Teaching Agents to Buy (TAB): An Incentive Aligned Decision Rule Elicitation Method

It has been well-documented that consumers use a variety of decision rules (compensatory and non compensatory) in purchasing decision making. In most cases, they first reduce the entire universe of product alternatives to a small consideration set (CS), and then purchase one from CS. This paper designs and tests an intuitive mechanism, Teach Agents to Buy (TAB), which incentive aligns respondents to truthfully reveal the decision rules they use in this process. An empirical study will be conducted to show TAB is able to elicit decision rules that can reduce consumer decision alternatives to a small consideration set more efficiently and accurately. When combined with state-ofart conjoint analysis, the predictive performance of TAB-Conjoint is substantially better than that by conjoint analysis alone. Based on its theoretical property and the empirical evidences, the researchers believe TAB is a valuable tool to practitioners, either as a standalone tool for uncovering decision rules and CS, and/or as an add-on to existing preference measurement methods such as conjoint analysis.