MMMs Answer Whether CTV Worked, Not Why

Marketing mix models can tell a brand whether its connected TV spend paid off. They cannot reliably explain what inside that spend actually drove the result — and that gap is becoming the central measurement problem as budgets shift toward CTV.

That is the argument laid out in a new AdExchanger analysis, which frames the current wave of MMM adoption as a net positive for an industry still over-crediting the click. The piece's core observation: a click can be accidental, reflect low intent, or simply be easy to count — none of which describes how viewers actually respond to ads on a television screen, where clicking was never the point.

The official argument, as the analysis presents it, runs in two parts.

First, MMMs deserve the attention they are getting. Used well, they pull brands out of a measurement culture that still gives too much credit to click-based channels. For practitioners, that means a well-specified model can reallocate credit toward channels like CTV, where exposure — not interaction — is the mechanism. That is a healthy shift for budget decisions that have historically underweighted upper-funnel video.

Second, and this is where the piece draws its line, MMMs are aggregate instruments. They detect correlation between spend levels and outcomes at the channel level. They do not isolate which creative, which publisher, which daypart, or which audience segment inside the CTV buy produced the lift. A brand can learn its CTV investment returned, say, more than it cost. It cannot learn from the model alone which half of that investment deserves the incremental dollar.

The practical implication for media teams is straightforward. MMM output tells you how much to spend on CTV in total. It does not tell you how to spend it — the questions of creative rotation, inventory mix, and targeting that determine whether that total spend performs at the top or bottom of its efficiency range. Teams that treat an MMM result as a complete CTV answer, rather than a starting constraint, risk optimizing a budget number while leaving the execution unexamined.

The analysis stops short of prescribing a full solution stack, but its framing points toward the standard hybrid logic: MMMs for the whether, and finer-grained measurement — incrementality tests, audience-level experiments, platform-level reporting — for the why. Brands running CTV at scale will increasingly need both layers, because neither one covers the other's job.

There is also a cautionary note embedded in the piece's tone. The same properties that make MMMs attractive — channel-level aggregation, resistance to click bias — are exactly the properties that hide within-channel dynamics. The models' growing credibility on budget questions should not be mistaken for diagnostic authority they do not have.

As CTV spend continues to grow and brands lean harder on MMMs to justify it, expect the measurement conversation to move past whether the channel works and toward the harder, still-unresolved question of what inside it is working.