Google just unveiled Topics, a new model for post-third-party-cookie ad targeting that replaces the company’s previous proposal, Federated Learning of Cohorts (FLoCs). We may remember FLoCs as the last stand of the cookie. The model proposed targeting groups of users based on their interests and behaviors — essentially, behavioral, cookie-based targeting as we know it for groups instead of individuals.
But Topics shows Google taking a markedly different approach to ad targeting. Topics will collect data on users’ interests over a rolling three-week period, assigning each user’s interests to one of 300 Topics. When advertisers want to reach users, Topics will randomly assign them three possible topics — one high-interest topic for each week of browsing data — that advertisers can use to target consumers.
