3 Debates Worth Having about Google’s Topics
Last week, Google announced Topics, a new data-driven ad targeting model to replace third-party cookies.
Topics succeeded Federated Learning of Cohorts, or FLoCs, which proposed grouping and targeting Chrome users based on interests, as observed by Google over long periods of time.
By contrast, Topics suggests collecting data on users’ interests on a rolling three-week basis and allowing advertisers to target users with ads relevant to three of those high-scoring interests. Google will determine Topics relevant to internet surfers based on the domains and subdomains they visit, and users will be able to opt out of Topics tracking.
The announcement set off a firestorm of debate over whether Topics goes far enough to secure user privacy, how effective Topics would be as an ad targeting methodology, and what the announcement portends for digital advertising as a whole. Here is a rundown of those three debates over Topics and the relevant perspectives adtech stakeholders should consider.
Does Google’s Topics go far enough on privacy?
The emerging data privacy era has effectively killed third-party cookies, which allow individualized tracking by obscure third parties, often without users’ knowledge or consent. FLoCs faced criticism for not deviating far enough from the norms of cookies. Sure, FLoCs proposed grouping users instead of enabling individualized tracking, but if Google is still tracking users across sites without opt-in consent, what’s the difference, some wondered.
Talk to some in the adtech industry, and they will tell you Topics is basically just behavioral advertising 101. The idea still involves tracking user browsing, probably without the knowledge or thoughtful consent of many (like so much internet tracking). The fact that Google will hypothetically limit behavioral tracking to three weeks doesn’t mean Google will have repented for adtech’s original sin (tracking users without their knowledge or active consent).
But others argue that Topics is very much a step in the right direction and that the methodology, as currently imagined, represents a meaningful improvement over third-party cookies and FLoCs. Limiting the duration of tracking means Google is restricting the treasure trove of granular data it collects on users, something most privacy-conscious consumers would welcome. Plus, providing users the ability to opt out of tracking is positive. Finally, many would say Topics resembles contextual targeting, which promises to match ad content to the content of users’ digital journeys — with the caveat that advertisers can harness three weeks of context to make a better match. I would agree that Topics shows Google understands the future of digital advertising will hue closer to the promise of context.
Ultimately, then, just how much privacy progress Topics delivers will depend on the implementation of the idea (should it ever come to fruition). This is true of much data privacy reform, which often sounds more meaningful in theory than it turns out to be in practice. For example, letting users opt out of third-party cookie tracking sounds good, but if in practice opting out is impossible or so cumbersome that it compromises the user experience, requiring sites to let consumers opt out is of little help. So it will go with Topics.
Of course, if Google really wanted to go all in on privacy in the manner of Apple, the company could require users to opt into Topics tracking. That would likely usher in a sea change in web-based ad targeting more similar to the shift Apple triggered when it forced mobile users to opt into cross-app targeting on iOS devices.
Will Topics be an effective ad targeting methodology?
Among the criticisms of FLoCs was that it would not be sufficiently effective to replace third-party cookies (despite Google’s reassurances that FLoCs would allow almost as much ad efficacy as cookies). Topics should and will face the same questions. So much of the open web depends on advertising to survive. Will Topics deliver?
I think it’s clear that Topics-based targeting will be less granular than that enabled by cookies. Indeed, Topics even falls short of the precision of much contextual targeting, as it will not analyze the content of the sites Chrome users visit — just the domains and subdomains.
The question, then, is not whether Topics targeting will be less granular than cookie-based targeting but whether hyper-personalized cookie-based targeting was ever as effective as advocates like to claim. Individualized targeting has faced criticism for serving users ads for products they already bought, creeping out customers, and for delivering inflated metrics because it targets people with ads for products they would have bought anyway.
Whenever Google gets around to implementing Topics, or whatever targeting model replaces third-party cookies, at scale, we will begin to get answers about the deeper question here: just how integral granular behavioral tracking is, and ever was, to digital ad effectiveness.
What does Topics tell us about the future of digital advertising?
In the long history of data privacy’s ramifications for digital advertising, we may look back on Topics as the last gasp of third-party behavioral ad targeting. I don’t expect the federal US government or even progressive state legislatures to “ban surveillance advertising” anytime soon, but I would not be surprised if foreign regulators, public sentiment, and the need to maintain pace with privacy bulls, most notably Apple, ultimately compel Google to tell advertisers, “Figure it out with contextual targeting, and add first- and second-party behavioral data to enrich contextual targeting where you can.”
Whether Google implements a targeting methodology like Topics or shifts more heavily toward context over the long term, it is clear that granular targeting will get harder, likely increasing the cost of user acquisition for advertisers and putting a premium on publisher inventory that comes paired with rich first-party audience data.
Publishers know they need to double down on first- and second-party data while strengthening the case for data collection to users to gain consent. Advertisers know that exploring contextual and other post-third-party-cookie targeting tools is key and that they, too, need ways to amass and consensually share first-party data.
In the short term, then, I see Topics benefiting those already capitalizing on the data privacy movement: adtech players devising ways to make the most of limited consumer data or target ads via other means and any data company helping marketers collect truly privacy-safe, consensual consumer data. How good Topics is for advertisers and publishers will hinge on a more complicated question — just how effective, and perhaps irreplaceable, third-party-cookie based targeting ever was.