January 26, 2022 Google announced its second attempt to allow advertisers to purchase ads based on users’ browsing interests without relying on invasive tracking cookies.

Google’s primary solution, known as Federated Learning of Cohorts, or FLoC, aims to divide people into large baskets determined by algorithms based on the websites that users visited in the previous week. Advertisers can display ads about a particular basket, but the people in them remain anonymous.

However, tests last year showed that some advertisers found FLoC to be less efficient than cookies for selecting users to target. It also carried the risk of revealing a user’s browsing history, said Google product director Vinay Goel.

Companies in the $250 billion global online advertising sector are concerned that the elimination of cookies will make them more dependent on buying ads from Google and Facebook, owing to their vast user databases.

This new browser system, called Topics, categorizes a user into up to 15 baskets of 350 human-designed choices such as “fitness” and “travel,” based on three weeks of browsing. Advertisers will be able to view up to three baskets per user while deciding whether to show that user an ad.

Goel said browsing is only tracked on websites that activate themes, and users could distance themselves from a topic or remove the technology altogether.

For more information, read the original story in Reuters.