Topics will allow Chrome to choose up to a maximum of five topics that represent a user’s top interest for a certain time period (three weeks), based on the websites they visited. Examples of Topics could be travel, music genre, books, team sports, etc.
Websites would opt-in to participate in Topics API to serve personalized ads. When a user visits a site, the API will select three topics (one from each of the past three weeks) and share them with the site and its advertising partners.
The three topics unique to the user will be available for the next three weeks. Afterward, they will be replaced by new topics.
Google plans to start with 350 topics to categorize user interest and activity. They will expand the set of topics as they continue to test and receive feedback. AdExchanger reports that Google may be seeking a third-party like IAB Tech Lab to maintain the taxonomy.
Improving on the shortcomings of FLoC, the new Topics API will address the sensitive category targeting issue by pulling topics curated by humans (instead of machines) and manually exclude sensitive categories like race, religion, or sexual orientation.
Also, to bypass the fingerprinting issue, Google will limit the number of topics shared of each user. Chrome users will have the option to disable the Topics feature completely, or see what topics are being shared about them.
Three topics per user every few weeks sounds limited, but Google is trying to counterbalance that by offering combined topics, for example, a book lover who also loves hip hop, to enable a deeper level of audience targeting.
Google’s Topics API will utilize machine learning via user devices. Take a look at this step-by-step breakdown: