I-Corps: Business Analytics for Large Scale Intelligence

Purdue University, West Lafayette IN

Investigators

Abstract

For the last 20 years, traditional "brick-and-mortar" stores have been trying to find strategies to compete with on-line store in understanding customers' behaviors and preferences. An on-line store can easily create personalized advertisements, based on what a customer has purchased and seen. In contrast, traditional stores do not have such information and cannot provide personalized experience attracting customers into the stores and encouraging customers to buy after they are in the stores. Understanding customers' needs are currently met by surveys, focus groups, in-store observers, sales records, and market trends. Surveys and focus groups are costly and disrupt customers' buying process. In-store observers are labor-intensive and cannot record many customers' activities at once. Sales records show what has been sold, without any information about why customers do or do not purchase products. This research team has created the Business Analytics for Large-Scale Intelligence (BALSI) system that allows retailers to make real time, dynamic, and personalized advertisements and promotions to attract customers into the stores. The BALSI system uses visual analysis of video stream from in-store camera systems to observe a shopper's behavior, movement, gaze and other actions. From such data, information can be mined to help the retailer better attract, retain, and satisfy customers. BALSI's software solution analyzes the videos captured in stores without interfering with the customers and does not require any additional employees. The software will analyze the characteristics of customers to presents trends and patterns to the clients and to create dynamic advertisements to attract the customers into stores. Inside the store, the software can adaptively try to increase the likelihood of a purchase by a shopper. BALSI's software also augments human sales people by alerting a store associate to help a customer that appears to be looking for something. The technology does not recognize faces and offers better privacy protection when compared with technologies that attempt do similar things within the retail area by tracking the signals coming from a customer's phone.