Not once, not twice, but three times, Google delayed the end of third-party cookies. This is what’s happening

Google has once again delayed the end of third-party cookies in Google Chrome. This cookie cutter, due to happen at the end of the year, ultimately won’t happen as the company faced a number of challenges to meet UK CMA requirements. Below we will look at everything that has happened to date, what consequences it has and what will happen next.

January 2020. Google kicked off 2020 by announcing it would remove support for third-party cookies from Google Chrome. It’s an interesting move, considering Google is a company whose revenue comes largely from online advertising, which relies on cookies on websites. The company intended to discontinue support within the next two years under an initiative called Privacy Sandbox. As Justin Shu, Chrome’s CTO explained:

“Google’s intention is to do this within the next two years. This is our strategy to redesign the standards of the Internet to make it private by default. There is a lot of focus on third party cookies and this is definitely one of them. tracking mechanisms, but it’s just a tracking mechanism, and we call it that because that’s what people pay the most attention to.”

FLOCK. The future of cookies for Google is FLoC (Federated Learning of Cohorts), a technology that has come head to head with other companies such as DuckDuckGo, Mozilla or Brave, and which has been investigated by the European Commission. According to the Electronic Frontier Foundation, it was a “terrible idea.” But what is “FLoC”? And why do we speak in the past tense?

FLoC was a token-based system that allowed user preferences to be identified while preserving privacy, or at least that was the idea. FLoC was based on assigning each user an anonymous cohort, that is, a group of thousands of other users sharing the same interests. Thus, when grouping interests, specific data about the user was not specified and his confidentiality was maintained, Google claims. The group classification will remain in the browser and only the cohort ID will be sent to the algorithm.

Example. Let’s say you love to read and frequent a book club in your city. There are several clubs, depending on the genre: fantasy, romance, adventure… and people go to the one they like best. Now imagine that a publisher wants to promote their new science fiction book. Well, instead of being handed a brochure on the street (and seeing your face, knowing what kind of person you are…), take a stack of brochures to the Science Fiction Lovers book club so people can see them. You fall into the right cohort, but you don’t know the people. It was essentially FLoC and, according to Google, it was effective for 95% of advertisers.

And why was it “terrible”? First, the proposal was not supported by the W3C, although Google held discussions with the World Wide Web Consortium, which is responsible for web standards. According to the Electronic Frontier Foundation, FLoC will “prevent privacy risks associated with third-party cookies while creating new ones. It can also exacerbate many of the worst problems associated with behavioral advertising, such as discrimination or predatory practices.” A simple example is how Google will discourage the creation of cohorts based on sensitive aspects such as race or sexual orientation.

From FLoC to themes. What if we used themes instead of cohorts? That’s exactly what Google thought when it announced Topics, the offering that will replace FLoC. Topics define locally (and without an external server) a number of topics, such as “fitness” or “travel,” that is, interests. If a user visits a website that supports the Themes API, the website receives three themes corresponding to the last three weeks. Site administrators, for their part, will be able to say whether they share this advertising information with their advertisers. The user, for his part, can remove interests from his browser or disable the function completely. As of June 2022, Topics had 300 different topics.

Delay number 1. Now that we know what FLoC and Topics are, and all the problems they entail, perhaps we are less surprised that their availability has been delayed several times. The first time was in June 2021, and at that time the first stage was considered at the end of 2022, and the second stage in mid-2023. The goal was to end the use of cookies at the end of 2023. But…

Delay number 2. In July 2022, Google, which had already received CMA approval for the Privacy Sandbox test, announced another delay. The reason, the Mountain View firm explained, is that advertisers, publishers and other ad industry players have asked for more time to test Privacy Sandbox technologies. This would be a huge change for an industry that turns over billions of euros annually and everything has to be measured down to the millimeter. As such, Google has decided to delay the phasing out of cookies until the second half of 2024. But…

Delay number 3. In mid-April, almost May 2024, Google just announced a third delay. As Google explained in its statement, “We recognize that challenges remain in reconciling the differing views of industry, regulators and developers, and we will continue to work closely with the entire ecosystem.” In addition, the CMA must have sufficient time to review the results of the Privacy Sandbox tests, and participating companies have until the end of June to submit them. Therefore, “given these two important considerations, we will not complete the removal of third-party cookies in the second half of the fourth quarter.”

New date. Given what we’ve seen, the plan is delayed, and essentially Google isn’t closing the door on it being delayed again. In their own words, “If we can reach an agreement, we plan to begin removing third-party cookies early next year.” This means that third-party cookies will remain safe until early 2025. Let’s see what happens next year.

Image | Unsplash edited by Xataka

In Hatak | Europe has found a data security goldmine. One of 2000 million euros per year