Since origin trials started in March, Google’s Federated Learning of Cohorts (FLoC) has sparked debate among adtech companies, advertisers, content creators, publishers, privacy advocates and browser makers. Reactions have been mixed so far: some have welcomed the more privacy-centric approach that FLoC embraces, some have raised concerns whether marketing will remain impactful enough to justify investment in content, while others still have questioned the current plan’s lack of transparency for users.
Can FLoC be declared a success or failure at this point? Of course, it’s too early to say. It’s difficult at present to make a definitive judgement on FLoC and whether it will set the standard in the post-third-party-cookie world or not. Putting all valid criticism aside, however, it’s positive to see that the general discussion is shifting toward a stronger emphasis on user privacy, while still ensuring that online content can be properly funded. Achieving this balance – where users, advertisers and publishers all have their needs met – is what will help us build a more sustainable internet in the long run.
A brief introduction to FLoC
FLoC is a proposed browser standard, created by Google, that promises to offer “interest-based advertising on the web”, but in a manner in which advertisers don’t know the identity of a user. Instead, they are grouped into “cohorts” created using Google algorithms, which are pools of users sharing similar interests and browsing habits. These cohorts are designed to be large enough to preserve user anonymity, while ensuring advertising can be carried out in a targeted way.
When a user visits a website, the Chrome browser tells the site which cohort that user is in, without revealing anything else about their browsing history. No content labels are assigned to these cohorts, meaning it is then up to the website and the rest of the adtech industry to work out the particular interests of that cohort, and use this information to coordinate advertising strategies.
FLoC in its infancy, but opinions aplenty
The data around the performance of FLoC is certainly not sufficient yet. Although Google claimed earlier this year that “when it comes to generating interest-based audiences, FLoC can provide an effective replacement signal for third-party cookies”, this statement should be taken with a pinch of salt.
Firstly, there is not yet any external indication for the reliability of these figures. At the moment, FLoC is only activated on around 0.5 percent of Chrome users in a small number of countries. Furthermore, analysis of this data requires adtech providers to correlate directly with the website to access this data, another barricade that is not always simple to cross.
Secondly, as stated by many other peers in the adtech sphere, Google’s statement elegantly ignores many other marketing use cases that so far were enabled by cookies, such as measurement and attribution, but are not supposed to be solved by FLoC, and which the advertising industry hopes will be solved by other Chrome Privacy Sandbox proposals.
While future proposals, especially FLEDGE – which will offer more granular data but will be restricted to fewer parties – could change the picture, there is a risk to online content and journalism. The simple maths is that the lack of third-party data will make first-party data more precious and attractive to potential ad buyers. The larger a publisher’s audience and the more diverse its content, the more valuable first-party data this publisher has. Potentially, smaller publishers, like blogs and even relatively new independent news sites, don’t have historical data, and will benefit from significantly fewer insights about their audiences than larger publishers like Google.
Is this enough to enable advertising to still fund content? And will smaller publishers have the same opportunities as the big tech companies, who own high-profile websites and can invest in sophisticated machine learning-based decision making? These questions still remain and will hopefully be addressed as part of future Privacy Sandbox proposals.
If you do something positive for users, why not involve them?
As hinted above, one notable concern is the lack of transparency and communication with the data subject, the user. Currently, FLoC doesn’t require user consent to operate, and doesn’t allow users to review, modify or correct the data collected.
This is not a small issue. It's most likely that FLoC is not currently being tested in the European Union due to doubts regarding its compliance with GDPR. The regulation requires the controlling party to ask for consent, make the data available for review and allow the data subject to ask for its modification and deletion. None of these options are currently available for the user.
With all that said, there is definitely a positive side. At the very least, it’s stimulating to see the discourse evolve towards a ‘sustainable privacy’ that respects users’ rights while being mindful of the need to fund content. Recent discourse on GitHub repositories and in vision papers have taken the spotlight and become a discussion topic for several stakeholders - including the non-technical ones from the advertising industry.
We should all aspire to a future online ecosystem that promotes user choice, while incentivizing content generation and journalism, and where innovation – and privacy innovation in particular – is feasible, even if not introduced by the big tech firms.
With all of this in mind, it is important to provide users with tools that ensure their privacy is respected, while making locally-processed anonymized data available for content funding. It is crucial to make sure that users get full transparency and control of their data, despite the fact that this data is anonymized.
All products offered to users should be straightforward and transparent if they wish to make any use of data. We believe that no solution can really last long without respecting users’ privacy above all. Only with that will we gain their trust.
We hope that some of these ideas will find their way into the Privacy Sandbox proposals. On the whole, we are encouraged and hopeful that a positive, balanced web economy can be achieved in the long run.
Rotem Dar, Director of Media Operations, eyeo