'Listening In' to Find Unmet Customer Needs and Solutions

Preprint
Publisher Website
Google Scholar
Abstract
We explore a practical methodology to identify new high-potential "fishing grounds" for product development by "listening in" to ongoing dialogues between customers and web-based virtual advisors. Customers naturally seek out these virtual advisors to aid their purchasing decisions, hence customers have the incentive to express their true needs to these advisors. We study a representative virtual advisor that uses Bayesian updating methods to select maximum probability recommendations for customers. This virtual advisor selects its questions efficiently through a two-step look ahead that maximizes the information potential of each question. We show that, for this virtual advisor, the maximum (forecast) probabilities of choice are indicators of underlying utility. Drops in underlying utility indicate that no existing product fulfills the conflicting needs that led to the drop. By defining a trigger mechanism and examining correlations among expressed customer desires, we identify the conflicting needs that remain unmet by the current marketplace. Once unmet needs are identified, an automated virtual engineer probes the antecedents of the unmet needs to gather information that is relevant to the product-development team. To explore the unmet needs further, a design palette enables the customer to express his or her own solutions. We examine the internal validity of the listening-in methodology with Monte Carlo simulation and demonstrate that the methodology can recover known unmet-need segments. We then apply the automated system by listening in to a "Truck Town" virtual advisor. This application, with over 1,000 web-based customers, identified at least two new opportunities for pickup-truck platforms. Together these opportunities represent significant incremental revenues for the truck manufacturer.