Dealing with Learning in eCommerce Product Navigation and Decision Support: The Teaching Salesman Problem

It is generally assumed that consumers in electronic shops feel more comfortable engaging in a dialog regarding their needs as opposed to inspecting detailed product features. However, empirical evaluations of needs-based recommender systems indicate that consumers are only confident about their buying decision if they can compare detailed information about product features. Ideally consumers should be supported in learning the relationship between their needs and product features. Current eCommerce-recommender systems do not support the required smooth transition between needs-oriented and feature-oriented interaction. In this paper we present a novel, homogeneous representation model for needs and feature preferences. The model enables a recommendation process that guides consumers from a needs-oriented to a feature-oriented interaction and thereby enables consumer learning and fosters confidence building. We illustrate our proposal with an eCommerce recommender system demonstrator for digital cameras.