Prediction of Product Purchase based on Conjoint Measurement with Process Simulation:

Manipulation with Photos and Confirmation of Improvement in Prediction Accuracy through Follow-up Survey

Abstract

The purpose of this study is to demonstrate that process simulation can improve the prediction accuracy of marketing research. First, we argue that photo-simulation-manipulation, proposed by Takeuchi (2015), is based on process simulation. Considering this argument, we conduct a randomized experiment in which participants are assigned to three groups (photo-simulation-manipulation vs. how-manipulation vs. non-manipulation). The participants answer the mind-set manipulation task based on the condition assigned to them and conjoint measurement in the main experiment. In addition, we investigate the products that the participants actually purchased in a follow-up survey. We find that in contrast to non-manipulation, photo-simulation-manipulation improves the prediction accuracy of participants' purchase deduced from conjoint measurement. We discuss the reason for this result and conclude that the participant responses better reflect actual purchase behavior due to activation of memory of past purchases and association with the target products caused by process simulation.