Moving towards inferential attribution modelling in a world without third-party cookies

Abstract

With the gradual disappearance of third-party cookies and Identifier for Advertisers (IDFA) tracking, marketers are becoming more restricted in their capacity to measure the performance of their marketing initiatives. Standard attribution models are currently based on user-level data to establish a one-to-one relationship between customer interactions and conversion goals. However, with user-level data about to become more difficult to access, marketers will need to embrace alternative ways to measure the effectiveness of their marketing efforts. This paper proposes inferential attribution modelling techniques as a potential alternative or complementary approach to user-level attribution techniques, and revisits older marketing techniques, such as media mix models, to address the upcoming changes in the marketing data ecosystem.