Marketing Mix Models Will Tell You If Your CTV Spend Flopped—But Don’t Expect a Postmortem

Let’s get this out of the way: anyone still reporting on click-through rates for CTV in autumn 2026 deserves the same respect we give to people selling fax machines on Flatbush. Clicks were never the point for TV, connected or otherwise, and the fact that agencies still try to shoehorn CTR into their CTV decks is pure cargo cult. So, yes, it’s a welcome shift that brands are finally waking up to marketing mix models (MMMs) to figure out if all those streaming ad dollars did more than just warm the pockets at Roku.

But here’s the problem: MMMs are blunt instruments. They’ll tell you—sort of—if your CTV campaign correlated with a sales bump last quarter. They won’t tell you if it was the 15-second pre-roll on Hulu or the weird mid-roll on Pluto TV that made the phone ring. And they definitely won’t explain why your CPMs went up 40% this season while your actual lift flatlined. It’s like getting a yes/no answer from a Magic 8 Ball when you need a transcript from the jury.

Let’s stop pretending this is some technical limitation that will vanish with a fresh batch of Python scripts. The opacity is built in. MMMs aggregate and smooth everything until you’re looking at a blurry impressionist painting of your budget. That’s fine if all you want is to tell your VP of Marketing, “CTV worked!” But if you want to actually optimize—kill bad creative, double down on the segment that converts, stop feeding the DSP grift—MMMs offer zero actionable insight. And don’t get me started on the consultants flogging black-box MMM solutions at $30k a pop. Their dashboards are prettier than what you get from the Sunday Times crossword, but just as inscrutable.

If you’re serious about accountability, you need to go full-on measurement psychopath: run incrementality tests, force your media partners to split their inventory, demand log-level data, and actually look at the overlap between your CTV and other channels. Spoiler: it’s going to be ugly, but at least you’ll know which half of your spend is lighting the autumn leaves on fire.

Here’s my uncomfortable fall recommendation: stop letting MMMs be your fig leaf. Use them for what they’re good at—blunt trend detection—and then do the hard work of experimentation, channel audits, and creative reviews. Or keep worshipping the MMM gods and enjoy another year of PowerPoint astrology. Your choice.