Managing conversions of anonymous online users: a privacy-compliant framework (JOBR-D-24-07633.R1)

Privacy concerns are on the rise. Privacy protection efforts and regulations pose technical challenges for extant digital marketing practices. The study finds that prospective customers tend to browse websites with different intents and therefore proposes a semi-supervised machine learning model to infer the browsing intent of each user while fully preserving the confidentiality of user-level browsing data. The authors field test the methodology at a large financial services firm and observe a significant lift in engagement and conversion rates relative to conventional approaches. The study contributes to the theory and practice of online marketing by (a) developing a mechanism to anonymously infer the browsing intent(s) of prospective customers, (b) further building upon the theory of flow, and (c) proposing a novel methodology-based framework to improve website performance in the age of increasingly stringent privacy regulations.