Lift Effect Analysis by Regression Discontinuity Design in Retargeting Ads

Abstract

Internet ad serving is one of the major businesses in the Internet industry. Among them, retargeting advertisements are known to be a particularly effective form of delivery. On the other hand, there is still an ongoing academic debate on the effectiveness of retargeting ad delivery. In this study, we estimated the causal effects of ad bidding and performance in retargeting advertising by applying a regression discontinuity design with the lowest bid price in ad delivery auctions as the threshold. As a result of the estimation using data from actual retargeting ad delivery platforms, we were able to confirm positive ad-lift effects on major KPIs in ad delivery, such as app launch and billing, for many advertisers.