And just like that, FLoC was finished. Following a year of industry-wide scrutiny, Google has scrapped its plans to replace third-party cookies with FLoC (or Federated Learning of Cohorts) and announced the launch of a new Topics API for Chrome. We’ve taken time to review the implications of this decision on the digital advertising industry and what it means for growing consumer privacy concerns. At first glance, it is encouraging that the Privacy Sandbox is responding to feedback and seems to be trying to find solutions that fit for as many concerned parties as possible; Topics is indeed a step towards greater consumer privacy protection. But not a large step...
As consumers’ digital behavioural signals will still be stored on the users’ machine and distributed at ad call (albeit with some randomisation), there is still a level of risk that these can be captured and stored, either against authenticated IDs or fingerprinting profiles.
There is certainly an argument that this might be mitigated by the breadth of the Topics themselves as the current spec suggests that they are assigned against a domain/sub-domain and do not provide granular intent or interest signals. While Google may adapt this to focus on a deeper level later, this trade-off toward privacy translates away from accuracy.
This is also echoed in the sparsity of the proposed Topic taxonomy (350 topics), which many have seized upon as a limitation of the solution, but being Google, this is unlikely to remain small for long – the question is, how long will the balance of privacy against the granularity of the behavioural signals remain in the favour of the consumer and not the advertising industry?
As the API will only store 21 days worth of data, it is obviously aimed at providing lower funnel behavioural targeting. A user will have 5 topics for each of the last three weeks and three are randomly selected when presented to a site for targeting, but due to this randomisation of the topics, targeting a high intent user based on this interest will be unreliably hit and miss.
Tracking failure
It seems that the digital advertising establishment’s desire to rebuild cookie-esque functionality within the bounds of the new privacy era leaves advertisers with solutions that are underwhelming compromises – bear in mind that this does nothing to address audience targeting outside of Chrome.
At Fifty, we know that there are far more effective ways to target audiences if our industry can let go of its obsession with tracking. By embracing new and known techniques in novel combinations, we build ways to effectively target advertising that don’t compromise on privacy but embody it; ways that aren’t limited to a single application or a single channel, ways that can provide a human understanding of real organic audiences and that can understand preferences over years and not weeks.
Our cookieless solution FiftyAurora makes waiting until the end 2023 pointless and we would love to show you how they can improve your digital targeting.
Let’s hope that Google’s ongoing consultation starts to lead the industry away from tracking users before we get lumped with cookies for even longer – in the meantime reach out to Fifty here or email [email protected].