Commerce Media KPIs: The Metrics That Actually Matter Beyond ROAS

ROAS is lying to you. Not maliciously — it's just that the metric was designed for a world where media and commerce were separate, and in commerce media, they're not.

Commerce media collapses the distance between content, intent signal, and purchase. That means the performance framework has to collapse too. Optimizing solely for reported ROAS in a commerce media campaign is like judging a restaurant by how many menus they hand out. It tells you something, but not whether anyone ate. Here's how senior practitioners actually measure commerce media performance — and the three numbers most brands are ignoring.

Incremental Revenue per Thousand Impressions (iRPM)

Most media buyers chase CPM or CPC as efficiency proxies. Those metrics tell you how cheaply you acquired an impression or click — they say nothing about whether that impression did anything that wouldn't have happened anyway.

iRPM — incremental revenue per thousand impressions — cuts through that. It asks: of the revenue attributed to this campaign, how much would have occurred even without the ad? Run a holdout test, compare conversion rates between exposed and unexposed audiences, and you get a number that reflects real commercial impact.

This matters especially in commerce media because the audiences are already high-intent. Someone searching for "best running shoes" was probably going to buy something regardless. The question is whether your ad changed what they bought, from whom, and at what margin.

• Run geo-based holdouts or PSA substitution tests quarterly to establish a baseline.
• Compare iRPM across placements to find where the ad is actually doing work versus riding intent that was already there.
• Use iRPM alongside ROAS, not instead of it — the ratio between them tells you how much of your reported return is real.

New Customer Acquisition Rate

Reported ROAS frequently flatters by over-indexing on existing customers. Retargeting campaigns, loyalty-adjacent placements, and branded search terms all drive high ROAS — and almost all of it is demand you already owned.

Commerce media should be acquiring net-new customers at a rate that justifies the investment in high-intent traffic. Track new-to-brand (NTB) orders as a percentage of total campaign conversions, and set a floor: if fewer than 40–50% of conversions are NTB, you're largely paying to remind people who were already coming back.

This metric also lets you model long-term value more honestly. A 2.5x ROAS from existing customers looks very different from a 2.5x ROAS from first-time buyers who carry a 180-day LTV three times higher.

Commerce media KPIs aren't just about what happened last click — they're about what changes in the customer relationship. If your metrics can't tell the difference between retention and acquisition, you're flying blind.

Assisted Revenue Share and Path Position

Commerce media often operates mid-funnel: it surfaces at the moment of intent but doesn't always capture the last click. Last-click attribution will systematically undervalue these placements, creating a perverse incentive to pull spend from exactly the campaigns that are priming your highest-value conversions.

Assisted revenue share measures how often a touchpoint appears somewhere in the conversion path — not just at the end. Pair this with path position analysis (first touch, middle, last) to understand which commerce media placements are seeding intent versus closing it. Impact, AWIN, and CJ all offer path-level data in their reporting suites if you configure the attribution window correctly.

• Don't defund a placement because it has low last-click ROAS if it carries strong assisted revenue share.
• Mid-funnel commerce media with high path presence can have 3–5x the influence suggested by last-click models.
• Set attribution window minimums at 7–14 days to capture delayed purchase behavior, especially for higher-consideration categories.

Category Entry Point Coverage

This is the commerce media KPI almost nobody tracks — and the one that best predicts long-term program success.

Category entry points (CEPs) are the moments when buyers enter consideration for a category: a specific search query, a comparison article, a product review. Commerce media that owns those moments shapes which brands get considered before any ad is served in the final purchase session. If you're only present at the bottom of the funnel, you're competing for buyers who've already shortlisted your competitors.

Build a quarterly CEP audit into your measurement practice: map your brand's presence across the top 20–30 intent signals in your category, identify gaps in coverage at high-volume entry points, and use that map to guide placement decisions across Google, Meta, Taboola, and your affiliate publisher mix.

The brands winning in commerce media aren't the ones with the highest reported ROAS. They're the ones tracking metrics that reflect actual commercial leverage: incremental revenue, new customer rates, assisted contribution, and category presence. ROAS matters — but only as one signal inside a measurement system built to distinguish real growth from cannibalized demand. That's the distinction that separates a commerce media program that compounds from one that flatlines while posting impressive-looking numbers.