top of page

Google's New "Topics API"

Consumer privacy has been a hot topic in the last five years. The European Union implemented its General Data Protection Regulation (GDPR) in 2018 for how websites collect and track personal information from online visitors. And, as we’re well-aware, that same consumer privacy concept translated over to the United States, when the state of California enacted the California Consumer Privacy Act (CCPA) the same year. Since then, several states are following up with similar regulations.

Meanwhile, individual tech companies have also formulated ways to protect individuals’ privacy through updates to their platforms and changes in the way they track consumers, as well as the way they do business with advertisers. This includes Google’s plan of eventually getting rid of 3rd-party cookies that track users’ behavior through their browsers, perpetuating the notion of requiring organizations to gather user consent as people surf the web.

With 3rd-party cookies slowly crumbling from multiple browsers and apps (including Google Chrome, Apple Safari, Microsoft Internet Explorer, and Mozilla Firefox), online users receive more confidentiality as their privacy is held at a higher standard. Google’s Privacy Sandbox gains more traction. Which is all great! However, brands and agencies are left scratching their heads – looking for new ways to personalize their advertising efforts, report using cross-site attribution, and track campaigns in general so they can optimize results effectively. What will replace the accurate targeting advertisers once had to track users across sites, bringing them to conversion? Targeted advertising is a HUGE business. So, Google must come up with a way to protect user privacy, while also providing a reasonable way advertisers can track users effectively as they market using their platform. The tech giant needs to maintain a steady balance between the two end users, or marketers will eventually find another workaround – potentially taking massive business away from Google.


This pushed the tech company to come out with their first solution called Google FLoCs – or Federated Learning of Cohorts. The FLoC (like the abbreviation looks and sounds) involved “flocks” or segments/pools of users that are assigned to a cohort – or group of other users. Instead of a marketer having the ability to track a user’s browsing data, that user’s data would be stored within its browser. All users would be assigned to a cohort, based on their browsing history, and an API would allow the browser to connect that cohort to the sites visited by these users. The privacy lied in the fact that only the cohort identifier could be retrieved from the browser – not the individual browsing history. In addition, users were reassigned to different cohorts every week.

Now this all sounds good upfront (minus the fact that view-thru conversions would not be trackable since 3rd-party cookies are needed for this, as well as a few other challenges). However, the main concern was that FLoCs could also be combined with other more evasive ways of advertising like digital fingerprinting – a way to identify individuals based on their browser history/setup and device settings. This would cancel out the entire reason Google was beta-testing FLoCs in the first place, as they are trying to prioritize the privacy of online users, not create a way that others could potentially combine methods to provide them with even more information on individual users than before FLoCs.


So, Google scratched the FLoC idea. It was time to figure out what would work better. The most recent solution in the Privacy Sandbox for targeting audiences while protecting the privacy of users – they’re calling Topics API. This idea involves pulling a user’s web history from the Chrome browser and placing them in up to five Topics (ie: books, auto and vehicles, fitness, rock music, etc.). Marketers can use these Topics to target consumers with relevant ads. New Topics are defined every week for up to three weeks, then after that, old Topics are deleted.


There are some upsides to this new solution. For marketers, Topics API is a way to provide strategic insights on consumer behavior, while looking at an overall larger audience-driven marketing strategy. Additionally, Publishers know how their sites fit into the Topics universe. And consumers feel good about their “gender” and race” being anonymized, as well as being more in control of their information online. Google is also going to allow users to opt out of Topics API and remove Topics information from their profile if they wish. It seems Google will be gathering its consumer behavior data from a website’s hostname alone, vs. from the full context of the page’s URL. It will take this data and segment them into 350 Topics. Of course, more testing and work needs to be done before full implementation, but this is what it’s looking like for now.

The downsides include the fact that even though you may group a user who visits Target or Amazon online into a “Shopping” Topic, there are still key details missing that could be important to marketers. For example, knowing a user visited provides much more context into the URL that advertisers would find extremely useful when targeting with specific ads that would produce conversions.

Looking ahead, it seems like contextual targeting is something that will continue to be examined. Topics API is pretty similar to our old friend contextual targeting, only with some additional browser information. But there’s one thing that holds true, owning your own data will be the big winner in all of this down the road. Once the 3rd-party cookie crumbles, First Party Data will be crowned king.


bottom of page