Four Challenges for IML Designers: Lessons of an Interactive Customer Segmentation Prototype in a Global Manufacturing Company

Abstract

Interactive Machine Learning (IML) apps permeate all aspects of businesses, including sales. Customer segmentation is integral for sales, to identify customer groups to serve them appropriately. However, the novelty of such apps in a sales context raises the question: what challenges designers of such apps will face, in a sales context? To explore this question, we report our reflections on an IML study for customer clustering. We used data from a global manufacturing company to cluster customers using the Recency, Frequency, and Monetary (RFM) method. We applied a machine learning clustering model (K-Means) and discussed with seven seasoned sales managers the interaction with clusters. We report four challenges and foresee that designers of such systems will face, in the context of sales operations.