Federated Learning of Cohorts is dead, Topics are the new Cookies. Is Google's latest solution to the privacy issue designed to benefit them, or publishers and advertisers?
Google's announcement this week of the planned introduction of the Topics API probably didn't come as that much of a surprise to anyone in publishing or advertising that has been following the Federated Learning of Cohorts saga.
Almost everything that we read about FLoC seemed to indicate a lack of conviction on the part of the Big G. Of course, Google will say, fairly, that the whole process of replacing cookies is just that - a process. A process with consultation. However, the very fact that FLoC was trailed in public means that Google did think it might fly. The FLoC has now been grounded forever. Clearly the feedback from industry and regulators wasn't favourable.
Google knows it is in the hot seat on this one. However much it uses carefully selected neutral language and the suggestion of impartiality towards the issue of balancing privacy and commerce, its dominance of search and the largely unaccountable and hugely profitable power it wields over every aspect of business online in the Western world is under scrutiny like never before.
A good indication of this, and a feather in the cap of a British government unit, comes in the form of Google's rare acknowledgment this week of co-operation with the UK's Competition and Markets Authority (CMA) on the proposed replacement of cookies.
It was the CMA, don't forget, that produced a beautifully comprehensive analysis of the damage the dominance of the US tech giants is doing to both business and consumers in every market where they are active. The CMA's report was so good, it was used as the basis for action by US regulatory bodies looking to limit the power of Big Tech, particularly the Google/Facebook duopoly. All credit to them.
So what of Topics? How does the system work? Here's what Google itself has said this week:
There are two main parts to Topics. First, the API labels each website with a recognizable, high-level topic. For example, the browser would match a sports website with the topic "Sports". Then, the browser collects a few of the most frequent topics associated with the websites you’ve visited. These topics are then shared (one new topic per week) with the sites you visit to help advertisers show you more relevant ads, without needing to know the specific sites you’ve visited.
When you visit a participating site, Topics picks just three topics, one topic from each of the past three weeks, to share with the site and its advertising partners. Topics enables browsers to give you meaningful transparency and control over this data, and in Chrome, we’re building user controls that let you see the topics, remove any you don’t like or disable the feature completely.
Google has even provided a diagram comparing devious, underhand third party cookies with clear, concise and cancellable Topics:
According to the documentation we're looking at around 350 topics to start off with, with examples given of "Fitness" or "Travel and Transportation". The intention is to expand the number of topics: "Chrome will share the list of topics publicly and expects it will evolve over time with the ecosystem feedback."
Google says that "Chrome intends to experiment with the Topics API in 2022. The technology is still in the early stages of development. The results of the trials and feedback from the web community will inform the timeline."
Glide Publishing Platform knows a great deal about taxonomy-driven systems such as Google's proposed Topics. GPP itself is built around a taxonomy system for content, based on our shop-floor experience working in publishing. We know how important order is - but we also know how complex taxonomy systems can get. Classification is not a simple game.
We have some distinct reactions to Google's proposals.
The first is why does a business that will benefit from the data it retains get to dictate what the rest of us can use. This is the fundamental issue facing regulators and legislators. Google is both poacher and gamekeeper.
The second is more specific - Topics sound broad enough that precise identification of an individual should be very difficult, yet as the list of Topics expands, and coupled with location data, a very good idea of who you are could possibly be assembled by a large advertising network.
Lastly, the whole Privacy Sandbox thing reminds us of the way AMP was presented as an industry initiative, not a Google one. And we all learned the truth of that.