FloC is dead. Long live Topics?
Some seven months after delaying the removal of third-party cookie tracking from its incredibly popular Chrome browser until 2023, Google introduced a new toy into its Privacy Sandbox: Topics API.
The search and ad behemoth’s new identity solution would pull in a user’s web history from Chrome, then utilize that information to assign that user up to five Topics (such as “Auto & Vehicles,” “Books & Literature,” or “Rock Music”) that advertisers could use for targeting relevant audiences. New topics would be determined every week, and after three weeks, old topics would be permanently deleted. Also of note: Google is officially putting an end to its FLoC experiment, thus ending the life of its previous (widely unpopular) proposed identity solution.
What does this mean for the advertising industry? Certainly, it’s far too premature to know whether Topics is really the future of targeting on the world’s most popular web browser (from the world’s most popular search engine and the world’s most profitable online advertising business), but here is a quick look at the new proposal as well as some immediate takeaways:
Let’s start with the upsides. In a lot of ways, Topics looks like Google’s attempt to placate as many FLoC-related fears as possible while still maintaining a means for cross-site tracking based on browser history. As a result, Topics seems like a solution that may ultimately appeal to more stakeholders—but, as you might expect, not all stakeholders.
For advertisers (and regulators), Topics does not look like the sort of “power grab” that made FLoC the boogeyman to many in the industry and around Washington, while still providing key insights on consumer behavior that can help compliment a larger audience-driven marketing strategy. Publishers, meanwhile, have a blueprint for how their sites fit into to the new topic categories.
As for consumers, the new API’s simplified groupings and the absence of more sensitive identifiers such as “gender” or “race” may appeal to privacy-minded individuals who wish to remain more anonymous while using Chrome and retain more control of what advertisers can see and learn about them online. As Basis SVP of Media Operations Zach Moore puts it, “Topics seems like a much easier way to explain to privacy-conscious consumers that they themselves have the power in their own hands when it comes to targeted ads, vs. having to explain FLoC’s cohorts and algorithms.” Additionally, Google plans to empower consumers to easily review and remove topics from their profile, or to opt-out of Topics API entirely by simply turning the feature off, which may help some skeptics feel a greater sense of power over their own privacy.
Of course, it can’t all be good news, and there are still many aspects of Topics that Google will need to address—or at least iron out a bit—before any widescale implementation or adoption. For publishers and advertisers, Google’s initial draft around how they plan to categorize interests for Topics may be a cause for some concern. It would appear, at least for now, that Google intends to gather its behavioral data based on a website’s hostname alone—rather than the full context of a page’s URL—before bucketing those interests into one of 350 topics. That number makes Google’s list fairly distinct from, say, the IAB’s much more expansive content taxonomy.
However, this strategy may inherently overlook some meaningful contextual information and granularity that could potentially benefit advertisers, publishers and willing consumers alike. For example, someone visiting a website like Amazon or Target could place that user into the Topic of “Shopping,” but that information is much less helpful than knowing the full context of a specific page’s URL (ex. https://www.amazon.com/diapers/s?k=diapers), with which advertisers can infer much more about someone’s behavior and actual needs.Additionally, there will likely be many within the industry who question Google’s initial plan to cap topics at five per week. As Basis VP of Media Systems Ken Rood puts it: “I contain multitudes. Grouping me into five broad topics is good for privacy, but it may not be good for showing me the most useful advertising.”
As for privacy advocates, while Topics is clearly a step in the right direction as compared to FLoC, any browser-based cross-site tracking solutions are likely to face extraordinary skepticism, particularly when they come from Google. It also does little to allay the concerns of those who fear Google having an outsized influence on the online advertising ecosystem. Whether that proves to be a vocal minority or a larger segment of the population remains to be seen, but it’s definitely something to keep an eye on in the months ahead.
In many ways, Topics looks an awful lot like another longtime advertising solution that has reemerged in recent years: contextual advertising. As we have previously detailed, contextual advertising shirks user data entirely and instead involves placing ads that are related to a webpage’s content. Topics, then, seems almost like a contextual advertising martini with a browser data twist. Whatever happens with Topics, the fact that it is so similar to contextual advertising only reenforces the important role that context will play in the future of digital advertising.
All in all, Topics may well be just another sign that there is no true “heir” to the third-party cookie throne. Instead, advertisers could be relying on a series of solutions (rather than one single solution) to stitch together their targeting strategies. One thing, however, still seems clear: in the land of the cookieless, first-party data will be king.
There will, of course, be much more to come on the topic of Topics, but the news is certainly a major development for the advertising industry and will undoubtedly be the subject of serious scrutiny in the months ahead.
In the meantime, media buying professionals are still looking for answers on the future of targeted advertising. For our perspective, check out Beyond Third-Party Cookies: Your Guide to Overcoming the Identity Crisis.