A 4x ROAS can still lose money because ROAS ignores your margin, and platform ROAS is inflated on top of that. What matters is POAS, profit on ad spend, the money you actually keep. At a 30% contribution margin, a reported 4x ROAS (roughly 2.6x real) loses money on every cedi of spend. Buy to profit, not to the dashboard.
ROAS measures revenue. Your bank measures profit.
ROAS, return on ad spend, only tells you how much revenue an ad brought in per cedi spent. It says nothing about what that revenue cost you to fulfil. A 4x ROAS on a product with fat margins is a great day. The same 4x on a thin-margin product can be a quiet loss.
Two things compound the problem in practice. First, margin: after cost of goods, shipping, payment and Mobile Money fees, and COD returns, the slice you actually keep is far smaller than the revenue figure. Second, inflation: as we covered in why your Meta ROAS is higher than your actual sales, the platform number overstates reality. Stack those together and a headline ROAS can be twice the truth.
A worked example, in Cedis
Say Meta reports a 4.0x ROAS. Allow for typical attribution inflation and the real figure is closer to 2.6x. Now run that real 2.6x against different contribution margins on GHS 10,000 of ad spend, and watch where profit actually lands:
| Contribution margin | Breakeven ROAS | Real 2.6x ROAS | Profit on GHS 10,000 |
|---|---|---|---|
| 40% | 2.50x | Above breakeven | +GHS 400 |
| 35% | 2.86x | Below breakeven | −GHS 900 |
| 30% | 3.33x | Below breakeven | −GHS 2,200 |
| 25% | 4.00x | Below breakeven | −GHS 3,500 |
What POAS actually is
POAS, profit on ad spend, is the metric ROAS should have been. Instead of revenue over spend, it measures the profit left after all variable costs, divided by the ad spend that produced it. A POAS above 1 means the campaign made money; below 1 means it lost money, no matter how flattering the ROAS looked.
Because it accounts for margin, POAS cannot be gamed by selling more low-margin volume. It is the number that agrees with your bank balance, which is exactly why it is the number we optimise toward.
How to buy on POAS when Meta only knows revenue
Meta's ad account does not natively know your margins, so you translate profit into a ROAS target it can optimise to. Work out your breakeven ROAS (one divided by contribution margin), add your target profit, and buy to that number. Our breakeven ROAS calculator does the maths for you.
Then close the loop: feed real order values back to Meta server-side through the Conversions API so the algorithm learns from money that actually banked, and judge the account on blended return against margin, not on the in-platform ROAS of any single campaign.
In Naira and Cedis, the margin math is unforgiving
African stores carry cost structures a US ROAS benchmark never sees: Mobile Money and card fees, higher delivery costs, and cash-on-delivery returns that wipe out a share of orders before they ever pay. Each one shaves points off contribution margin, which pushes your breakeven ROAS higher.
That is why POAS discipline matters more here, not less. The stores that scale profitably in Ghana and Nigeria are the ones that know their real margin, translate it into a ROAS target, and hold spend to it, while the ones chasing a vanity 4x quietly fund their own losses.