Finding the successor to the tracking cookie is difficult. Google’s “Federal Learning of Cohorts” approach, FLoC for short, is not really popular with browser manufacturers and websites . Also browsers based on the Chromium engine, such as Vivaldi , do not want to allow FLoC. Google has a market share of almost 70% with its in-house Chrome browser ( Source Statista April 2021 ) and therefore a good user base, but this could also crumble if other browsers offer more data protection with the same level of convenience.
The Documentation on FLoC leaves many questions unanswered. FLoC should use the browser history (i.e. the pages visited) to classify the user in a so-called cohort. Users with similar browsing history end up in the same cohort. This is identified by an ID. When visiting a website that displays advertising, this ID can be obtained in order to display targeted advertisements.
These cohorts should include several thousand users, to make it impossible to identify a single user.
The browser assigns the user to a cohort. For this purpose, the browser receives a cohort model, which then uses algorithms to calculate the cohort identifier from the browser history. The calculation should only take place in the user’s browser, the history is not transferred. Visiting other websites can of course change the assignment to a cohort.
Where is the problem?
Until now, tracking technologies could only track users where advertising material or tracking pixels were integrated – and only if this technology ran on their own platform. The entire browser history is analyzed with FLoC – even websites that do not use tracking techniques are now part of the cohort calculation via FLoC.
Websites can block the recording for FLoC via a special header, but this requires active action and a little technical knowledge ( WordPress could soon send this header by default).
The profiles with FLoC could therefore be significantly more precise, i.e. the previous cookie solution – simply because the entire browser history flows into the profile.
Of course, apart from the browser manufacturer, nobody can say how big such a cohort is or how it is calculated. For a browser manufacturer who earns their money with advertising and targeting at the same time, a smaller cohort size should be an advantage.
There is also little or no information about the controllability by the user. Can I actively influence which pages are considered for my cohorts? Can I see which pages are relevant for the calculation? Is there a way to reset my cohort?
Privacy and advertising?
Both could basically work together – but in recent years the advertising industry has made use of all the possibilities that technology offers on the web.
Anyone who runs a poster campaign knows where their poster is and how many people see it every day (roughly) – but they don’t know how many of them are looking for their business, how many are looking at products, subscribing to the newsletter or actually placing an order.
These key figures are available in online marketing and are used extensively to optimize campaigns and to create reports for customers. So customers are used to being fully informed about the effectiveness of their advertising measures. The industry’s fear is that poorer measurability will also reduce advertising spend.
Recognition of the user also plays a significant role in the area of targeting and especially retargeting. With FLoC, these areas could at least rudimentarily continue, depending on how precisely the cohorts are formed. The finer the cohort, the worse the data protection.
Search marketing or content marketing show that effective online advertising is also possible without user targeting. Targeted advertising based on the search term or the environment in which the advertisement is displayed. No user data, cohorts or cookies are required for this.
There are enough solutions for measuring the performance of a campaign that advertisers can integrate directly and in compliance with data protection regulations on their website or shop. I would be happy to support!
