Google’s FLoC Proposal
Why are cookies on Chrome going away, creating this need for new targeting solutions?
In mid-2019, Google announced their initiative called the “Privacy Sandbox” in order to foster greater web privacy standards for users; a solution to keep up with increasing online data privacy regulations like we’ve seen with Apple. By blocking onsite cookies, Google is removing data that’s collected from unknowing people when they visit web pages. This is an effort to create a web environment where user data is collected with full consent. The tradeoff for the loss of cookies is an expected negative impact in advertiser targeting integrity, as advertisers have long relied on cookies to gain an understanding of audiences based on their online activity.
How does the proposed initiative work?
Providing browser access to the web with increased privacy regulations and relying on large-scale ad revenue to support their business puts Google between a rock and a hard place. The Privacy Sandbox initiative is designed to comply with the former, and the FLoC proposal is meant to be a solve for the latter. Ultimately, Google wants to maintain a leadership position for advertiser solutions but not interrupt ad revenue.
With the core issue of cookies gaining a 1:1 understanding of an individual’s browsing data, the FLoC approach removes the 1:1 aspect of identification without losing affinity signals advertisers want to reach high-intent audiences. What this proposal seeks to accomplish is being able to take similar Chrome browser affinity signals and place them in a cluster (or cohort) to inform ad calls that there’s a high intent pool of interest-based users, as opposed to a high intent interest-based individual that advertisers with matching affinity target layers could serve to.
Google’s self-graded initial testing of this FLoC approach states that advertisers should expect this method to be 95% as effective as 1:1 cookie-based advertising in terms of conversions to dollars spent. Naturally, there is some variance based on the specific interest being targeted and size of the audience cluster. Google expects to gain additional data on this approach’s effectiveness following public testing in March 2021 and initial advertiser rollout within Google Ads the following quarter.
What are the privacy hurdles this solution needs to prove it can clear?
While testing is on the horizon, there are components of this proposal that need to check the compliance boxes in order to be a feasible advertiser targeting method, complying with both GDPR and Privacy Sandbox standards. The nature of the FLoC proposal itself presents questions about how it will solve for these standards. It will need to prove it can satisfy key privacy standards while maintaining effectiveness as an ad platform:
- Ensure the intent signals used aren’t able to identify an individual when placing them into a cohort
- Ensure that an individual isn’t 1:1 identified as “belonging” to a cohort category
- Ensure that cohorts don’t reveal sensitive information about their category
What is our takeaway at present?
Prior to the FLoC approach being implemented with subsequent results to learn and benchmark from, we can’t place an immediate reaction on the effectiveness of this cookie-less solution. If this solution is in fact almost as effective as cookie-based advertising, then we can hope the negative impact of cookie removal mandates is minimal. That said, this FLoC proposal shouldn’t be considered to be the end-all-be-all solution that solves all post-cookie targeting problems, nor should it be ignored that more solutions should be expected to emerge. A perfect example? The TradeDesk’s Unified ID 2.0, which we discuss in another blog post.
In the meantime, it is critical to develop strategies that get ahead of regulations. Strategies that value and retain existing customers — specifically thoughts that are geared toward developing first-party datasets and utilizing first- or zero-party data for targeting efforts — can expect to see less impact from these changes in privacy compliance.