TL;DR
Platform ROAS is what Meta's dashboard reports: attributed revenue divided by Meta ad spend, inside your attribution window. Blended ROAS is total business revenue divided by total ad spend, no attribution needed. They diverge because attribution is an estimate, not a measurement: windows miss late conversions, view-through gets modeled, and Meta takes credit for sales that would have happened anyway. Use platform ROAS for comparing creatives against each other, blended ROAS for deciding whether the business can afford to scale. Trusting either one alone is how you make confident, wrong decisions.
Your dashboard says 4.2x. Your bank says 2.1x.
Every advertiser hits this moment. The Meta dashboard shows a beautiful ROAS. The finance team shows a spreadsheet that disagrees. Someone is wrong, and everyone has a theory about who.
The Meta-believers say the dashboard is undercounting (privacy gaps, missed attribution). The finance-believers say the dashboard is overcounting (Meta taking credit for everything). They're both right, which is why this argument never ends and why you need both numbers instead of picking a side.
Let's define them properly, because half the confusion in this debate comes from people using the terms loosely.
What platform ROAS actually measures
Platform ROAS is simple arithmetic on top of complicated estimation: revenue Meta attributes to your ads, divided by what you spent on Meta, within your chosen attribution window (commonly 7-day click, 1-day view).
The key word is "attributes." Meta doesn't observe every sale your ads caused. It observes clicks and views, observes some conversions through the pixel and Conversions API, models the rest, and then assigns credit according to its rules. Change the attribution window and the same account shows a different ROAS. That's not a rounding error. That's the number being a product of your settings.
Platform ROAS is also, not coincidentally, the number Meta most wants you to look at. It's the number that makes the platform look good. That doesn't make it useless. It makes it interested. Treat it like a witness with a stake in the outcome: useful testimony, verify independently.
What blended ROAS actually measures
Blended ROAS skips attribution entirely: total revenue divided by total ad spend, usually across all channels, over the same period.
No windows. No models. No arguments about whether that sale "belongs" to Meta or Google or the email that arrived three days earlier. Just money in versus money out. (When people say MER, marketing efficiency ratio, they usually mean the same idea extended to all marketing spend, not just media.)
Blended is harder to game and impossible to attribute. It tells you whether the business is healthy. It tells you nothing about which ad deserves the credit, which is exactly why you can't run an account on it alone.
Why they diverge
The gap between the two numbers isn't a bug. It's the sum of everything attribution can't see and everything it overclaims.
Attribution windows miss late conversions. On a 7-day click window, a customer who clicks on Monday and buys the following Tuesday doesn't exist in platform ROAS. For products with consideration cycles longer than the window, platform ROAS systematically undercounts. Your ads look worse than they are.
View-through and modeled conversions inflate. On the other side, 1-day view attribution hands Meta credit for sales where the "view" was a video autoplaying below the fold while someone scrolled past. Some of those sales were genuinely influenced. Many weren't. For brands with strong organic or repeat-purchase behavior, platform ROAS systematically overcounts. Your ads look better than they are.
Incrementality is the ghost in the machine. The deepest problem: neither number measures what would have happened without the ads. A customer who was going to buy anyway, clicks a retargeting ad out of habit, and converts: platform ROAS counts it as a win. Blended ROAS counts the revenue but can't isolate the cause either. True incrementality requires holdout tests, which almost nobody runs, which is why almost everybody argues about the wrong numbers instead.
iOS and privacy gaps. Since ATT, a meaningful share of conversions never gets reported back to the platform at all. This mostly pushes platform ROAS down relative to reality, which is why the "dashboard is undercounting" camp has a point. But it's not the whole story, because the overcounting mechanisms above are still running at full speed.
Net effect: the two numbers can diverge by 30 to 50 percent or more in either direction, and both directions are "correct" given what each one measures.
Which one to trust (trick question)
Neither. You use each for what it's good at.
Platform ROAS is for relative decisions. Comparing creative A against creative B. Judging whether the new angle beats the old one. Deciding which ad set gets more budget within Meta. For these, the absolute number doesn't matter much. What matters is that the measurement bias applies roughly equally to everything being compared, so the ranking is informative even if the values are fuzzy.
Blended ROAS is for absolute decisions. Can we afford to scale spend 50 percent? Is the business actually growing? Should marketing get more budget next quarter? These are money-in-money-out questions, and only blended answers them honestly.
The classic failure modes: using platform ROAS to decide scale ("dashboard says 5x, let's triple the budget" while blended is 1.8x and the business is bleeding), and using blended ROAS to judge creative ("blended dropped, must be the new ads" when actually the email program paused and Meta had nothing to do with it).
How to use both without going insane
Track both, weekly, side by side. The absolute values matter less than the trend and the ratio between them. A stable ratio means your measurement world is stable. A shifting ratio means something changed in attribution, incrementality, or the business, and that's worth investigating.
If platform ROAS is flat but blended drops: something outside Meta broke, or incrementality collapsed. Check other channels, check the site, check the offer. Don't rewrite your ads.
If blended is flat but platform ROAS drops: usually a Meta-side issue. Creative fatigue, auction pressure, tracking problems, or an attribution setting someone changed. This is a platform problem until proven otherwise.
If both drop together: congratulations, it's real and it's everywhere. Start with the diagnostic playbook: offer, site, tracking, then creative.
Set targets on blended, manage to platform. Your business target ("we need 2.5x blended to grow") is the boss. Your day-to-day levers (which creative wins, which ad set scales) run on platform metrics. Translate between them with your historical ratio, and recheck the translation quarterly because it drifts.
FAQ
What is the difference between blended ROAS and platform ROAS? Platform ROAS is attributed revenue divided by Meta spend, inside your attribution window. Blended ROAS is total revenue divided by total ad spend, with no attribution at all. One measures what the platform claims; the other measures what the business received.
Why is my Meta ROAS higher than my actual ROAS? Usually a mix of view-through attribution taking credit for sales it barely influenced, repeat customers converting through retargeting ads they would have clicked anyway, and short attribution windows interacting with your buying cycle. Check the blended number before celebrating.
Can platform ROAS be lower than true performance? Yes. iOS privacy gaps mean some conversions never get reported, and short attribution windows miss late buyers. For long consideration cycles, the dashboard routinely undercounts real impact.
What is a good blended ROAS? It depends entirely on your margins and growth goals. A 2x blended ROAS is fantastic at 70 percent margins and fatal at 20 percent. Work backwards from unit economics, not from benchmarks.
Should I optimize my account toward blended ROAS? You can't optimize directly to it inside the ad account, since Meta only sees its own attributed data. Use blended as the business target and platform metrics as the steering wheel, and check regularly that the two still agree with each other.
Written by the MetaMaxd team. We run Meta ads for brands that are tired of agency theater: no vanity metrics, no "hacks," just accounts engineered to make money. If your dashboard and your bank account are telling different stories, that's what our audits are for.
