Thoughts on Google's Federated Learning of Cohorts (FLoC)
Updated: May 7, 2021
On January 25th Google has shared progress on eliminating 3rd party cookies.
The proposed approach is Federated Learning of Cohorts (FLoC). While there is a lot to process, I wanted to share some thoughts on the approach and potential implications for user privacy, digital advertising and other areas of digital marketing.
In this post:
What is Federated Learning of Cohorts (FLoC)?
FloC is a new potential way to target users online without using 3rd party cookies issued by Google, Facebook or any other vendor. FLoC is exploring an idea of targeting users not based on a specific interest of one user (3rd party cookie), but rather an interest of a group of over 1,000 users who have similar browsing history.
FLoC vs 3rd Party Cookie, What's Different?
According to Google, FLoC is different because it "...effectively hides individuals in the crowd” and uses on-device processing to keep a person’s web history private on the browser." In other words, FLoC proposes to move data processing to the User's browser, as opposed to sending a third party cookie to User's browser.
Inside the browser, based on criteria defined in Chrome's FLoC a User will be categorized into different audiences, say "Car Enthusiast" or "Coffee Lover" together with at least 1,000 users with a similar behaviour.
This means that from perspective of an advertising vendor you will now be a "Coffee Lover" (or another audience for that matter), as opposed to a holder of their cookie that they can read information from from.
From my perspective, it's a compromise to balance between targeting User's based on their interest and still protecting their privacy. Google has long had a number of audiences available for targeting through Google Display network, say In-Market Audiences, Affinity Audiences etc.
Potential Implications of FLoC for Advertising Vendors
The thinking process here is - once FLoC is rolled out, Chrome (which holds ~66% of global browser share) may restrict or limit the third-party cookies altogether.
Now who this can really impact are vendors and publishers. From the vendor perspective, many rely on shared ID that they use to identify a User who comes to a website.
While from publisher perspective Google's study has shown Visits without a cookie resulted in about 52% less revenue for the publishers.
Potential Implications of FLoC to Targeting
According to Google, they were able to achieve "at least 95% of conversions per dollar spent" on Google Ads using this approach to targeting.
Let's imagine, a group of users researching a purchase, say, furniture or a roof. I would assume that the consumer journey to purchase be longer in this market. So there will be more than enough time to collect the data and group users under a Cohort ID.
Many of our small business clients rely primarily of Google Ads and SEO to generate a large portion of their revenue. So I don't think this change will have much impact on their results. At least, in most cases.
How Might FLoC Change Online Advertising/Marketing?
We don't know yet. But we can make a couple of educated guesses:
Facebook - may try and shift to keeping more users inside their platform to compensate for targeting precision from 3rd party cookies. But it's not new, facebook has been trying to keep users inside the app for years.
Advertising Vendors - coming continue coming up with workarounds.
Businesses may shift focus towards SEO which does make sense considering it's by far the highest ROI channel.
I hope you find this helpful. Reach out if you have any questions!
Sasha Matviienko is an Analytics & SEO Consultant in Analytics & Digital Marketing consultant, founder of growth360. Sasha has 10 years of Digital Marketing experience from Paid Search, SEO and Display to advanced Data Analytics and Conversion Optimization experience working with clients of all shapes and sizes in Toronto, Canada and Worldwide.