If you're weighing commerce media vs. programmatic display, here's the short version: one is built to capture purchase intent, the other is built to buy impressions at scale. They are not interchangeable line items on a media plan.
Plenty of e-commerce brands still treat them that way. The programmatic display budget gets renewed every year because the CPMs are cheap, the reach numbers look impressive, and the retargeting ROAS looks great on a dashboard. Meanwhile, commerce media — placements built around shopping context and paid on performance — gets evaluated with ten times the scrutiny. That asymmetry costs brands real money, because the two channels answer completely different questions.
Programmatic display is an efficiency machine for one thing: impressions. Through DSPs like DV360, you bid on ad slots across millions of pages, layered with audience data of varying quality. It's excellent at broad reach and frequency, and at low CPMs it can look like a bargain.
The problems show up when you ask what those impressions do:
Commerce media flips the model. Instead of chasing audiences across random inventory, it places your brand inside shopping contexts — comparison content, buying guides, deal placements, curated product discovery — where the reader is already in purchase mode. And critically, much of it is transacted on performance: you pay when a sale happens, not when a pixel loads.
That structural difference changes the economics:
Programmatic display asks "who can we show this to cheaply?" Commerce media asks "who is about to buy, and where are they deciding?" Those are not the same business.
Display's measurement problem is well documented: view-through attribution gives credit for impressions nobody noticed, and retargeting takes credit for conversions that were already coming. Run an incrementality test on a mature retargeting program and the lift is often a fraction of what the attribution dashboard claims.
Commerce media isn't automatically incremental either — a placement in front of a shopper who already chose you is still captured demand. But because the channel is transaction-based, testing is more honest by default. You can compare cohorts, measure new-customer share, and structure payouts so you're rewarding acquisition rather than interception. The channel's economics make it easier to demand proof, and harder for waste to hide.
This isn't an argument to zero out display. It's an argument to stop grading the two channels on the same curve.
The brands winning right now aren't the ones with the most impressions. They're the ones who stopped paying for reach they couldn't tie to revenue and moved budget to channels where intent, placement, and payment are aligned. That's the entire case for commerce media: fewer vanity metrics, more transactions you can actually count — and prove you wouldn't have gotten anyway.