Topics API (Application Programming Interface)

The Topics API is a browser feature in Google's Privacy Sandbox that lets a website infer a user's coarse interests for advertising without tracking that user across the web. Instead of following a person from site to site with a third-party cookie, the browser itself observes the sites someone visits, assigns a handful of broad interest topics from a public, human-readable list, categories like fitness or travel, and shares only a few recent topics with sites that ask, through the application programming interface. The topics are computed and stored on the user's own device, only broad categories are exposed, sensitive subjects are excluded, and older topics expire, so the signal is far less identifying than a cross-site cookie. The aim is to keep interest-based advertising possible while removing the pervasive individual tracking that third-party cookies enabled.

The Topics API arrived as part of a larger, much-delayed effort to change how the web handles ad targeting, and the honest picture is that the ground has shifted repeatedly. Google originally proposed phasing out third-party cookies in Chrome and building Privacy Sandbox APIs, including Topics, to replace some of their functions. That timeline slipped several times, and Google ultimately stepped back from simply removing third-party cookies, moving instead toward approaches that leave user choice in the picture rather than deleting the cookie outright. So the Topics API should be understood as one privacy-oriented proposal among several, developed against a backdrop of an evolving and contested plan, not as a settled, universal replacement that swept cookies away. Its real-world weight depends on decisions and adoption that have kept changing, which is exactly why any description of it has to be dated and provisional.

The Topics API is best understood against the two things it relates to: FLoC, which it replaced, and third-party cookies, which it was meant to reduce reliance on. FLoC, or Federated Learning of Cohorts, was Google's earlier Privacy Sandbox proposal for cookieless interest targeting. It grouped users into large cohorts of people with similar browsing and exposed a cohort identifier to sites. It drew heavy criticism, that a cohort identifier could aid fingerprinting, that it might reveal sensitive traits, and that grouping users this way carried its own privacy and fairness problems, and Google abandoned it. The Topics API was designed as the successor, exposing a few broad, curated topics from a public taxonomy rather than a cohort identifier, with sensitive categories filtered out, in an attempt to answer those criticisms.

Third-party cookies are the older technology both proposals were reacting to. A third-party cookie set by a domain other than the site you are visiting can follow you across many sites, building a detailed, often individually identifying profile for ad targeting and measurement, powerful for advertisers and long the backbone of programmatic targeting, but corrosive to privacy. The Topics API differs sharply in intent: it deliberately exposes only coarse interest categories computed on the device, not a cross-site identifier, so it cannot rebuild the granular, person-level tracking a cookie enables. That is the whole trade-off. Third-party cookies offer rich, precise, privacy-invasive targeting; FLoC tried a cohort-based middle path and failed; the Topics API offers a deliberately blunter, more private signal. Which of these actually dominates the market has depended on regulatory pressure, industry adoption, and Google's own shifting decisions.

For marketers, using the Topics API well starts with keeping perspective on where it sits. Treat it as one signal in a privacy-conscious toolkit, not as a like-for-like replacement for the individual tracking third-party cookies allowed, because it is intentionally coarser and will not reproduce cookie-level precision. Because its status and adoption have moved repeatedly, build an advertising approach that does not depend on any single mechanism surviving unchanged: invest in first-party data and consented relationships, in contextual targeting that needs no cross-site identity, and in measurement methods that tolerate less granular signals. Where the Topics API is available and relevant, use it for the broad interest-based reach it is designed to provide, while understanding that it deliberately trades precision for privacy.

The failures come from treating a moving, provisional technology as fixed. Assuming the Topics API is a full replacement for third-party cookies sets up disappointment, because it is deliberately less precise and was never meant to match cookie-level targeting. Building a strategy that hinges on one API surviving a timeline that has already changed many times is fragile. Confusing the Topics API with FLoC misreads what it does, since FLoC exposed a cohort identifier and was abandoned, while Topics exposes a few curated categories. And ignoring the honest uncertainty around third-party cookies, in either direction, leads to bad planning. The discipline is to describe the Topics API accurately and provisionally, lean on first-party and contextual approaches that do not depend on it, and revisit the plan as the ground keeps shifting.

The Topics API is part of Google's Privacy Sandbox initiative and succeeded the abandoned FLoC proposal as a more private approach to interest-based advertising.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where topics api (application programming interface) is a core concern: