Understanding product returns: A systematic literature review using machine learning and bibliometric analysis

Product Returns (PR) are an inevitable yet costly process in business, especially in the online marketplace. How to deal with the conundrums has attracted a great deal of attention from both practitioners and researchers. This paper aims to synthesise research developments in the PR domain in order to provide an insightful picture of current research and explore future directions for the research community. To ensure research rigour, we adapt a six-step framework - defining the topic, searching databases, cleaning and clustering data, paper selection, content analysis, and discussion. A hybrid approach is adopted for clustering and identifying the distribution and themes in a large number of publications collected from academic databases. The hybrid approach combines machine learning topic modelling and bibliometric analysis. The machine learning results indicate that the overall research can be clustered into three groups: (1) operations management of PR, covering (re)manufacturing network design, product recovery, reverse distribution, and quality of cores; (2) retailer and (re)manufacturer issues including return policy, channel, inventory, pricing, and information strategies; and (3) customer's psychology, experience, and perception on marketing-operation interface. Furthermore, from the content analysis, five potential future directions are discussed, namely digitalisation in the context of PR; globalisation versus localisation in the context of PR; multi-layer (i.e., retailer, manufacturer, logistics provider, online platform) and multi-channel (i.e., online, offline, dual and omni channel) oriented bespoke return policy; understanding and predicting customer return behaviour via online footprints; and customer return perception across the marketing–operations interface.