Analysis: Will Google's FLoC proposal be yet another walled garden?

Just last week, Google shared that its Federated Learning of Cohorts (FLoC) proposal, which is part of the Privacy Sandbox introduced by Chrome, which it claims “can provide an effective replacement signal for third-party cookies”. This comes a year after it announced that it will be phasing out third-party cookies, following in the footsteps of Safari and Firefox.

The FLoC application programming interface (API) was first proposed in January 2020 and Google said in a recent blog post that new data shows its tests of FLoC to reach in-market and affinity Google Audiences will enable advertisers to see at least 95% of the conversions per dollar spent when compared to cookie-based advertising. The specific result depends on the strength of the clustering algorithm that FLoC uses and the type of audience being reached.

Google said FLoC puts forth a new way for businesses to target consumers with relevant content and ads by clustering large groups of individuals with similar interests.

According to Google, this approach “effectively hides in the crowd” and uses on-device processing to keep a person’s web history private on a browser.  Chrome plans to make FLoC-based cohorts available for public testing through origin trials with its next release in March and it expects to begin testing FLoC-based cohorts with advertisers in Google Ads in the second quarter of this year. 

The FLoC API relies on a cohort assignment algorithm, meaning a function that allocates a cohort ID to a user based on their browsing history. To ensure privacy, Chrome requires this cohort ID to be shared by a certain amount of distinct users. According to Google in a whitepaper, a cohort should comprise users with similar browsing behaviour and the cohort ID should prevent individual cross-site tracking.

The more users share a cohort ID, the harder it is to use this signal to derive individual user's behaviour from across the web.

While this benefits privacy, the trade-off is that a large cohort is more likely to have a diverse set of users, thus making it harder for advertisers to use this information for fine-grained ads personalisation purposes. 

An ideal cohort assignment is one that generates cohorts by grouping together a large number of users interested in similar things. Also, as it is simple for an individual's cohort to be computed, Google said this allows for the FLoC API to be implemented in a browser with low system requirements.

what is floc fig 1

Will FLoC further solidify the walled gardens?

At first glance, the latest development concerning FLoC might seem impressive, given advertisers can expect to obtain at least 95% of the conversions per dollar spent compared to cookie-based advertising. However, Gowthaman Ragothaman, CEO of Aqilliz, said even though FLoC is open sourced, it is still owned and managed by Chrome and built for the Chrome browser only. This is compared to cookies which has no one single authority that manages or owns it. Also, while FLoC claims to achieve 95% conversion, it is done in a privacy compliant manner, which is the price advertisers now have to pay for privacy over personalisation. 

That said, if FLoC does become mainstream, Ragothaman said the industry will be entering "a whole new world of digital marketing", adding:

The open web as we see it will gradually subsume into a bunch of walled gardens, each having its own methodology to safeguard privacy and having its own strike rate on the extent of personalisation.

"Brands have always had very clear business outcomes, which either increase penetration or increase consumption. All other metrics are intermediary variables that were stitched up to, and somehow get attributed to these two overall business outcomes," Ragothaman said. He foresees a situation where the industry will have to find few more surrogates to connect metrics from different walled gardens to eventually arrive at the overall productivity from marketing ROI.

Hence, Ragothaman said it is now time for marketers to really focus on a structured approach to cross media measurement, and they can start with cross-garden measurement, at the least among digital platforms. Since Google's announcement regarding third-party cookies last year, measurement firms such as Nielsen and Comscore have launched their own cookie-less targeting solution. Nielsen's ID resolution system, for example, claims to support cross-media measurement, including digital, in a cookie-less world. Concurrently, Nielsen is also working with The Trade Desk and the industry on its unified ID resolution system, which has also received support from other adtech firms in the industry including PubMatic, Magnite, Criteo and LiveRamp. 

Meanwhile, the development of FLoC, along with the deprecation of third-party cookies and Apple's Identifier for Advertisers mean that brands, media platforms, agencies are "coming to a gradual acceptance" that targeting contextually rather than individually will be the way of the future, Bobbie Gersbach, planning director, Asia Pacific for M&C Saatchi Performance said. 

However, one of the first challenges which comes to mind from a media buying perspective is that buying into a certain context, for example automotive, business and finance has traditionally come at a premium. Hence, Gersbach said achieving efficient scale can be a challenge.

"In theory, what FLoC could provide is a way of targeting interest-based cohorts, across a broader set of online inventory to improve efficiency. That being, a cohort identified as a group rather than via individual identifiers to protect user privacy," she added.

Metrics in a post-cookie world

While Google has said it is developing measurement APIs associated with the Chrome browser or FLoC, which replicate elements of existing metrics such as post-click events, Gersbach said it is likely that this will be at an aggregate view. According to her, this could potentially be similar to Apple’s SKAdNetwork, which provides an aggregated attribution solution, albeit with time-based and granularity-based limitations.

"How actual performance for advertisers will be affected remains to be seen. Google’s initial testing suggests that advertisers will see at least 95% of the conversions per dollar spent when compared to cookie-based advertising. However, this testing occurred across Google’s own ad products and is probably not representative of the overall media mix," she explained.

Agreeing with Gersbach on contextual targeting is Integral Ad Science's MD, Southeast Asia Laura Quigley, who said in a post-cookie world, marketers will substitute audience data with contextual intelligence. 

The new metric will be time-in-view which is a proxy for attention - a metric that will have more marketers turning to understand the engagement levels of the cohorts. 

"Rather than having to support all the data management and privacy regulations that go along with capturing and leaving personally identifiable information, many marketers will switch to advertising in environments that are contextually relevant and use that as a proxy for the audience," she explained. At the same time, Quigley said evolving mediums such as connected TV will only accelerate the marrying of the content, which is interest-based, with the relevant ads.

"In the case of FLoC, conversion will seem like a strong metric to earn the trust of advertisers. Nothing concrete is set and several options are being considered. In the report, it recommends implementing site-wide tagging or using Google Tag Manager now to prepare for the future of how that mechanism might work," she added.

While the development of FLoC is still in the early days, she said this still seems like a good step in the right direction for consumer privacy as well as enabling the brands and publishers to reach the right audience in the right environments. According to her, the efficiencies of FLoC will depend on marrying privacy and monetisation with optimal cluster size, a strong clustering algorithm that the API uses, and the type of audience being reached. "It will be interesting to see how the algorithm helps uncover newer audience interest groups," she said.

Photo courtesy: 123RF

Related articles:
Google to scrap third-party cookies: Will it crumble parts of the digital ad world?
Build a data reservoir: Regional marketers on Google cutting third-party cookies
Are you truly prepared for the cookie-less world?