Return on ad spend is the most quoted number in performance marketing and the most misread. It divides revenue by spend, and revenue is not money you keep.
ROAS = revenue ÷ ad spend
Six thousand dollars of revenue on $1,500 of spend is a ROAS of 4.00, which the ROAS calculator also expresses as a 300 per cent return. Both figures are correct and neither tells you whether the campaign made money, because neither has met your cost of goods.
The number that decides it
The threshold is the point where gross profit equals ad spend, and it falls out of the margin alone.
break-even ROAS = 1 ÷ profit margin
On a 30 per cent margin, the break-even ROAS calculator gives 3.33. On 20 per cent it gives 5.00. On 60 per cent it gives 1.67, and on 80 per cent, the kind of margin a software business runs, it gives 1.25.
So the same 4.00 ROAS is a comfortable win for the software business, a solid result at a 30 per cent margin, and a loss at 20 per cent. Put the numbers together: $6,000 of revenue at a 20 per cent margin is $1,200 of gross profit, against $1,500 of ad spend. The campaign returned four times its spend and lost $300.
That is the whole trap. ROAS has no idea what you sell. A furniture retailer and a jeweller running identical campaigns with identical ROAS are in completely different positions, and nothing on the ads dashboard distinguishes them.
Target CPA is the same statement in a more useful unit
Margin and order value together give you a maximum acceptable cost per acquisition, which is easier to act on than a ratio. At a 30 per cent margin on an $80 order, the break-even cost per acquisition is $24. At 20 per cent it is $16, and at 60 per cent it is $48.
Those are numbers a buyer can work with directly, because the cost per acquisition calculator reports actual CPA in the same unit: $2,000 of spend across 40 conversions is $50 each. Against a $24 target that is unambiguous in a way that a ratio against a benchmark never is.
Three things that make it worse
Margin is usually taken from a blended figure rather than from the products the ads actually sell, and ads tend to sell the discounted lines. Shipping, payment processing and returns often sit outside the margin number people quote, and all three scale with orders. And attribution generally credits the platform with revenue that would have arrived anyway, which inflates the numerator before any of this starts.
None of that is an argument against measuring ROAS. It is an argument for knowing your break-even first, so the number has something to be compared against. A campaign is not good because the ratio is large; it is good because the ratio is larger than one divided by your margin.
For the margin figure itself, there is the margin and markup calculator, which is also where the other classic version of this mistake lives: a 50 per cent markup is a 33 per cent margin, and feeding the markup in by mistake puts the break-even ROAS at 2.00 when it should be 3.00.