The Great Unbundling of Brand Truth: Why Marketing's Measurement Complex Is Due for Disruption
The companies abandoning perfect attribution for directional accuracy are seeing 3x better returns
The Attribution Theater We All Participate In
A CMO friend recently showed me their attribution model. Two years in the making. Millions invested. It confidently declared their Super Bowl ad was worth... nothing. Zero. Meanwhile, Google Shopping was apparently responsible for 97% of sales.
The board loved it. Clear numbers! Data-driven decisions!
Nobody mentioned that brand searches went up 400% after the Super Bowl. Or that their stores were packed the following week. The model couldn't see it, so it didn't exist.
This is where we are. We've built elaborate systems to measure the measurable and ignore everything else. It's like looking for your keys under the streetlight because that's where you can see, even though you dropped them in the dark.
How Privacy Changes Became an Opportunity
Meta basically gave up on attribution. Their new Advantage+ campaigns come with a simple message: "Stop trying to figure it out. Just trust us."
You know what? It's working. Adoption is up 300% because advertisers are exhausted. They're tired of pretending their attribution models mean something.
The brands succeeding on Meta aren't the ones with fancy measurement. They're the ones who accept uncertainty. They know their ads work-ish. They know the direction, not the decimals. And that's enough.
Take Outdoor Voices. They scrapped their attribution model and started doing something stupidly simple: turning off all marketing in Denver for two weeks to see what happens. Stone age measurement? Maybe. But it actually tells them what their fancy model never could—what happens when the ads stop.
The Rise of Probabilistic Decision Making
Walmart gets it. Their DSP doesn't promise perfect attribution. It offers incrementality testing. Same with Amazon's Marketing Cloud. They're not tracking people anymore. They're running experiments.
The privacy apocalypse everyone feared? It freed us from measurement theater. When you can't track individuals, you have to think bigger. You have to ask better questions.
Look at retail media. Target's network doesn't pretend to know which ad view led to which purchase. They just know you saw an ad and bought the thing. Simple. Clean. Honest.
Ricola (the cough drop company) went all in on this approach. One metric: incremental revenue per marketing dollar. They measure it through holdout tests—some people see ads, some don't, compare the difference. Their CMO told me they spent years perfecting attribution models before realizing they were "optimizing for the model, not for growth."
What Happens When We Stop Pretending
New tools are making this easier. Companies like Recast and Measured use the same statistical methods drug companies use to test medicines. No tracking pixels. No cookies. Just math that answers one question: did this work?
P&G's Marc Pritchard said something brilliant recently: "We've been trying to impose precise measurement on an imprecise medium." They cut their KPIs from 300 to 12. Revenue went up.
Netflix never measures the ROI of individual shows. They measure subscriber retention. They get that content works as a portfolio. Some shows attract, some retain, some do both. The magic is in the mix, not the parts.
Here's what's coming: by end of 2025, a major brand will publicly abandon attribution modeling. The ad tech world will lose their minds. Everyone else will wonder why it took so long.