Privacy-enhancing technologies are reshaping online advertising, making it increasingly difficult for advertisers to track customer journeys and link marketing actions to economic outcomes. As a result, interest in probabilistic measurement techniques such as media and marketing mix modeling (m/MMM) has surged, particularly among digital-first advertisers. Many of these advertisers are small and midsize firms that rely heavily on digital channels yet lack the resources to commission bespoke proprietary models. Against this backdrop, grassroots analytics communities and large online advertising platforms such as Meta Platforms and Google have started developing open-source packages for m/MMM. The packages share core principles but differ on important aspects such as their degree of automation, ease of use, and modeling paradigms. This article provides a narrative review and practice-oriented synthesis of these developments and discusses the challenges and opportunities that come with “packaged up” m/MMM. It concludes with an outline of topics that marketing scholars can engage with to vet and extend open-source m/MMM, and to align its continued development with its users’ needs.
Runge et al. (Wed,) studied this question.