MER, ROAS and paid cost per new customer: which one runs the budget
ROAS is a platform's own claim. MER is all of the store's revenue over all ad spend. Paid cost per new customer is Meta and Google spend over the new customers the store recorded. Convey Growth reports all three every week, labeled, and moves budget only on the third, against a ceiling written for the brand before spend goes up.
The three numbers.
ROAS, return on ad spend. The revenue a platform credits to its own ads, divided by what you spent there, inside the platform's own window. Good for comparing one ad against another in the same platform. Wrong for the total budget, because every platform claims the same orders.
MER, marketing efficiency ratio. All of the store's revenue over all ad spend: repeat orders, subscriptions and email included. Good for watching whether the business as a whole gets more efficient. It isn't a return. Most of that revenue would have come in anyway, so a healthy MER can hide ads that lose money.
Paid cost per new customer. Meta and Google spend over the new customers the store itself recorded in the same window. Nothing else folded in. It's the one number no platform can inflate, and it decides whether the next dollar goes in.
A ceiling, from one brand's record.
At a bakery we run, the ceiling on paid cost per new customer was written before spend went up. The week the number went over it, the budget step waited.
| What | Value | Window |
|---|---|---|
| paid spend per new customer, the first week after launch | $65 | Aug 28 to Sep 3, 2026 |
| paid spend per new customer, since launch | $55 | Aug 28 to Oct 8, 2026 |
| the ceiling on it, written before spend went up | $55 | since launch |
| the latest week, when the budget step was held | $61 | Oct 2 to 8, 2026 |
Observed results from the store's own record, not isolated causal lift.
Where the ceiling comes from.
From the brand's own margins. Take what a new customer is worth over time after product, shipping and fees, and set the ceiling below it. A brand with strong subscriptions can afford a higher ceiling than one that sells a single gift box.
The ceiling is written down before spend goes up, so nobody sets it after seeing the results.
Read the next dollar, not the average.
An account can average a fine number while the last dollar added buys customers at twice the ceiling. The average mixes the best audiences with the newest test. So the read for a budget decision is the campaign getting the money, not the account.
When paid cost per new customer on that campaign holds under the ceiling for a full week, spend steps up. When it doesn't, the step waits. That's the whole rule.
Questions.
Is a high MER good?
Not on its own. It depends on margins and on how much of the revenue is repeat. A brand with a big subscriber base can show a strong MER while its ads lose money on every new customer. Read it beside paid cost per new customer.
Why not use the platform's cost per purchase?
It counts the purchases the platform claims, repeat buyers and orders another channel also claimed included. The store's own count of new customers is the one the platforms can't move.
Do you report ROAS?
Yes, as each platform's own claim, labeled with its window, next to MER and paid cost per new customer. Three numbers, three labels, every week.
Send us your URL.
Get a free written read in 48 hours. No login, no cost, no call.
Sent. The read lands in 48 hours.