Causally Motivated Attribution Methodology for Online Advertising

This paper is a research article that presents a causally motivated methodology for conversion attribution in online advertising campaigns. It discusses the challenges associated with assigning conversion credit to multiple advertising channels that serve ads to users. The authors emphasize the need for standardization in attribution measurement and propose three guiding principles to aid in this endeavor. They frame attribution as a causal estimation problem and introduce two approximation methods for situations where full causal estimation is not feasible. These methods are informed by cooperative game theory and can be interpreted as variable importance measures when causal assumptions are not met. The paper also provides empirical examples of attribution measurement using various online advertising campaign datasets, highlighting the limitations of the current last-touch attribution model and the importance of developing a more robust framework for measuring the effectiveness of advertising channels.