Platform-Reported ROAS vs Actual ROAS: Why Rising CPMs Make the Gap Costlier Right Now
Meta CPMs are up roughly 20% year over year, and Google Search CPCs have climbed about 12%.
That alone would be enough to squeeze margins on any ecommerce account. But there’s a second, quieter shift happening at the same time: platform-reported ROAS is getting more generous, not less.
The gap between platform reported ROAS vs actual ROAS was already a problem. Rising media costs just made ignoring it a lot more expensive.
Meta and Google Both Got More Expensive in 2026
Start with the cost side, because it’s the part every media buyer already feels. The all-industry average Meta CPM climbed from $11.82 to $14.19 this year, a 20% jump, according to Ryze AI’s 2026 benchmark data.
Some of that is a straightforward supply and demand story. As Google Search has gotten pricier, advertisers have been shifting budget into Meta, where CPMs still run far below Google’s per-click costs.
That migration crowds Meta’s auction and pushes its own CPMs up in turn, per Threadpoint’s account-level analysis.
Google didn’t get cheaper to compensate.
Cross-industry Search CPC rose about 12% year over year to roughly $2.96 in Q1 2026, the steepest annual increase since 2021, driven partly by AI Overviews pulling organic clicks out of the funnel and pushing more of that traffic into paid results, according to Digital Applied’s 2026 Google Ads benchmarks.
Put those two together and the practical effect is simple.
The same targeting, the same budget, and the same creative now buy less reach and fewer clicks than they did a year ago. Every downstream metric, CPA, CAC, and yes, ROAS, is built on top of that more expensive foundation.
At the Same Time, Reported ROAS Got More Generous
Here’s where it gets tricky.
Rising costs should, in theory, show up as pressure on reported ROAS. Instead, several platform-side changes have been working in the opposite direction, padding the number rather than letting it reflect the real cost increase.
Meta expanded engaged-view attribution in 2026 so that video ads now count a conversion after just a 5-second view followed by a purchase within a day, even with no click at all.
That change alone is estimated to have inflated reported ROAS by 15% to 25% on many accounts, simply by counting more activity as ad-driven, according to 1ClickReport’s analysis of the update.
Short-form video ads under 15 seconds see the biggest lift, since more of their traffic now clears the engagement bar.
This isn’t a new trick, it’s the same mechanic behind AdBeacon’s hero example of a client whose Meta dashboard showed 3.23x ROAS while independent, click-based measurement put the real number at 0.93x.
The gap comes down to what counts as “credit.”
Meta’s recent attribution changes keep adding more ways for an ad to get credit for a sale it may or may not have actually influenced, and every platform defines a “view” a little differently. None of it is independently verifiable, which is exactly the point of tracking it separately.
Why the Combination Is the Real Problem
Rising costs alone are manageable. Inflated reporting alone is manageable, if you know to discount it. Together, they hide each other, and that’s what makes this moment worth flagging.
Industry-wide, actual ecommerce ROAS dropped to 2.87x in 2026, down 4% year over year, with rising CPMs cited as a primary driver, according to Improvado’s 2026 ROAS analysis.
But that decline is happening underneath dashboards that, thanks to broader attribution windows, often look flat or even improved.
A media buyer glancing at a platform’s native ROAS column has no easy way to see that the number is now doing more work to look the same than it was twelve months ago.
This is also why “my ROAS looks great but the bank account doesn’t agree” has become such a familiar complaint among practitioners.
Forums full of media buyers describe watching a dashboard show strong returns while contribution margin tells a completely different story, and the disconnect usually traces back to the same root cause: a revenue-based metric with inflated attribution standing in for a profit-based one.
It’s an easy trap to fall into precisely because the dashboard number hasn’t done anything unusual. It just quietly absorbed a cost increase that should have shown up as a warning sign.
What to Actually Do About It
None of this means Meta or Google are broken, or that you should stop advertising where your customers are. It means the reported number needs a second, independent read before you trust it enough to scale a budget.
Compare platform ROAS to blended MER, not to itself.
Marketing efficiency ratio is calculated off total revenue and total spend, so it can’t be inflated by an attribution window change the way platform ROAS can. If Meta’s reported ROAS is climbing while blended MER is flat or falling, that gap is your signal, not the dashboard number itself.
Watch your own CPM and CPC trend line, not just the industry average.
Benchmarks are useful for context, but a 20% industry-wide CPM increase means little if you haven’t checked whether your account moved with it. Pull a rolling 30-day view and compare it to where you sat a quarter ago.
Separate click-based results from view-based credit.
Whatever attribution window a platform defaults to, know which conversions came from an actual click and which came from a view or an engagement. Understanding what’s actually driving your ROAS starts with knowing which part of it you can verify.
Re-check breakeven math before scaling.
With CPMs and CPCs both up, the ROAS you needed to hit breakeven six months ago may already be higher than it was. Scaling a campaign that clears an old breakeven number, calculated against an old cost basis, is how margin quietly disappears even while the dashboard says everything is working.
The businesses handling this well aren’t the ones panicking over rising costs. They’re the ones who stopped trusting a single platform’s version of ROAS to tell them whether a campaign is actually working. If you want to see what independent, click-based attribution looks like next to your own Meta and Google numbers, book a live AdBeacon demo and bring your current dashboard.
FAQ
Why is my Meta ROAS higher than it used to be even though CPMs are rising?
Meta expanded engaged-view attribution in 2026, so more conversions now qualify for credit after a short video view with no click. That change can inflate reported ROAS by 15% to 25% independent of any real performance improvement.
What’s the difference between platform-reported ROAS and actual ROAS?
Platform-reported ROAS includes conversions the platform credits itself with, including views and engagements it can’t independently verify. Actual ROAS, measured through click-based, first-party attribution, only counts conversions tied to a verified click.
Is a 20% CPM increase on Meta something to worry about?
Not automatically. A higher CPM paired with strong conversion rates can still be efficient. The risk is scaling spend based on a ROAS number that’s rising for reporting reasons rather than performance reasons while your real costs go up.
What should I track instead of platform ROAS alone?
Blended MER, calculated off total revenue and total spend rather than platform attribution, is harder to inflate and gives a more honest read on whether rising costs are actually being covered by results.
Why did Google Search CPCs rise even though Meta CPMs also went up?
Both platforms are absorbing more advertiser competition. On Google, AI Overviews have reduced organic click volume, pushing more advertisers into paid results for the same queries, while Meta is absorbing budget shifting away from increasingly expensive Google Search campaigns.