ROAS is reported by the ad platform and counts revenue it decided to claim, so three platforms will between them claim more revenue than your store actually took. MER divides your own revenue by your own total spend, so nothing in it can be claimed twice. POAS replaces revenue with contribution margin, so it knows the difference between selling something and selling something worth selling. Steer the business on MER and POAS; keep ROAS for comparing two creatives inside one account on one day.
Why platform ROAS adds up to more than you sold
Ask three ad platforms how much revenue they produced last month and the total will usually be larger than your Shopify revenue for the same month. Nobody is lying. Each platform counts a conversion it can plausibly attribute inside its own window, and a customer who saw an Instagram ad on Tuesday, searched your brand on Thursday and bought on Friday is a legitimate claim for two platforms and arguably three.
That overlap is not a bug in their reporting, it is what attribution windows do, and it is why summing platform ROAS is arithmetic that cannot work. There is no shared numerator. Add to that the fact that each platform's window and model changes without notice, and you have a number that moves for reasons unrelated to anything you did.
None of that makes ROAS worthless. Inside one account, on one day, comparing two creatives measured the same way, it is a perfectly good relative signal. The mistake is promoting it to a business metric.
MER: the ratio that cannot double count
The marketing efficiency ratio is your revenue divided by your total advertising spend across every source, taken from your own store rather than from the platforms. One numerator, one denominator, no attribution model in sight. It is deliberately blunt: it will not tell you which campaign worked, and it is not meant to.
Which revenue figure goes on top matters more than most articles admit. Gross revenue includes VAT you never keep, so a German store computing MER on gross is flattering itself by roughly 19%. Net revenue, after discounts, returns and VAT, is the honest denominator for every cost ratio in the business, and using it consistently is what makes MER comparable across months in which your discount depth or return rate changed.
On the example store used throughout this site, an apparel catalogue doing about €10,6M a year, net revenue over fourteen months was €12.400.114,85 against €3.144.072,34 of total ad spend. That is a MER of 3,94: every euro advertised was accompanied by €3,94 of net revenue.
POAS: the version that knows what you kept
MER still has a blind spot, and it is a big one on a mixed catalogue. Two products can produce identical revenue and hand you completely different amounts of margin. A campaign pushing the thin one looks excellent on any revenue-based ratio while losing money on every order it produces.
Profit on ad spend fixes that by replacing revenue with contribution margin. Which margin you use is a real choice: CM2, after goods, fulfilment and payment fees, is the sharpest one, because it is everything the order earned before you paid to acquire it. The same example store carried €6.060.903,38 of CM2 against that €3.144.072,34 of spend, which is a POAS of 1,93. Every euro advertised returned €1,93 of contribution margin.
ROAS asks whether the ad sold something. POAS asks whether selling it was worth doing.
A POAS above 1 means the advertising paid for itself out of contribution margin. What it needs to be depends entirely on the overhead it also has to cover, which is why it belongs beside your fixed costs and not on a leaderboard. The ladder those margins come from is set out in CM1, CM2 and CM3 explained.
Blended CAC, the third one worth having
Blended customer acquisition cost is total marketing spend divided by the number of new customers won in the same period. Every euro, every new customer, no platform claims involved. The example store's blended CAC in August 2026 was €35,46.
The number on its own means nothing. Compare it against what a first order actually contributes, not against what it sells for. If the first order's CM2 is below your CAC, you are buying customers at a loss and betting on the second order, which is a strategy rather than a mistake, but only if you know you are doing it.
The three side by side
| Ratio | Formula | Value | Answers |
|---|---|---|---|
| MER | Net revenue ÷ total ad spend | 3,94 | Is the machine running at a ratio we can afford |
| POAS | CM2 ÷ total ad spend | 1,93 | Is what we sold worth having sold |
| Blended CAC | Total ad spend ÷ new customers | €35,46 in Aug | What one new customer costs |
| Platform ROAS | Attributed revenue ÷ spend | Not comparable | Which creative beat which, inside one account |
What to do on Monday morning
Set a MER floor and a POAS floor for the business, derived from your fixed costs rather than from an article. Work out what CM3 has to be to cover a month of overhead, then work backwards to the spend and the ratio that gets you there. The MER, POAS and blended CAC calculator does that arithmetic on your own spend and margin, and gives you two numbers that mean something specific to your store rather than a benchmark from someone else's catalogue.
Then use the platforms for what they are good at. Inside Meta, compare Meta creatives. Inside Google, compare Google campaigns. Do not add their revenue figures together, do not report them upwards, and do not let a platform ROAS of 4 talk you into a budget increase that your blended numbers say you cannot afford.
One honest caveat about all of this. Blended ratios tell you whether marketing overall is profitable. They will not tell you which channel to cut, and no amount of arithmetic on your own data can, which is the trade blended measurement asks you to make. For most stores under €10M, knowing reliably that the whole machine is profitable beats knowing unreliably which part of it was responsible.