BB-FLoC: A Blockchain-based Targeted Advertisement Scheme with K-Anonymity

New data protection regulations, e.g., General Data Protection Regulation (GDPR), enforced advertisement providers to amend their conventional approaches, enhancing users’ data privacy. As a result, major Internet browsers such as Apple Safari and Firefox were quick to announce their plans to remove third-party cookies from their browsers entirely. In an effort to preserve conventional advertising practices, Google proposed a Federated Learning of Cohorts (FLoC) system to deliver higher privacy guarantees to users whilst also providing interest-based advertising. In FLoC, users sharing similar browsing histories are put into cohorts, and thus advertisements can be targeted to them as a group, rather than individually. Since each user independently calculates their cohort group, a minimum cohort size cannot be enforced, making them vulnerable to identification and tracking. To address this issue, in this paper, a blockchain-based FLoC (BB-FLoC) system is proposed that guarantees k-anonymity for its users whilst at the same time allowing for effective personalised advertising. We further evaluate the operational feasibility of such a design and demonstrate that k-anonymity guarantees can be fulfilled in a fully decentralized manner in the proposed system. The proposed system is relatively lightweight, showcasing that it can be adapted for low-end devices such as mobile phones.