[1K3-GS-3] Knowledge utilization and sharing

[1K3-GS-3-05] Estimating the Optimal Ad Serving Media with Hierarchical Bayesian Model Using Customer Attribute Data

〇Reina Komoda¹, Haruka Yamashita¹ (1. Sophia University)

Keywords:Hierarchical Bayesian Model, MCMC, Internet Advertising, Customer Attribute, Advertising Effectiveness

Internet advertising expenditures have been increasing in recent years, and the market is expected to grow in the future. Among these, most of the market is dominated by managed advertising, which enables targeting based on consumer attributes and the determination of advertising distribution media. Therefore, it would be a great advantage for companies if they can clarify the optimal advertisement delivery destination for each consumer and improve the advertising effectiveness of managed advertisements. In this study, we construct a model for estimating the optimal Internet ad serving media using a hierarchical Bayesian model that enables flexible model construction that considers differences in consumer attributes. Based on the assumption that the distribution media with the greatest advertising effectiveness differs depending on consumer attributes, we propose an analytical model that introduces a hierarchical Bayesian framework. First, the hierarchical Bayesian model is used to analyze the relationship between consumer attributes, changes in purchase intention for a given product, and the frequency of use of each Web medium. Furthermore, the obtained parameter values are used to calculate the effectiveness of each medium in terms of the attributes of the target consumer, and the optimal destination of advertisements is determined.