Let’s Change “Topics”: Announcing The Post-Cookie Targeting Analysis Study

Does anyone remember what happened on March 3, 2021? Since the onset of the pandemic, time has simultaneously slowed down and sped up in mind-bending ways, to the point that eleven months can sometimes seem like eleven years. Couple that with the fact that memories are short in this fast-paced industry of ours, and it’s not hard to see how one can forget that this was the date that Google announced plans to add more substance to their Federated Learning of Cohorts (FLoC) initiative. Or, in their words, chart a course towards a more privacy-first web.
As per usual with Google’s self-serving chess moves in the privacy game, the March ’21 announcement elicited panic from most of the industry and eye-rolls from some of the more seasoned (i.e., cynical) among us. Apple had long since pulled the plug on 30% of browser personalization with Intelligent Tracking Prevention (ITP), and those of us heavily involved in the world of data & identity were more than familiar with the capricious nature of the browser gods. At the time, we surmised that the announcement was a thinly disguised strong-arm move to keep marketers and advertisers from migrating away from its stack due to post-GDPR data restrictions, which is exactly what one would and should expect from Google HQ.
Fast forward to last week’s announcement of Google Topics, and it is clear that the more things change, the more they stay the same. Despite earlier claims of FLoC being 95% effective (provided, of course, that advertisers spend the entirety of their budget within the confines of Google’s ecosystem and O&O inventory), the Google Privacy Sandbox has a new (and arguably much worse) plan to continue their media hegemony. Earlier concerns regarding fingerprinting with the FLoC ID have been addressed by associating users in broad segments with 350 preset “Topics”. While a positive move in terms of consumer privacy, Topics presents advertisers with even greater challenges; the granularity of targeting is significantly diminished, the ability to target niche segments is lost, and with only the top three topics delivered in a randomized fashion, there is a risk of high-value users remaining uncaptured. Further complicating matters is the admitted 5% possibility that a user might be associated with a topic at random and the frustratingly slow nature of the new process overall—i.e., any participating site is limited to learning only one new topic per week from new users, supposedly an additional privacy precaution. For many smart veterans in our industry, it’s Welcome to 2002 with all the weaker personalization and ROI, plus 40x+ more ad revenue market share going to one company.
This brings us to the real conundrum—will publishers opt in? Users who are aware of these changes and have the option to opt out can do so via Chrome’s Privacy settings. Publishers must opt in by integrating the Topics API. “Why should the publishers bother?” is a fair question. Topics takes the anonymous group-targeting of FLoC to a whole new level with a lifespan of only three weeks per topic, drastically diminishing the granularity of interest-based targeting. This might not have such a significant impact for premium publishers, but smaller ones are sure to struggle with making their inventory as compelling as it used to be. Worse yet, without cookies, frequency capping has the potential to become a logistical nightmare, and publishers that truly value a positive site experience will be wary of alienating their users with ad fatigue. Perhaps most frustrating of all, Google’s FLEDGE and Core Attribution Reporting APIs remain in nascent test phases, placing severe limitations on an advertiser’s ability to retarget and measure conversions. In light of all these roadblocks, we may see Topics accelerating the adoption of Publisher-Provided Identifiers (PPIDs) and increasing focus on the expansion of PMPs. In comparison to the lower quality, less valuable targeting set of Topics, PPIDs offer publishers more control and more granularity and it will be very interesting to see what happens when advertisers start running tests of PPID-targeted vs. Topics-targeted campaigns. No doubt it will spark a whole new debate on measurement and tracking methodologies (which we look forward to with pleasure) and, if it accelerates the investments publishers are making in their own data products today, it might just have a positive outcome (for everyone but Google, that is). That said, a focus on first-party data and second-party data sharing in clean rooms doesn’t come cheap; it won’t be a fast and easy option for most publishers and perhaps not a viable one at all for many.
The final question is, “who wins?” and in this case, we’re afraid the answer is “nobody”, except, if things keep going this way, perhaps 4 companies globally. Users will enjoy a modicum of additional privacy, but that might prove a small comfort when they are bombarded over and over with much less relevant ads. Advertisers don’t win with less data for targeting, minimal or no frequency capping, and significantly constrained measurement. Publishers arguably fare the worst of all as these factors converge to exert downward pressure on ad pricing. All things considered, it’s hard to see this latest incarnation of FLoC (because that’s what it really is) as anything less than the adtech equivalent of rearranging deck chairs on The Titanic.
We believe there is a better way. A methodology where everyone wins, proven out in the past year with various publisher, brand, and tech client engagements we’ve been leading, and through more than 500 conversations we’ve had globally on the topic, just since 3/3/21. Our upcoming Post Cookie Targeting Analysis (PCTA) study, aka CounterFLoCWise, was conceived to illustrate how effective advertising can be when distributed appropriately across all inventory, making use of data beyond what AdsDataHub decides to let you use. Together with some of our most valued publisher, data, identity, and measurement partners, we’re aiming to demonstrate how campaigns perform across multiple methodologies to uncover best practices for targeting strategies in a post-cookie world.
Learn more about the PCTA Study and get on the list to receive the final report.