Profit-First Media Buying

Platform ROAS is a diagnostic. It tells you which ad set is outperforming which. It is not a business result, and running a company on it is how brands end up growing revenue while losing money.

Why platform ROAS overstates

Meta counts a conversion when someone who saw or clicked an ad later buys, inside an attribution window it defines. It has every incentive to claim credit generously, no visibility into your margins, and no idea that the customer was already going to buy after four emails and a Google search for your brand name.

None of that is scandalous — it is just what the tool is for. The error only becomes expensive when you treat the number as the scoreboard rather than the instrument panel.

The three numbers that replace it

1. Contribution margin per order

Selling price minus product cost, shipping, pick-and-pack, payment fees and a returns allowance. This is the money an order actually contributes before any fixed cost. It sets the ceiling on what you can pay to acquire a customer, and it is the input to your break-even ROAS.

2. Blended MER

Total revenue divided by total advertising spend. Every order, every channel, one number. It cannot be inflated by attribution windows because it does not use them — it only knows what came in and what went out.

Blended MER is deliberately crude. It will not tell you which campaign to turn off. What it will tell you, unarguably, is whether the machine as a whole is working.

3. Payback window

How long it takes a cohort of new customers to return the cash you spent acquiring them. A brand paying back in three weeks can scale aggressively on the same margins that would bankrupt a brand paying back in nine months. Cash timing, not just margin, decides how fast you are allowed to grow.

Setting your MER floor

Work backwards from the P&L rather than forwards from ambition:

1. Take monthly fixed costs — salaries, software, rent, agency fees.
2. Divide by your contribution margin percentage. That is the revenue you need before you make a penny.
3. Add your target profit, then divide again.
4. Divide that revenue figure by the ad spend you can afford. That is your MER floor.

Below the floor you are shrinking. At the floor you are static. Above it you are compounding — and the gap between your current MER and the floor is your permission to spend more.

The weekly operating rhythm

Cadence	What you look at	What you may change
Daily	Spend pacing, delivery breakages, anything obviously broken	Nothing structural. Fix errors only.
Weekly	Blended MER, new-customer CAC, creative winners and losers	Budget shifts, creative cuts, new tests live
Monthly	Contribution margin, cohort payback, channel mix	Target resets, structural changes, offer changes
Quarterly	LTV curves, category expansion, pricing	Anything

The discipline is in the right-hand column. Most accounts are damaged by daily changes made against weekly-scale noise — every intervention resets learning and the account never stabilises long enough to read.

Scale, hold or cut

• Scale when blended MER is comfortably above the floor and has held there for at least a week, and there is creative in the pipeline to support more spend.
• Hold when MER is near the floor, or when the last change is less than a week old and has not yet had a fair read.
• Cut when MER has been below the floor for two consecutive weeks and the cause is not a known one-off. Cut spend, then diagnose — not the other way round.

What this looks like in practice

You stop asking the media buyer for the ROAS. You ask what blended MER was, what it costs to acquire a new customer, and how that compares to the floor. The answer to “should we spend more?” becomes arithmetic instead of an argument.