Direct-response advertising has a comforting feature: someone clicks, someone buys, and a number lands in a dashboard. Brand advertising doesn't work that way. A pre-roll spot is seen by a million people; six weeks later, sales are up. Whether those two facts are connected — and how strongly — is the question that decides whether next quarter's brand budget gets approved.
For a direct-to-consumer brand we worked with, this was a real budget conversation. The team's growth had been built on direct-response performance media, where attribution was clear. They were now investing in upper-funnel brand media — TV, audio, premium video — and finance was reasonably asking what it was buying. The standard performance dashboards weren't going to answer that.
Three Layers, Not One
The instinct in this kind of situation is to reach for marketing mix modeling and call it done. MMM is part of the answer, but it's not the whole answer. Brand media affects the business through three different mechanisms, and a measurement framework that captures only one of them will systematically underclaim what brand is doing.
If your only brand-measurement tool is a sales-attribution model, you'll always conclude that brand drives less than it does — because most of what brand does happens upstream of the sale.
Layer one: brand metrics. Aided awareness, consideration, preference — the soft metrics that move first when brand media works. The brand had an existing third-party tracker, but trackers are quarterly at best. We supplemented it with a continuous online survey, social-listening data, and search-trend signals, then statistically related week-to-week movement to media activity. The output was a near-real-time read on whether brand media was actually shifting how the audience thought about the brand.
Layer two: direct sales lift. A marketing mix model attributing weekly sales variation to each media channel, including brand. This is the standard MMM work, but we treated it as one input rather than the whole answer. The contribution that surfaces here is the slice that even a skeptic would credit to brand media.
Layer three: indirect impact on other channels. The most consistently underappreciated effect. When TV is on the air, search volume rises. When a podcast campaign runs, retargeting performance improves. The brand layer creates demand that the response layer captures and converts. Treating those as independent channels — which is what most measurement setups do — credits the response channel for sales the brand campaign actually generated.
What This Changes
The combined framework produces a more honest picture of brand contribution than any single layer would on its own. Brand media looks weaker than it is in a pure MMM (because the model can't see the indirect lift) and stronger than it is in a pure brand tracker (because awareness gains aren't dollars). Adding the layers together gets you closer to the truth.
For the team, the immediate output was a defensible answer to the budget question. The downstream output was a planning tool: a scenario simulator that took the model's coefficients and let the team test reallocation ideas without re-running the full analysis.
Any business with both upper-funnel and lower-funnel media has the same measurement problem. Subscription services, retailers, B2B SaaS with brand campaigns, hospitality, financial services. Wherever brand and response live in the same plan, measuring only the response side underclaims brand by exactly the amount of the indirect lift — which can be substantial. The hard part isn't the modeling. It's the willingness to maintain three measurement layers when one would be simpler. Simpler isn't always more accurate.
Trying to defend or measure brand media in a performance-dominant org? Say hello.