Resource Library Calendar Career Management Community
About The AMA Search
Login

The AMA connects you to a world of resources that deliver results, and help you succeed today and into the future. Join the AMA, and put the power of AMA membership to work for you.


Join AMA

About AMA

Email Print page

An Incentive-Aligned Mechanism for Conjoint Analysis 

Min Ding

Executive Summary
Building on the mechanism design literature, this article specifies a truth-telling mechanism that embeds standard conjoint studies in an incomplete information game and proves that it is the Bayesian Nash equilibrium for participants to reveal their true preferences in conjoint studies. In addition to its rigorous theoretical foundation, this mechanism contains several desirable features that will facilitate its adoption among practitioners. It does not require any changes in existing conjoint methodologies and could be used for all of them (e.g., rating, choice, polyhedral).

As a result, a practitioner could rely on his or her expertise in any specific conjoint methodology and perform the same data collection and analysis. Equally important, this mechanism removes the onerous burden of requiring that all product variations be available at the time of the experiment (as is required by existing incentive-alignment guidelines), such that only one product variation is needed at the time of the conjoint study.

In terms of additional effort, the only major nonfinancial burden it imposes on practitioners is to calculate each participant's willingness to pay for a product variation using the conjoint results after experiment. The additional financial burden is also limited; practitioners will need to provide real products as the prize of random drawings but only to the extent that the expected value of this random drawing for each participant is higher than his or her opportunity cost (e.g., a study of a $200 television may require a 1 in 10 chance of winning, whereas a study of a $2000 refrigerator may require only a 1 in 100 chance of winning). Finally, this mechanism does not impose any additional burdens on participants, other than needing to read extended (but easy-to-understand) instructions (note that a participant needs to learn only once if he or she participates in many different conjoint studies containing this mechanism).

The empirical tests conducted using the iPod Shuffle (Experiment 1) and iPod Nano (Experiment 2) packages demonstrate the superior external validity of the truth-telling mechanism and show that such improvement can be achieved using lottery incentives, which reduces the financial cost associated with the implementation of this mechanism. Using a realistic holdout task (16 different iPod packages and the no-purchase option), the mechanism can correctly predict the choices of 36% of the participants in the truth-telling mechanism compared with only 17% in the control condition in Experiment 1. Similar improvement is observed in Experiment 2 (34% versus 21%).

Biography
Min Ding received his doctoral degree in marketing in 2001 from the University of Pennsylvania and his doctoral degree in molecular, cellular, and developmental biology in 1996 from Ohio State University. He is Assistant Professor of Marketing in the Smeal College of Business at Pennsylvania State University. His current research focuses on two areas in new product development. The first area is new product development challenges in high-risk, high-return industries epitomized by the ethical drug industry—for example, pipeline and portfolio, prelaunch forecasting, and alliance. The second area is designing incentive-aligned mechanisms to elicit better responses from customers in new product research, such as conjoint analysis. He has published in journals such as Management Science, Journal of Marketing Research, and Journal of Biological Chemistry, among others.

Journal of Marketing Research, Vol. XLIV, No. 2, May 2007
View Table of Contents.

AMA IconPowered by the American Marketing Association | Copyright © 2010 MarketingPower, Inc. The site content may not be copied, reproduced, or redistributed without prior written permission from the American Marketing Association or its affiliates.