Unveiling Consumer Preference from Filtering Choices Using a Bayesian Dynamic Approach
(A2024-119481)

AUTHORS

Zoey Hu, University of Michigan; Xiaojing Dong, Santa Clara University

ABSTRACT

Making inferences about consumer preferences has been instrumental in marketing. Traditional methods rely on historical consumer-level data, a resource becoming scarce due to burgeoning privacy regulations. In this research, we introduce a Bayesian dynamic approach to decode consumer preferences by harnessing their search filtering choices such as adjustments on price ranges to refine search results. This approach allows us to infer consumer preferences with limited data. This inference holds the potential to bolster a company's capability to provide timely recommendations while adhering to the constraints of privacy regulations. Through a utility model tailored for price-quality tradeoffs, our model provides a closed-form solution, which brings insights into how various factors influence consumer choices. We validate the model's efficacy through simulations and actual data from a prominent travel agency, deploying the Markov chain Monte Carlo technique.