Chrome kept cookie choice in its settings.

What changed on 22 April 2025

Google said it would maintain Chrome’s existing approach to user choice for third-party cookies and would not introduce a new standalone prompt. The company said users could continue to choose in Chrome’s Privacy and Security settings. This was a change in the direction of a long-running browser-policy transition, not a finding about a publisher’s revenue or a promise that identifiers would be available to every ad call.

The same post said Chrome Incognito already blocked third-party cookies by default and that Google planned to continue tracking protections there. It also said the Privacy Sandbox APIs could take on a different role, with a future roadmap to follow. The dated record matters because browser-policy plans need the policy statement and the date it applied, rather than a generic “cookies survived” label.

Source notes: Next steps for Privacy Sandbox and tracking protections in Chrome

What a publisher could infer

The useful inference was narrow: a team could record that Chrome would keep its existing user-choice approach at that moment. It could not infer that the same audience would remain addressable, that every buyer would use cookies in the same way, or that a consent-management design had become optional. Browser setting choices, blocking, private browsing, other browsers, partner behavior, and geography still create different delivery conditions.

For a small publisher, the immediate practical work was to name what each tag actually needed: a purpose, a recipient, a data path, a fall-back behavior, and a report that stayed intelligible when an identifier was absent. That inventory is useful regardless of future browser announcements because it separates a commercial question from a technical assumption.

The limits of the announcement

This was Google’s own product-policy statement. It did not test publisher earnings, establish a universal privacy rule, or settle what any regulator, platform, buyer, or browser would require. It also did not mean that a historical cookie-based measurement result could be compared cleanly with a future report assembled under different settings and filters.

A forecast that uses this update should therefore show the dependency it is assuming: browser, region, consent state, identifier availability, and the specific reporting denominator. Calling all of that simply “cookie traffic” hides the conditions that decide whether two numbers are comparable.

A practical question

Which revenue report would become misleading first if identifiers disappeared for a meaningful share of visits: audience reach, frequency, attribution, fill, or a sponsor’s conversion claim? Trace that report back to its inputs. Then decide what can be reported directly, what needs a qualified estimate, and what cannot be claimed. The answer is more useful than a generic “cookieless readiness” label.