"How much should I spend?" has no single answer, but it does have a logic. Spending too little starves the algorithm; spending too much on too little creative just burns money faster. Here is how to reason about it.
The floor: enough to learn
Meta's algorithm needs conversions to optimise. As a rough floor, a main conversion campaign wants on the order of 50 conversions per week to exit the learning phase and stabilise. Work backwards from your cost per purchase to find the spend that produces that. Below it, the account never settles and results stay volatile. This is the single most common reason small budgets underperform: not that the ads are bad, but that there is not enough spend for the system to learn.
The ceiling: how much creative can you feed?
Budget and creative volume are linked. More spend means the algorithm exhausts each ad faster, so it needs more fresh creative to keep learning. If you raise budget without raising creative volume, frequency climbs and costs rise. So your practical ceiling is set by how many quality ads you can produce, not by how much cash you are willing to spend.
The rough tiers we see
- ~5,000 EUR/month: enough to learn on one or two products, needs ~24 ads/month to stay fed.
- ~15,000 EUR/month: room to test properly across angles, needs ~75 ads/month.
- 50,000 EUR+/month: serious scale, needs 120+ ads/month and multi-channel structure.
Margins decide how aggressive you can be
A brand with high margins and strong repeat purchase can tolerate a higher cost per acquisition, because the lifetime value carries it. A thin-margin, one-time-purchase product cannot. Before setting a budget, know your margin and your repeat rate, because they decide how much you can afford to pay for a customer, which decides how far you can scale.
A simple way to set it
- Find your cost per purchase.
- Multiply by ~50 to get a weekly floor for one campaign, then by four for a monthly floor.
- Check you can produce enough creative to feed that spend (use the angles times concepts math).
- Scale up only as fast as both conversions and creative volume allow.
If you want the creative side of this worked out, see how many ads to test per month.