Every new ad forces the same decision: how long do you let it run before you judge it? Cut too early and you'll bury creatives that would have worked once the algorithm settled. Wait too long and you're just paying to confirm a loser. There's no magic number, but there is a principled way to decide, and it's tied to your break-even CPA, not your gut.
Why "give it 24 hours" is the wrong frame
Time isn't what makes a result trustworthy, data is. An ad that spent $4 in a day has told you almost nothing, whether that took one hour or twelve. The right question isn't "how long has it run?" but "has it spent enough to show something real?"
Small numbers lie
Early results swing wildly because the sample is tiny. With two purchases, one refund or one lucky order changes your CPA by 50%. This is ordinary statistical noise: small samples are volatile, and the metrics haven't had enough events to stabilize. So the first job of a test isn't to find a winner, it's to spend enough that the numbers stop bouncing.
Anchor the decision to break-even CPA
A useful, simple rule of thumb: give an ad roughly one to three times your break-even CPA in spend before you judge it. The logic is that your break-even CPA is, by definition, the spend level at which you'd expect about one conversion if the ad were viable. Spending only a fraction of it and seeing zero sales tells you very little; spending a multiple of it and still seeing zero is a real signal.
If you don't know your break-even CPA yet, stop and get it first with the break-even ROAS calculator, everything here depends on that number.
| Break-even CPA | Rough spend before judging |
|---|---|
| $20 | ~$20–$60 per ad |
| $40 | ~$40–$120 per ad |
| $75 | ~$75–$225 per ad |
Use leading indicators to stop early
You don't always have to wait for the full budget. Some signals let you cut sooner without much risk:
- Very low CTR at meaningful impressions usually means the creative won't earn clicks, you can often cut before spending a full test budget.
- Lots of clicks, zero add-to-carts points at the store or offer, not the ad, and may warrant fixing the page rather than killing the creative.
If you're unsure which number is failing, read what each metric is telling you, it maps each symptom to a cause.
A simple testing rule you can actually follow
- Know your break-even CPA before you launch.
- Budget each new ad ~1–3× that number.
- Cut early only on a clear leading indicator (e.g. dismal CTR at volume).
- Otherwise, let it reach the spend threshold before deciding.
- Judge the final CPA against break-even, then cut, keep, or scale.
The point isn't to be patient or ruthless. It's to stop making decisions on samples too small to mean anything.
Learn this on live campaigns, with mentorship
The Masterclass covers testing, reading metrics, and when to cut or scale, inside a private Discord you can ask questions in before you buy.
Written by Duper. Duper runs the Shopify Scale Masterclass and teaches the TikTok Ads side of the program.