The ad budget has doubled in a year, every platform's dashboard reports a healthy return, and the store's actual profit has gone down. Someone proposes cutting the ads; someone else points at the dashboards; the argument runs monthly and settles nothing, because both sides are reading numbers that cannot answer the question. Whether advertising is eating your profit is not a matter of opinion or attribution, it is one division with a threshold, and this post builds it: the blended view that no platform can inflate, the MER floor your own unit economics set, the difference between the average euro and the marginal one, and the order in which to cut when cutting is the answer, which it is less often than the bank balance suggests.
- Never judge ad profitability from the platforms' own reports: each one claims the conversions it saw, the claims overlap, and their sum routinely exceeds the store's takings.
- Your unit economics set a hard threshold: on the example store, every ad euro must bring €1,90 of net revenue just to break even on an order, its MER floor.
- A healthy average hides a losing edge: the last euro of a scaled budget buys less than the first, so judge budget changes by the marginal euro, never the average.
- Half of all 'ad problems' are margin problems wearing a media hat: check goods drift, basket size and returns before touching a campaign that used to work.
Stop asking the platforms whether the ads work
Every platform's return metric is computed from the conversions its own tracking claims, every platform claims everything it plausibly touched, and the same order is routinely claimed by two or three of them. Their reports are useful for comparing campaigns inside one platform; as a verdict on the whole budget they are structurally incapable of honesty, not from malice but from arithmetic. The blended view fixes it in one line: the store's actual revenue over the total actually spent, both from systems that cannot double-count. Which ratio to standardise on, and what each one is good for, is its own subject, MER, POAS or ROAS, but the principle carries regardless: numerator from your orders, denominator from the billing pages, nothing from an attribution model.
Your MER floor: the number that ends the debate
The floor comes from what an order keeps before advertising. On the example store, an €85 order contributes €44,68 after goods, parcel and payment fee, about 52,6% of its net revenue. Invert that share and you have the threshold: the store needs roughly €1,90 of net revenue for every ad euro simply to stand still, because at that ratio the contribution the revenue brings exactly equals the euro that bought it. Blended MER above the floor: the ads are building profit, whatever the bank feels like this week. Below it: every ad euro buys a loss, whatever the dashboards claim. One afternoon with your own margins produces your own floor, and a MER calculator does the division and the what-ifs interactively. The full arithmetic from ad spend down to the statement is the companion guide on profit after ad spend.
The average hides the marginal euro
A blended MER of 2,4 against a floor of 1,90 sounds comfortable, and can still contain a loss, because the average blends cheap demand with expensive demand. The first euros of budget harvest the people already looking for you; the last euros of a scaled budget chase colder audiences at worse rates. Scaling therefore decays the marginal return long before it drags the average below the floor, and a store that only watches the average will overspend for months while feeling fine. The test is empirical, not clever: change the budget in steps, hold each step for long enough to read, and compare the extra revenue each step bought against the extra spend it cost. Where the marginal ratio crosses your floor is your budget's honest ceiling, and it moves with seasons, creatives and competition, so it is a measurement to repeat, not a constant to find once.
The diagnosis: is it the ads at all?
Before any campaign is touched, rule out the impostors, because ad spend is the most visible cost in the building and gets blamed for leaks it did not cause. A rising customer acquisition cost with steady margins is genuinely an ad problem. But if acquisition costs are flat while profit falls, the leak is elsewhere: goods creeping, baskets shrinking under the per-order cost stack, or returns handing wins back, the case covered in returns killing your margins. And if profit per order was thin all along, growth simply multiplied it into visibility, the pattern behind high revenue and no profit. The table sorts it:
| Symptom | Actually | Check first |
|---|---|---|
| CAC rising, margins steady | an ad problem | marginal MER by budget step; creative fatigue; audience saturation |
| CAC flat, profit falling | a margin problem | goods cost drift, basket size, fee and parcel share |
| Revenue growing, profit flat | thin unit economics, multiplied | the per-order walk, before and after the ad share |
| Strong months, weak follow-ons | returns maturing late | refund totals against the prior month's sales |
The cut order, when cutting is the answer
- Creatives first: fatigue is the commonest cause of marginal decay, and refreshing them costs production money, not reach.
- Then audiences and campaigns: kill the segments whose measured contribution sits under the floor, keep the ones above it, and resist averaging the two into a verdict on 'the ads'.
- Then scheduling and geography: money spent in hours or markets that never convert is the cheapest cut in the account.
- Budget last, in steps, reading the marginal ratio at each step down exactly as you would on the way up.
- Never the blind pause: switching everything off to 'see what happens' destroys the measurement along with the spend, and what happens is usually a quieter month you learn nothing from.
Three questions that always come next
Should returning customers count in the ratio? Compute both. The blended figure over all revenue is the store's reality and the number the floor applies to. The new-customer version, spend over first-order revenue, is stricter and more honest about acquisition specifically, since repeat orders would have arrived largely without this month's ads. A store healthy on blended and sick on new-customer is living off its past, which is fine as a season and dangerous as a strategy.
What about brand campaigns that pay off later? They exist, and they are also the last refuge of every underperforming budget, so give them the discipline their defenders claim they deserve: a stated horizon, a stated expected effect, and a review date on the calendar. Meanwhile the floor still applies to the account as a whole across a quarter; a budget persistently below it is not investing in the future, it is mispricing the present.
Does the floor change? Constantly, because it is made of margins: a supplier increase, a heavier product mix or a returns spike all raise the revenue each ad euro must bring, and a price rise or cheaper parcel lowers it. That is the deepest reason ad decisions belong on a daily statement rather than a quarterly review: the floor your budget answered to in March may simply no longer exist in June, and nouz recomputes it nightly from the same orders and costs the rest of the statement reads.
The argument, retired
Ads eating your profit is a checkable claim, not a mood. Build the floor from your own margins, hold the blended ratio against it, test the marginal euro at the edges of the budget, and rule out the margin problems wearing a media hat. Sometimes the answer really is cut, and now it comes with a place to cut first. More often the answer is that the ads are fine and something quieter is eating the store, and the monthly argument finally has somewhere better to look.