How the Topics API works
The idea is to move interest targeting out of the ad network and into the browser. Instead of a third party watching which sites someone visits and building a profile on a server, Chrome would work out broad subject labels from the sites visited recently, hold them on the device, and hand a small selection of those labels to a site that asks — with no identifier attached and no browsing history revealed.
The labels come from a fixed, published list of general subjects, deliberately coarse and deliberately free of sensitive categories such as health, religion or ethnicity. They expire as browsing moves on, so the signal reflects recent interest rather than a permanent profile, and a site only receives labels connected with its own subject area rather than everything the browser knows.
It sits inside Google’s Privacy Sandbox programme, and it inherits that programme’s uncertainty. The direction of the proposals has changed more than once and published sources disagree on what is currently live. Check Google’s own developer documentation before citing its status anywhere a client will read it.
Why the Topics API matters
It is worth understanding because it shows what interest targeting looks like when the browser, not the advertiser, holds the data. The signal is coarser than a cookie-based audience by design. You would be reaching people who have recently shown interest in a broad subject, not people who viewed a specific product on a specific site.
That trade-off is the point rather than a flaw, and it hints at where open-web targeting is heading whatever happens to this particular API. Precise cross-site profiles are becoming harder to build, while contextual relevance, first-party audiences and creative that speaks clearly to a broad interest are becoming more valuable.
Where the Topics API goes wrong
The most common mistake is describing it to a client as a replacement for remarketing. It is not. Remarketing is about people who already touched your business; a broad interest label is closer to an old-fashioned magazine placement.
The second is treating it as settled. Anything written about it may predate a change in the programme, and repeating an out-of-date summary in a proposal is an avoidable way to lose credibility.
The third is expecting it to matter to your account. This is browser and ad-technology plumbing. Unless you build ad tech yourself, you would not switch it on, tune it or see it as a setting in an advertising interface.
What to do about it
Know what it is so you can answer the question, and leave it there. If someone asks whether you are using Topics, the honest answer is that it is a browser-level proposal, not a campaign lever, and that its status is worth verifying at Google’s documentation rather than in a blog summary.
The practical work sits elsewhere. Collect first-party data with clear permission, keep your customer records tidy and connected, and make sure conversions are measured in a way that does not depend on a third-party cookie surviving. Whether Topics or something else ends up carrying interest signals, a business that owns its own audience data and can count its own results independently is in a strong position either way.
