Google’s abandonment of FLoC (Federated Learning of Cohorts) as a replacement for tracking cookies and the introduction of the Topics API is consistent with the company’s belief that the browser holds the keys to how interest-based targeting will work in the future.
The impact of today’s Google announcement on news publishers appears to be:
- Certainty: If the Topics API sticks, Google’s aim is to bring certainty back to the advertising ecosystem.
- Calming regulators: This move could cool the waters of discontent among regulators, yet it won't likely appease most regulators or resolve fundamental consumer privacy concerns.
- Publisher advantage: Publishers that have invested in first-party data strategies may have an enhanced competitive advantage, depending on how well developed their audience cohorts are.
How Topics works
Here is what we know so far about the Topics API:
- This is a new system for interest-based advertising.
- It works by identifying five of a person’s interests in the past week, such as “Sports” or “Travel” as a person moves around the Web.
- The Chrome browser stores these interests for three weeks, then deletes them.
- The categories are selected on a person’s device. Google emphasises that no external servers are involved.
- When a person visits a Web site, the Topics API will show the site and advertising partners three interests — one topic from each of the past three weeks.
- There are about 350 topics in Google’s advertising categories.
- The timetable for phasing out third-party cookies for Topics remains the same as FLoC: late 2023.
Google noted that it had spoken with many advertising industry stakeholders — including INMA members in two town halls last year.
The critique of the previously proposed system, FLoC, came from privacy advocates who said data could be used to build profiles of people. FLoC was seen as a step back from today’s third-party cookies that help build more detailed profiles of people and behaviours — trigger regulation and ad blockers worldwide.
What this move means for publishers
I visited today with the new lead of INMA’s Smart Data Initiative, Ariane Bernard, and here were her takeaways from the Tuesday announcement:
- It’s still about the browser and interest-based targeting: Google isn’t backtracking on its general approach to replacing cookies. The company is keeping within similar lines of what it previously conceptualised. This is about the Chrome browser and interest-based targeting.
- Topics will make targeting fuzzier: Topics only proposes a random slice of a browser’s observed top topics of interest. Said Bernard: “FLoC schematically would have made targeting fuzzy and users harder to identify by describing a person as part of a group. The approach would have been to be as descriptive of a user’s profile as possible but fuzzy on who they were. Topics would do this by subtracting part of the picture that makes the user’s interests, rather than only focusing on identity.”
Publishers may be uncomfortable if Google is too successful at making targeting fuzzy because this likely will be done at the expense of CPMs. “This is because what Google is trying to do here is to approach same-or-better ad effectiveness with methods that conceptually are like putting a light piece of cloth on the eyes of the sharpshooter,” Bernard said.
Google and publishers: diverging priorities
While publishers and Google both want the highest CPMs they can get, they have different risk scenarios.
Google’s risk isn’t a few percentage points of CPM effectiveness, “while these few percentage points are possibly a dire difference for publishers who don’t sit on coffers of cash,” Bernard said. Instead, Google is under pressure to find a framework that delivers an adequate level of user privacy to thwart privacy or anti-trust regulation.
In other words: Publishers need ad CPM optimisation. Google needs to fend off regulators.
Whether FLoC or Topics, Bernard believes Google and publishers will land in a similar place: “Where the mechanism to deliver a targeted ad is dulled up, campaigns are naturally less effective. Wherever advertising is strictly display-based, this may not be as problematic. But as the bulk of advertising dollars are performance-based, less effective campaigns mean smaller revenues.”
Publishers and a first-party data strategy
Whether FLoC or Topics, publishers that invest in first-party data that advertisers can leverage have emerging advantages such as using their data in addition to the Google Topics API.
Said Bernard: “Google may be obfuscating part of the picture of a user’s specific profile or interests, but for a publisher with a good store of user data and a registered user on the site, their first-party data remains as useful as it ever was. Possibly more, in fact, relative to the rest of the data picture being used for ad serving. In this respect, where Google may offer less, publishers have their own card to play with their owned knowledge of their known users.”
The direction for publishers remains clear: Get as many registrations and reader subscriptions, together with as many consumer permissions as possible, to sell targeted ads against each of them. Build your walled-garden, then join like-minded publishers to create scale.
Meanwhile, the flagged changes, while lacking detail, don’t appear to threaten a publisher’s revenue streams from re-selling Google ad products.
Regulation and privacy
On the regulatory and privacy fronts, Topics is an improvement over FLoC, which was an improvement over today’s third-party cookie ecosystem.
Yet asked to assess the likely international regulatory reaction to the latest evolution, INMA Digital Platform Initiative Lead Robert Whitehead told me that today’s announcement wo’t appease most regulators. “This move from cohorts to topics will not be enough to get most regulators off the Google case. It still strengthens Google’s hold on data, even though there are several thousand fewer topics envisaged than cohorts.”
Meanwhile, Whitehead said the move won’t resolve fundamental privacy concerns. “As Apple has emphatically demonstrated, when consumers are given a choice to switch off tracking, they switch off in large numbers.” Privacy as a core proposition is a movement that isn’t going away.
Quick industry reactions
The reaction to the Google announcement was muted today as the ad tech community and privacy advocates delved into the details and ramifications. From the privacy perspective, the Topics move may be a middle ground that will never satisfy some – though there is consensus that it’s better than FLoC. And ad tech executives suggest that moving from highly targeted ads to Topics might be sub-optimal (as Bernard noted, too).
The publishers with whom INMA talked today were processing the information before commenting on what, if any, impact the shift from FLoC to Topics means. Yet they were unanimous that publishers adopt their own first-party data strategy and accelerate those plans before late 2023 remains imperative.