Here’s Why FLoCs Are a Step Backwards in Data Personalization (And What to Do Instead)

Tim Glomb, VP of content and data at Cheetah Digital

Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Tim Glomb, VP of content and data at Cheetah Digital.

Did you know that, according to a study by Accenture, 91% of consumers want a personalized online shopping experience?

But at the same time that buyers want personalization, they also want privacy. Consumers are growing more concerned about companies tracking their online activity across channels: The same study found that 35% of consumers said it’s “creepy” when a company send ads to social sites after a user browses its website.

To put it simply: There’s a fine line between anticipating customers’ needs and overdoing it with an invasion of online privacy.

With these concerns in mind, Google recently opted to get rid of third-party cookies in favor of a new system called Federated Learning of Cohorts, or FLoC. Instead of the cookies that often work in the background and track users as they travel from site to site across the internet, FLoC groups people together by their shared interests. This grants users a sense of anonymity online. Google, meanwhile, claims that it will still be able to offer personalization.

But that’s not possible. In fact, a FLoC-based strategy is a step backwards in data personalization.

Consumers Want a Targeted Customer Service Experience

What it comes down to is that FLoCs eschew layers of personalization in favor of broad generalizations. While users will have more privacy than they get with cookies, they’re more likely to receive inaccurate recommendations, because information about them is merely inferred.

Consumers want specificity from the brands they interact with, to improve the shopping experience. A fitting analogy is the in-store shopping experience, at least if the store has good customer service. Clients looking to buy specific items will be met by sales representatives, asked what they’re looking for, and offered solutions directly. That sort of precision and personalization is what consumers want, on a digital platform no less than in a store.

Don’t Overlook the Simple Solution

Privacy concerns have arisen in recent years largely due to those aforementioned “creepy” ad strategies. Ultimately, though, we’ve seen that customers are willing to share a little information about themselves in the interests of a better shopping experience. Think of customer surveys that offer discounts or chances to win, or consider style quizzes that help fashion-minded shoppers pinpoint exactly what they are looking for.

FLoCs are a quick fix, one that doesn’t acknowledge the core issue: Consumers want to be heard, and they want their voices to be respected. Putting them into larger groups doesn’t fulfill either of those desires.

There’s a simpler solution, though: zero-party data. A zero-party data strategy directly asks the buyer for information. This makes it clear from the start who’s gathering information and what for. Consumers understand exactly how their data is being used, allowing them to express their preferences and maintain privacy.

Zero-party data is the better solution for brands, too, because they get the chance to own their data. No longer do they have to “rent” it from third-party sites in a way that gives them no control over its use. You can do more with data you own because you can control where it’s stored and what’s done with it. Rather than lumping consumers together, you can now segment them for ad targeting on the basis of your own data points as well as deliver hyper-personalized email and SMS offers at scale.

If we want to move the needle forward in data personalization, and we do, we need to respect the customer’s wishes and change our data strategy accordingly. Zero-party data is the simplest way to acknowledge clients’ desires for privacy while ensuring that they still benefit from a customized experience.


Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Three thoughts on FLoCs:

    1. Consumers want personalization AND privacy — This is very similar to the trick question of “what happens when an irresistible force meets an unmovable object?”. Well. Nothing. They can’t coexist in the same universe. The moment you have data, you have hackers, you have people who will misuse the data, you have people who will abuse the data.

    2. “Zero-party Data”. This is totally semantics heavily coated by BS. Does anyone doubt that the system will be gamed, corrupted and abused in a New York second? I don’t. I would not believe for a moment (as a consumer) that some well-meaning, totally-transparent entity is going to be benign in their collection of data.

    3. Consumers are “willing to exchange their data for personalization”, a belief that has been battered into some unrecognizable shape by all the unrelated questions one faces. Do you really need my zip code to use the “build your own” feature in a car site? Having my zip code makes sense only 2 ways: (1) I am inquiring about offers in my area (which I’m not as a I build my dream Giulia) or (2) you are collecting geographical data to see what people in 33146 like vs. people in 33010.

    To me, this sounds like just more lipstick on the pig.

  2. You are describing 1st party data, not zero party data. You’re not the 1st party to get this wrong, as a simple search will dozens of mistaken references to this term as akin to ‘consent-based’ data collection. The reality is unique user data would only be ‘zero party’ if the ‘1st party’ never gets to ‘control’ the data. These other companies may get to use the data for various purposes, but the use has to be exclusively within the control of the individual who ‘owns’ it (or it belongs to depending on your perspective), such as through a secure wallet or possibly through a clean room/homomorphic encryption. A preference center could be a wallet which is controlled by the user with a realtime api to the data lake/DMP, perhaps such as what Killi (and many others have tried/will try – now with blockchain!)