Google Retires FLoC and Launches Topics API
Published: February 2, 2022
Author: Joe Kerschbaum
Privacy Sandbox Announcement & Explanation
The depreciation of third-party cookies continues to speed up and 2022 is shaping up to be a fast-paced year in this area. Effective digital marketing in a privacy-first environment is going to require a new perspective on measurement, and advertisers need to acclimate and embrace these changes soon. By the end of next year, Chrome plans to block ALL third-party cookies and Google is working furiously to have privacy-focused solutions in place before that deadline.
Landing on the best solution will require trial and error, and Google has determined the results of its initial public proposal, the Federated Learning of Cohorts (FLoC). This week Google announced that it is retiring its Privacy Sandbox proposal of FLoC and turning their development efforts toward a new project, Topics API.
The purpose of the Federated Learning of Cohorts was to analyze consumer browsing habits and use artificial intelligence to group consumers into themed cohorts. Advertisers would have been able to show relevant ads to these themed cohorts without having any identifiable information on any individual in the group. However, FLoC did not fly (pardon the pun) with privacy groups who were testing and evaluating the technology.
Underlying concerns kept arising that FLoC was not comprehensive enough when protecting user privacy. Vinay Goel, Google product director, told Reuters that in tests last year, some advertisers found FLoC less effective than cookies for choosing users to target and the system carried the risk of exposing an individual’s browsing history. This development does not surprise 3Q, as our agency was uncertain of FLoC’s ability to replace or replicate third-party cookies. According to the Topics API tech sheet, the main causes for concern included:
- Adding too much fingerprinting data to the ecosystem
- The need for more user transparency and user controls
- It was possible that topics, or groups of topics, are statistically correlated to sensitive categories
The second half of Google’s announcement was the launch of its new Privacy Sandbox proposal, Topics API. This new system groups each user in up to 15 baskets out of about 350 human-designed choices, such as “fitness” and “travel” based on three weeks of browsing. Here are a few details on how this works, according to Google:
- Topics are kept for only three weeks, and old topics are deleted
- Topics are selected entirely on your device without involving any external servers
- When you visit a participating site, Topics picks just three topics – one from each of the past three weeks – to share with the site and its advertising partners
- Topics enables browsers to give consumers meaningful transparency and control over this data
- In Chrome, Google is building user controls that let you see the topics, remove any you don’t like, or disable the feature completely
If you’re thinking FloC and Topics API sound similar you aren’t entirely wrong, but there are noticeable differences. With FLoC, browsers would gather historical data to group users into cohorts. Topics API does not create cohorts, but rather determines specific topics based on browsing history that is analyzed from the most recent three weeks. FLoC would share a cohort ID with sites and advertisers, but Topics API will select three topics to share. The initial taxonomy of topics within the API will include 350 subjects such as Arts and Entertainment, Business and Industrial, and Travel and Transportation. These consumer interests are derived from a list or model that maps website hostnames to topics.
Topics API – or at least the idea of algorithmically assigning topics anonymously to individuals – isn’t a completely new concept from Google. In 2019, Privacy Sandbox proposed “Private Interest Including Noise,” otherwise known as PIGIN (which was a terrible name and acronym). The Electronic Frontier Foundation (EFF) stated “While FLoC promises to match each user with a single, opaque group identifier, PIGIN would have each browser track a set of “interest groups” that it believes its user belongs to. Then, whenever the browser makes a request to an advertiser, it can send along a list of the user’s “interests” to enable better targeting.” Google is circling around a long-term service by experimenting with various solutions including FloC, Topics API, PIGIN, and FLEDGE.
Determining privacy-first targeting and measurement solutions are critical for brands. Digital advertising will continue to be an important element in every media plan, especially given the amount of time people spend online, but it will also force us to change the way we measure success. While Google works to find a macro-level solution to these challenges, this should remind advertisers that they need to develop systems and processes for audience development and targeting, and durable attribution, that is built on their own first-party data. Also, advertisers need to collaborate across various partner platforms to develop solutions that will meet their needs. 3Q continues to build our collaborative audience strategy and measurement approach with our partners including The Trade Desk, Lotame, and GWI.
Historically, Google has rolled out large-scale changes in waves and usually within a three-month (approximate) timeframe. They will not launch a large-scale update in Q4 2023 due to the holidays, which means they will likely implement this change beforehand, probably in Q3 2023. From our estimations, Google will announce the “official” date of the Chrome change in Q3 2022 with the actual change taking effect in Summer/Fall of 2023. That may sound far away, but because advertisers need to reimagine their attribution and measurement strategies, everyone needs to get started right now.
The longer it takes for Google to develop a solution, the more time it gives competitors to develop more robust solutions. As brands build their first-party data, the less need they will have for platforms like Google to provide solutions. Furthermore, as this decentralization accelerates, the more likely it becomes that a brand with a robust first-party database will bypass Google and build direct relationships with creators or media owners. The stakes are high for Google from both a media and technology standpoint, with the greatest risk to their tech stack and DV360/GDN.
3Q continues to make consumer privacy a priority as we develop our own in-house solutions to help advertisers meet these challenges and opportunities. The privacy-focused future is here, and it’s time to evolve!