Most scaling advice sounds the same. Raise the budget 20% a day, duplicate the winners, accept that ROAS will drop as you go. We do it differently. Before a single rupee is added, the client and we agree on a floor: the return below which the account does not go. Then we scale only as far as the floor allows.
Setting the floor
The floor is not a target and not a wish. It comes from the unit economics: product margin, delivery cost, payment gateway fees, the cancellation rate, and the profit the founder actually needs per order. For a wooden toy brand we work with, that number came out at 4.0x. Anything above it is growth. Anything below it is a paid hobby.
How the week runs
- Monday: pull last week’s spend and delivered revenue by campaign. Not platform revenue, delivered revenue.
- Campaigns above the floor with headroom get a budget increase, capped at what the audience can absorb without frequency climbing.
- Campaigns near the floor hold. We test new creative inside them instead of feeding them more money.
- Campaigns below the floor for two consecutive weeks get cut, however good they looked in month one.
Why it works
Meta’s algorithm is very good at spending whatever you give it. It is not good at knowing when your margin has run out. The floor is the piece of information the platform does not have, and once it is written down, the budget conversation stops being emotional. Nobody argues about whether an ad feels like it is working. It is either above the line or it is not.
The result for that toy brand over a marathon season: budget up week on week, return never under 4.0x. Growth with a floor under it.