[1I3-GS-2] Machine learning: Market analysis

[1I3-GS-2-02] Analysis of Advertisement's Effects by Products and Customer Groups Using Propensity Score

〇Takaaki Ishiwatari¹, Takahumi Okukubo¹, Risa Honda¹ (1. Graduate School of Systems and Information Engineering, University of Tsukuba)

Keywords:Propensity Score, Advertisement's Effects, Latent Class Analysis

In the commoditizing market, it is difficult to make profits by decreasing the price of the product due to the excessive price competition. This study aims to analyze the advertisement's effectiveness for product and customer classes with the propensity score method and to examine measures for de-commodification. Firstly, latent class analysis is applied to classify customers based on their genders, ages, and consumption values. Secondly, we selected the covariates by employing the regression analysis method and then measured the advertisement's effectiveness for each class with the propensity scores. As a result, it becomes clear that there are some differences among the classes for which the advertisement effectiveness is recognized in the commoditizing market. We also discussed the possibility of realizing effective advertisements for each product and various customer groups to differentiate the products from the competitors’ markets and to realize the de-commodification of the market.