Assessing an on-site customer profiling and hyper-personalization system prototype based on a deep learning approach

Abstract

• This article examines the perceptions of early adopters of an AI-based on-site customer profiling and hyper-personalization system (OSCPHPS). • The article employs a qualitative comparative analysis to outline the causal recipes clustered in three categories: facial expression recognition, automatic age and gender recognition and in-store customer tracking, enabling a high intention to test OSCPHPS prototype. • The findings offer strategic guidance to marketing managers of retail stores in order to capture the great hyper-personalization opportunities created by AI technologies. • We prove that deep algorithms and neural convolutional networks are highly valuable in analyzing customer profiles and offering hyper-personalized products and services to customers. The development of artificial intelligence (AI) technologies is proceeding fast across many fields. Based on a deep learning approach, we propose a prototype of an on-site customer profiling and hyper-personalization system (OSCPHPS) targeted at marketing professionals. We propose an AI platform to create customer profiles during their physical presence in stores. The idea of the OSCPHPS prototype is to automatically detect and gather customer data directly from the store, essentially completing customer profiles containing gender, age, personality, emotions, and products they interacted with or bought, irrespective of where they are in the store. Each buying operation could generate an anonymous customer profile. Therefore, for every product sold, the system will track multiple customer-generated profiles of the people who bought that product. These kinds of data offer endless further possibilities for the business. Through a configurational study conducted via fsQCA methodology, we assessed the interest in the OSCPHPS prototype on the part of marketing managers of clothes & fashion stores located in different European countries. Based on these live generated profiles, we could further enhance the OSCPHPS system by adding support for customer segmentation, strategic product campaigns, live product recommendation, analysis of emotions toward a product or a category of products, sales forecasts, and personalized store space enhancements based on augmented reality, customer exploratory statistics and customer purchasing patterns.