The diagnosis says increase budget. I can't. What now?

When budget is below the learning threshold and can't move, you have two levers: reduce what you're asking the algorithm to do, or change what you're optimising for.

1. Consolidate to one ad set. Split budgets multiply the problem — each ad set needs 50 conversions/week independently. One ad set at your full budget is always better than two at half.
2. Move up the funnel. If Purchase isn't hitting 50/week, switch to Add to Cart or Initiate Checkout. These fire 3–5x more often and give the algorithm enough signal to stabilise.
3. Raise your CPA target temporarily. If your break-even CPA is €40 and you're targeting €25, the math doesn't close at any budget. Test at break-even CPA for 2 weeks before judging.
4. Accept longer evaluation windows. Below threshold, week-over-week results are noise. Judge performance over 3–4 weeks minimum, not 7 days.
5. Kill underperforming creatives faster. At low volume every impression matters. One weak creative dragging down the ad set costs more proportionally than at scale.