Google Abandons FLoC in Favor of New Topics API, a Replacement for Third-Party Cookies

Google is burying FLoC (Federated Learning of Cohorts) in its sea of abandoned experiments. FLoC’s proposed mechanism for replacing third-party cookies grouped people together and labeled them using machine learning. The controversial origin trial caused a group of WordPress contributors to propose blocking it in core before it was even out of the experimental stage.

In July 2021, Google concluded FLoC’s origin trial and removed the project from the testing phase while analyzing feedback. The company did not elaborate on why it dropped FLoC but had previously stated an intention not to share any feedback from the project.

FLoC ran in limited markets and received overwhelmingly negative feedback from the tech industry that left Google with an uphill battle to get enough buy-in to proceed. Google was not able to get any major browsers on board and Amazon, GitHub, Firefox, Vivaldi, Drupal, Joomla, DuckDuckGo, and other major tech companies and open source projects had opted to block FLoC by default.

As an alternative, Google is now proposing Topics, an interest-based advertising approach, which it says was informed by community feedback from the FLoC trials:

With Topics, your browser determines a handful of topics, like “Fitness” or “Travel & Transportation,” that represent your top interests for that week based on your browsing history. Topics are kept for only three weeks and old topics are deleted. Topics are selected entirely on your device without involving any external servers, including Google servers. When you visit a participating site, Topics picks just three topics, one topic from each of the past three weeks, to share with the site and its advertising partners. 

Google contends that this new system would give users more control and plans to build controls into Chrome for viewing the topics and removing irrelevant ones, as well as disabling the feature completely.

The Topics API would be implemented at the browser level and curated to exclude topics like gender or race, data which could be deemed sensitive.

Chrome’s weighty market share would position Topics to be widely used if Google decides to move forward with it, but one hurdle with industry-wide adoption would be convincing other browsers to invest in implementing new ad tech.

A list of proposed taxonomies is available on the Topics API GitHub repository. The README file includes a technical rundown of how the API will work. It also explains Topics’ evolution from FLoC and lists FLoC’s various deficiencies – i.e. adding too much fingerprinting data to the ecosystem, lack of transparency in the API for users, inclusion of sensitive data, among others.

Google plans to launch a developer trial of Topics in Chrome that will include user controls and allow developers and the ad industry to test it. The technical details of the API are expected to change as Google gets more feedback.


3 responses to “Google Abandons FLoC in Favor of New Topics API, a Replacement for Third-Party Cookies”

  1. That was a pretty quick death but not surprising considering virtually everybody in the universe was opposed to FloC.

    Let’s see Topics fail too.

    How about let people opt in and specify their own interests, then give them some type of reward for doing so? That might not be good for the bottom line but I don’t think it’s too much to ask.

    Collecting information about people without them “really” knowing violates “Don’t Be Evil”, no matter how dumbed down the approach is.

    • Indeed. And while a more direct reward would be nice, not seeing ads that are irrelevant or maybe even objectionable could be considered as one.

      I mean, there may be times when you might even want to see some ads, when you’re looking for something, and most definitely there are many things you don’t want to see. So don’t track, but instead let people freely choose what they’re interested in AND also what they do NOT want to see, and then pick the ads first according to the topics of interest, then if there aren’t enough from those go through all others while avoiding those specified as not interested (though since the ad space remains the same, if there won’t be enough ads available without avoiding those, or for those who’d put pretty much everything in that category, some will out of necessity be picked from it too).

  2. I cannot imagine why I would want to run a web browser that contains components that are only there in order to inform third parties how to better target me with advertising. That’s because I conceive of software as a tool I choose to deploy in order to serve me and my ends. Google apparently look at the purpose of software quite differently.
    I mean, if someone really wants their web browser to do that, then by all means, let Google provide an extension for it, and people can opt-in to use it. I would expect that the low numbers of people doing so would be very revealing.


Subscribe Via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.