Google Plans New Cookie-Replacement

Google has rethought its plans to overhaul the way targeted advertising works online. It's the latest attempt to balance user privacy and accurate targeting.

At the moment a large proportion of online ads are powered by third-party cookies. That's where an advertiser or ad network operator uses cookies to track a user's online activity and try to figure out their tastes and interests.

This data then powers the ads they see on many websites, which are shown specifically for them. It's good news for advertisers as they can theoretically do a better job of reaching suitable customers, but bad news for users who don't like being tracked.

Google has already said Chrome will start blocking third-party cookies next year. Because it's also a key player in advertising, it's now looking for a replacement targeting method. Without one, advertisers would have to revert to simply posting ads that related to the topic a particular website or the content of a particular page.

User Groups Too Revealing

Instead, Google wants to continue the idea of analyzing user activity and interest, but run it entirely in their browser rather than on remote servers or by third parties. The idea is to find a way for the browser to pass on enough details for targeted advertising without revealing too much about the user's activity.

The original idea, dubbed Federated Learning of Cohorts (FLoC), would have meant users being put into groups of people with a similar interest. Advertisers could then place an ad to be shown to members of this group, but wouldn't have any details about who the individual users were or what they'd done to get in the group.

Originally Google seemed happy with this technology, claiming advertisers could get targeting 95% as accurate as using cookie tracking. Now it's revised the plans following critical analysis that said FLoC still didn't offer enough anonymity. (Source: theregister.com)

Three Topics Sum Up User

The revised system, dubbed "Topics," means the browser simply looks at the names of the websites the user has visited in the past week. It uses a machine learning algorithm to try to figure out the subjects of these sites from the names, then picks out one topic which it thinks is most relevant for the user.

Instead of getting full browsing history or a profile of the user, advertisers and ad networks will simply see the three topics chosen for the past three weeks. They'll have to use this data (along with the details of the page where the ad will appear) to decide what to show them.

Google says this will greatly increase user privacy, particularly as it will let users see which topics have been selected and let them remove any they don't like. (Source: blog.google)

The question now is whether advertisers will feel this offers enough targeting that they are willing to continue paying publishers and ad networks - including Google - at the same rate.

What's Your Opinion?

Does Google's new system sound fair? Given the chance, would you block any information about your browsing or interests going to advertisers? Would you prefer better-targeted advertising rather than seeing less relevant ads?