How much should you spend testing ad creatives? Learning-phase maths, per-creative kill thresholds and a monthly testing plan. Free.
"How much should I spend testing creatives?" is one of the most-asked questions in every ads community β and most answers are vibes. This calculator uses the actual mechanics of ad platforms: the ~50-conversion learning phase, the 1β3Γ CPA kill rule that media buyers actually use, and your own margins. The result is a concrete monthly testing budget, a weekly cadence, and kill/scale rules you can hand to anyone running your ads.
Meta and Google need roughly 50 optimization events per ad set before their algorithms stabilize. An ad set that never exits learning will always look worse than it is β you'll kill good creatives and scale noise. That's why the plan budgets learning-phase spend per ad set separately: it's the price of trustworthy data, not waste.
The kill rules are the most valuable part of this page. A creative with zero conversions after 1Γ CPA spend is telling you something β listen. Every day you let a loser spend is budget stolen from the next test. Print the rules, stick to them for a month, and your cost per acquisition will thank you.
Ad platforms need ~50 conversions per ad set before their delivery algorithms stabilize. Until then, performance data is noisy β budget for it deliberately instead of judging creatives early.
It depends on budget, not ambition: each creative needs 1β3Γ your target CPA before you can judge it. Divide your testing budget by that number β that's your real creative capacity.
1Γ is ruthless and best for tight budgets; 2Γ is the industry standard; 3Γ suits high-ticket products where conversions are rare and noisy. Pick one and apply it consistently.
They do different jobs: prospecting finds new customers (test here), retargeting harvests warm traffic (scale winners here). Mixing them muddies your test reads.
Explore our full collection of free online tools β from image compressors to business calculators, all free and no signup required.