Cost-Effective Coupon Targeting Method Based On Uplift Modeling

Abstract

Since the Internet has become popular in the world, electrical commerce(e-commerce) has made a major advance and established an enormous market. According to Ministry of Internal Affairs and Communications in Japan, it is reported that global sales of e-commerce reached 4.25 trillion USD in 2020 and expected that the global sales keep increasing. In order to acquire customers, e-commerce companies have been taking various actions. Discount price coupon is one of the effective ways for customer acquisition. Previous studies show that coupons increase purchase intention. However, they have not only positive side. Compensation for discount price reduces revenue. In addition, coupon targeting is limited by budget. For these reasons, we set our goal to invent a cost-effective coupon targeting method under budget constraint. In this article, we proposed a cost-effective coupon targeting method based on uplift modeling. The proposed method decides coupon users with an indicator of cost-effectiveness estimated by machine learning. By coupon targeting simulation and comparison with a conventional method, we verified that the proposed method outperforms the conventional method.