Google's Topics won't save marketers, here's why

Marketing is about influencing people. Not cookies, not audience segments, not visitors to publisher sites—but people.

Real people and households will make the individual decisions that are right for them. As marketers, we hope to influence those decisions in our favor, but in the end, it’s always their decision.

The way to influence people is to understand them: having rich data that helps with targeting and personalization—that’s what drives performance. There is no point in talking to an uninterested audience or saying the wrong thing to an interested audience.

Epsilon is fortunate to have access to some of the best assets for improving the performance of digital messaging, thanks to our clients who empower us to apply both their first-party data and our own extensive data assets to execute their marketing initiatives. We use these diverse assets to scale the peaks of performance and deliver real, transparently measurable results for our clients—so we are keenly aware of which assets make a difference to performance, and which don’t.

It’s true that having more data points is always better, but some data points are much more effective than others. We know that the best data, by far, is relevant intent data (recent search for a product or service) and relevant behavioral data (having previously taken a relevant action such as purchase of a product or service).

So, with all the talk about Google’s third-party cookie deprecation and the broader impacts on the industry, the key question becomes: Will Google’s new Topics offering be effective in improving performance?

No, it won’t—and here’s why.

Google’s Topics isn’t enough

Topics is a particularly limited form of contextual data, based on classifying the sites rather than the pages that the user visits, and will just give a little insight into where the user goes online.

We can expect essentially the same approach applied to app usage instead of web site visitation when Google extends Topics to Android. Topics doesn’t push performance to the top of the mountain—it leads (at most) to a hillock and often to no more than a bump in the road.

In fact, the entire discussion about browser-based replacements for third-party cookies leaves me cold because it’s talking about the wrong things. Performance requires prediction: Who is likely to buy? Who can be influenced to buy?

Contextual data, and especially site-level contextual data, is not very predictive because where a user browses tells you little about them and even less about what they buy. For example, visiting sports sites does not predict your likelihood of buying toothpaste or even a jogging suit, and reading about politics does not predict your likelihood of traveling to Washington DC. Your past purchases—in the form of first-party data—are far more relevant.

Contextual data is not very predictive because where a user browses tells you little about them and even less about what they buy.

So, how do third-party cookies help performance, if not through tracking users across the internet? They’re valuable to marketers because it’s possible to attach a rich, anonymized profile to them and then use that information to reach that individual online. Third-party cookies are valuable for reach, not tracking. The same is true for Android’s mobile Ad ID (MAID), which Google also intends to deprecate.

But if online browsing data has limited value for performance, what about the performance that Google touted for the now-retired FLoCs? The search giant said it would be “at least 95% of the conversions per dollar spent when compared to cookie-based advertising,” but Google never released its methodology, and independent testing showed much lower performance. Our best understanding now is that Google’s results were based on a scenario that is not quite the same as the one that marketers will face after third-party cookies are deprecated.

There is a better future for marketers

Even Google isn’t betting on Topics as a key component of their own ad technology, as they did not bet on FLoCs. Eric Schmitt, research director at Gartner, noted that third-party cookie deprecation will have little impact on Google’s most important capabilities in DV360. Today, Google relies primarily on its own first-party identity and other solutions and will continue to do so after deprecating third-party cookies in Chrome and Android’s MAID.

At Epsilon, our extensive experience has taught us that the foundation for delivering performance is rich profiles that are anchored in first-party data. Our recommendations to marketers are based on that deep understanding:

  • Implement or enhance a CRM solution that creates your foundational first-party data asset in a consented, privacy-safe way.
  • Build a deep understanding of your customers by analyzing and developing insights from your first-party data.
  • Partner with publishers to reach your customers and prospects wherever they spend their time (in a consented and privacy-respectful way) in exchange for helping the publishers receive fair pay for their content quality.
  • Deliver messaging to hold on to your customers—and to find new customers like them.
  • Work with technology partners that feed all the insights gained through messaging back to you, so you keep learning about your customers and prospects.
  • Use the insights to improve your products and personalize your messaging.

Most importantly, you want to keep a keen eye on your marketing performance and use these recommendations to make it a little stronger every day. It can feel like a daunting task, but don’t ever stop striving for that peak—you’ll reach it.  

This article was originally published on Adweek, March 2022.