Chrome Turns Off Third-Party Cookies for 1% of Users

What Changed

Google announced the test in December 2023 and started it on January 4, 2024. Chrome began restricting third-party cookies by default for 1% of its browsers worldwide, on desktop and Android, chosen at random. Google called the feature Tracking Protection.^1,2

People in the test were told when they opened Chrome. If a site seemed broken without cookies, for example by repeatedly reloading, Chrome could prompt them to turn cookies back on for that site, and they could also do it themselves from an eye icon in the address bar, for 90 days at a time.^1,2 Sites that needed cookies for things other than advertising, such as sign-ins or embedded services, could apply for extra time through a deprecation trial.²

Google presented the 1% as the first step to a full phase-out from the third quarter of 2024, subject to addressing any remaining competition concerns of the UK Competition and Markets Authority (CMA).²

The test also had a second purpose: measurement. With the replacement APIs now in Chrome, Google labelled slices of Chrome traffic so that ad tech companies could compare results with cookies, without them, and with the Privacy Sandbox replacements, and share their findings with the CMA.^3,4

Why It Mattered

It was the first time Chrome actually switched third-party cookies off for real users, after two delays. For most advertisers the direct effect was small: 1% of Chrome traffic showing up as new or unattributed visitors. The results of the testing mattered more.

It was also a preview of what would and wouldn't break. Conversion tracking on an advertiser's own site mostly relies on first-party cookies: Google's conversion linker stores ad-click information in cookies on the advertiser's domain, and Meta's pixel stores its browser ID and click ID the same way.^5,6 What depends on third-party cookies is anything that follows a person across other sites: retargeting audiences built by ad networks, frequency caps across publishers, and conversions credited to display ads people saw elsewhere. Those were the functions the Privacy Sandbox had to rebuild, and an industry review by the IAB Tech Lab in February 2024 found several of them, including frequency capping and audience building, unsupported or degraded.⁷

The test results weren't encouraging. In June 2024 the ad tech company Criteo reported that, in an eight-week test across its advertisers and publishers, publishers that had fully adopted the Privacy Sandbox would lose about 60% of their revenue on average compared with third-party cookies, and that ads took longer to appear.⁸ The CMA's own summary of the industry tests, published in June 2025, found a smaller but still large effect: publisher revenue per impression was around 30% lower without third-party cookies, even with the Privacy Sandbox available.⁴

By then the plan had changed. Google delayed the full phase-out in April 2024, partly so the CMA could review these results, and three months later dropped it.