Google yesterday threw its weight behind the Federated Learning of Cohorts (FLoC) API, a tool developed within its Privacy Sandbox designed to replace third-party cookies. Google is bullish on FLoC, saying it’s 95 percent as effective as third-party cookies when used for interest-based targeting.
But is the wider industry ready to embrace FLoC in place of third-party cookies? And will the ‘cohorts’ model of targeting really allow privacy-preserving targeting that still works well?
I don’t often find myself in agreement with Google on issues of digital ad targeting, but I applaud the company’s recent move to embrace generalized cohort targeting as a replacement to individualized ad targeting through cookies and similar technologies.
Cohort based targeting is much better for users. it dispenses with the collection and storage of so much personal information. It is much more transparent. Users are either in cohorts – believed to share an interest with others in the cohort – or they are not. It is much, much less complicated. The digital ad industry has created a massive programmatic and identity industrial complex that taxes everyone in the business.
Cohort targeting is simple and should mean that a massive amount of ad targeting technology overhead can go away. Finally, cohort targeting works the same as personalized targeting for all intents and purposes. Google’s research puts it at 95 percent. Losing a tiny fraction of effectiveness to protect people’s privacy and save us from the digital infrastructure tax sounds like a great deal to me.
Oh, and by the way, cohort targeting is where the online ad business started its targeting work back in the late 1990’s …. what’s old is new again.
Privacy Sandbox is a false good idea that pretends to protect users while in fact stripping them of their right to have a say in how their personal data is used.
Privacy Sandbox is also a way to further cement Google’s dominant position in digital advertising. After years of harnessing cross-site tracking technologies to fuel the largest personal data collection engine in history, they want to stop the clock and limit data usage for everybody else. Google wants to prevent publishers from leveraging their legitimate relationship with users to create data assets that would enable them to compete with them.
Privacy Sandbox is both unethical and anti-competitive.
Google’s post certainly caught most people off-guard. While FLoC has been circulated and discussed in various forums for over a year now, it hasn’t been seen as the only solution. This and the other bird-based acronyms are still seen as viable solutions, as we expect brands and publishers to be needing a cocktail of deterministic, probabilistic, and contextual solutions to scale their Identity, in and out of walled gardens, going forward.
So, what just changed with this announcement? As far as we know — nothing. While advertising industry groups (i.e. PRAM, Rearc) and internet consortiums (i.e. W3C) have been reviewing options, nothing has been finalised. At this point, it seems that Google (Chrome and Ads divisions) have determined this is the best way forward, for them and the industry. That does not feel like collaboration for something so deeply entrenched in the ad ecosystem.
Firstly, FLoC does not solve everything. It covers an aspect which is the interest-based segmentation. It also appears that Google determines the type of segments and granularity that will be created. The methodology is not externally known, reviewed, or finalized. So, it feels that Google is leveraging its power here.
Secondly, this is something that is “adopted” by Chrome. The other browsers are not following this methodology. As such, there can now be competing solutions from other platforms which further creates challenges in holistic solutions, measurement, targeting ,and attribution. So, while privacy might be improved many other things became immensely complicated.
Thirdly, the value of walled garden’s data (of which Google is one of the largest) will continue to be the prized data. This does not degrade their data solutions and now increases that value, continuing to support any spending occurring within their garden but not allowing brands to leverage any intelligence earned from their ad spending elsewhere, restricting better personalisation for consumers.
And to add more, it doesn’t fix everything. There are other workarounds that will be created to prevent user identification (ex. masking/altering IP addresses) and finding other solutions for retargeting solutions and look-alike modelling.
There needs to be better dialogue and industry consensus vs. Google selecting its own preferred approach.
Ari Paparo, CEO, Beeswax
It is exciting that Google has made some progress on the sandbox proposals and published data on effectiveness. Floc seems like a viable path forward for interest-based advertising, though the measurement of results remains a key question.