Meta’s dashboard says 6x. Google says 4x. Add the revenue those platforms each claim and you’ve apparently made more money than your store actually took in. Every platform grades its own homework — and if you’re making budget decisions off in-platform ROAS alone, you’re almost certainly over-spending somewhere and starving the channel that’s quietly doing the work. Here’s how we reconcile the numbers for brands, and the simple measurement stack that ends the arguments.
Why the platforms disagree with your bank account
Ad platforms attribute on clicks and views, each using its own window. A customer who saw an Instagram ad on Tuesday, clicked a Google Search ad on Thursday and bought on Friday shows up as a full conversion in both dashboards. Neither platform is technically wrong — they’re answering the question “did we touch this sale?”, not “did we cause it?”. Stack three or four channels and the double-counting compounds until your reported revenue is 140–180% of reality.
That’s not a reason to abandon platform metrics. It’s a reason to know what each number is for.
The three-layer measurement stack
Layer 1: Blended ROAS — the truth layer
Total revenue ÷ total ad spend, across everything. No attribution model, no arguments. It can’t tell you which channel to scale, but it tells you whether marketing is working at all — and it’s the only ROAS number your finance team should ever see. Watch it weekly as a trend line, not a daily target: blended ROAS moves slowly, and reacting to daily wobbles causes more damage than it prevents.
Layer 2: Platform ROAS — the steering layer
In-platform numbers are inflated, but they’re consistently inflated. That makes them useless in absolute terms and very useful in relative terms: campaign A versus campaign B inside the same account, this month versus last month, this creative versus that one. Use platform ROAS to steer within a channel, never to compare across channels.
Layer 3: Incrementality checks — the honesty layer
A few times a year, test what happens when you turn something off. Pause branded search in one region for two weeks. Cut retargeting spend by half and watch whether revenue actually drops. Geo holdouts and spend-step tests are the closest most brands get to ground truth — and they routinely reveal that the “best performing” campaign in the dashboards is mostly harvesting sales that would have happened anyway.
Set one north-star target in blended terms — a MER or blended ROAS floor tied to your margins — then let each platform fight for budget with its own relative numbers. When a channel manager says “my ROAS is 6x, give me more budget”, the question is always: did blended move when we scaled you last time?
The retargeting trap
Nowhere is the gap between platform and blended ROAS wider than retargeting. Serving ads to people already on their way to buy produces spectacular dashboard numbers — 15x, 20x — and very little incremental revenue. If your retargeting budget has crept above 20–30% of spend and your blended ROAS hasn’t improved, you’re paying to interrupt people who were already coming back. Shift that budget to prospecting and give it a quarter; most brands see blended hold steady while new-customer revenue grows.
What good looks like in practice
- One weekly scorecard with blended ROAS / MER, new-customer CAC, and total revenue — the numbers that survive contact with a P&L.
- Platform dashboards used for optimization only — creative winners, audience fatigue, bid strategy health.
- A quarterly incrementality test on your biggest line item. One honest holdout beats a year of attribution debates.
- UTMs and server-side tracking kept clean, because every measurement layer above depends on the plumbing underneath.
None of this requires an enterprise data stack. It requires agreeing, once, on which number means what — and having a performance marketing partner disciplined enough not to grade their own homework.
Frequently asked questions
What is a good blended ROAS for an e-commerce brand?
It depends entirely on gross margin. A 70%-margin brand can be profitable at 2.5x blended while a 30%-margin brand may need 4x+. Work backwards from contribution margin — the right target is the one where an incremental order still makes money after product, shipping and ad costs.
Why is my Meta ROAS higher than my actual revenue suggests?
Meta counts view-through conversions and claims credit for sales other channels also touched. The number is consistent enough to compare campaign against campaign — but it should never be read as literal cash generated.
Should I use Google Analytics as the source of truth instead?
GA4 is last-click biased, so it undercounts view-influenced channels like Meta and YouTube — it swings the error the other way. Use blended ROAS from real revenue as truth, platform numbers for steering, and GA4 as a tie-breaking third opinion.
How much should retargeting be of my total ad budget?
For most established brands, 10–20% of paid social spend is plenty. Above 30%, you’re usually paying to interrupt people who would have converted anyway — and starving the prospecting that actually grows the business.
The takeaway
Platform ROAS steers the ship; blended ROAS tells you whether it’s moving. Keep finance on blended, keep channel managers on relative platform numbers, and audit the whole thing with an occasional holdout test. The brands that scale profitably aren’t the ones with the prettiest dashboards — they’re the ones that know which of their numbers is lying, and by how much.