An Incentive-Aligned Mechanism for Conjoint Analysis

Abstract

This article specifies, analyzes, and validates a rigorous and practical truth-telling mechanism (game) for conjoint applications. The mechanism requires only one real product variation and has truth telling in conjoint as its Bayesian Nash equilibrium, thus making it possible to incentive align participants in most conjoint applications. Using the iPod package as the context, the author shows empirically that the mechanism substantially improves purchase prediction compared with a standard conjoint procedure.