Measuring media channels in a vacuum can easily skew marketing budgets, but marketing mix modelling (MMM) is increasingly being used to inform rational decision-making. Tobin Thomas, CEO and Founder of Lifesight, breaks down some uncomfortable yet valuable insights gleaned from MMM.
Here’s the uncomfortable thing about marketing measurement: the channels that are easiest to track usually get more credit than they’ve earned, and the ones doing the heavy lifting get shortchanged. This is not because anyone is cooking the books, but because the dashboard rewards whatever sits closest to the click.
Marketing mix modelling (MMM) exists to answer a question that sounds simple and almost never is: what’s actually driving growth? The answers deserve to be known, even if they tend to annoy people. Here are five examples:
1. Brand is more measurable than you’ve been told
Contrary to popular opinion, TV, video and audio measure fine; they just don’t fit click attribution. The fastest proof of this is when a brand pulls back on upper-funnel spend, branded search softens and the baseline quietly drifts down weeks later. That lag is exactly why last-click attribution misses brand impact – exactly what MMM is built to catch.
2. Your best-performing channel probably isn’t your most valuable one
This is the one that starts arguments in QBRs. A retargeting line posting 8x ROAS feels untouchable until a geo holdout shows its incrementality to be near zero. It was intercepting people already walking to the register; high ROAS, low contribution. Meanwhile, an upper-funnel channel creates demand that gets redistributed to search and direct, where it shows up as someone else’s win.
Efficiency and growth are not the same metric. It is when they get confused that budgets are optimised straight into a plateau.
3. Channels are a system, not a leaderboard
We evaluate channels in isolation, but customers experience them all at once. TV lifts search conversion, while brand makes retail media work harder, and social raises the ceiling downstream. Score each channel solo and you will underfund the ones whose value only shows up in combination, and you may never notice, because the credit landed elsewhere.
4. Spending more can make things worse
Every channel has a point where the next dollar does less than the last. Past that point, you’re buying diminishing returns and calling it scale. The eleventh dollar into a saturated Meta account does not perform like the first. MMM shows you where the curve bends, so you can move money out before efficiency erodes. The win doesn’t usually come from a bigger budget, but the same budget allocated like someone was paying attention.
5. Cross-media measurement isn’t aspirational anymore
The journey is fragmented across CTV, video, social, retail media, search and offline, so measuring any one in a vacuum tells you very little. You need one framework that holds all of them. No method is perfect, and anyone who says theirs is should be handled with suspicion. But modern MMM — richer data, faster refresh, better math — is good enough to make real decisions on.
The point was never a prettier dashboard. It’s the willingness to defund your favourite channel when the model says it isn’t pulling its weight. The teams that win with MMM aren’t the ones with the best charts; they’re the ones willing to act on an answer they didn’t want.
Follow VideoWeek on LinkedIn.
