Reddit Is Targeting on Taste, Not Tracking. Where People Go Is the Most Honest Taste Signal.
Tell me you like lattes without telling me you like lattes…
Reddit wants to know you like lattes without knowing who you are. This week it partnered with Qloo, gaining access to more than 10 years of “taste” data that links restaurants, artists and brands by shared attributes.
The pitch: serve ads that fit what people care about, without personal data, search queries or browsing history.
Interesting direction. And a good idea. Affinity beats surveillance. But it begs the question: which taste signal is the most honest?
Solving the cold targeting start
Most new campaigns start blind, or at least hazy. Until now, Reddit may have needed days or weeks of engagement data before it could optimize a new advertiser’s ads. Qloo’s graph lets it find a proxy for the brand and optimize from day one. Someone reading about coffee machines might see an ad for a table to set one on, not another coffee machine.
This isn’t a new concept or execution. Put a nail gun in your cart at Amazon and they’ll suggest a box of nails that will fit the gun.
Predictive signals let brands skip the learning period. But the outcome is only as good as the signals guiding it.
Targeting on what people say vs. what people do
Online taste graphs are built from reviews and metadata: what people write, rate and tag. That’s stated taste, and it’s certainly useful. For decades, in-person focus groups hung on every word of anyone willing to sit in a room for two hours and spout opinions for a couple bucks and a few slices of pizza.
But stated taste can also be aspirational, selective and skewed toward the loudest voices. The person who reviews a natural wine bar once is not the person who goes every Thursday.
To reveal deeper affinity and taste, where people spend their time is the golden ticket.
Nobody visits a climbing gym three times a week to impress a review site. Real-world visits cost time, money and effort. And this makes them the clearest, most definitive signal of what someone actually values. When it comes to targeting, what people do and where they go is who they are.
Affinity audiences built on real-world behavior
That’s the brains and brawn behind Cuebiq Audiences. We build segments from anonymized, opt-in location data, matched to stores, restaurants, venues and other points of interest, then aggregate them so no individual is singled out.
Unlike some location data providers who compile signals from old, shady sources, Cuebiq data is farm-to-table fresh and fully opted-in. Every signal is privacy-first—fueled by direct relationships with app publishers,
The result is affinity without identity: independent coffee shop regulars, gym-goers, late-night diners, loyal shoppers of a specific brand. Here’s how our behavior-based audiences come together, from ready-to-use lifestyle and dining segments to custom combinations based on your specific asks.
Visit data can also show where stated and actual affinities split: the brands an audience truly spends it’s time, energy, and money on versus the ones it simply talks about.
Put the signals head to head
Taste-based targeting is a welcome step away from tracking. But does stated affinity hold the same targeting value as IRL behavior? In your next campaign, run a visitation-based affinity segment against a contextual one, then measure the real-world lift to your locations each delivers. Want help setting up the test? Book a demo and we can run a quick feasibility check.
