there and extract interesting insights. This paper aims to present the hardware infrastructure of an embedded sensor network devoted to real time shelf out-of-stock management and to demonstrate the feasibility and the scalability of the system in providing a lot of data and interesting insights for store team and brand's marketing team.