Free tool · Loki
Use this creative testing budget calculator to estimate how much spend you need to properly test ad creative concepts and variants across paid social platforms.
Suggested test budget
£14,400
Spend per variant
£1,200
Spend per concept
£3,600
Test matrix — 4 concepts × 3 variants = 12 cells
Underpowered — needs ~29 days at this budget
Enter your monthly budget and the shape of your test — how many concepts and variants, over how many days. Set your target CPA (or a target ROAS and order value), and the minimum conversions each variant needs before you trust it. The calculator returns the suggested test budget, the spend per concept and per variant, the minimum viable period, and a clear winner threshold.
Creative testing only works when each concept and variant receives enough budget to generate a meaningful signal. If budget is spread too thinly across too many assets, performance data can become noisy and the test may not produce a reliable winner. The maths is simple: every cell in your test matrix needs roughly its minimum conversions at your target cost, so total budget scales with the number of variants — not just the number of ideas.
Spend per variant is your minimum conversions multiplied by your target cost per conversion (for a ROAS target, that cost is order value ÷ target ROAS). Multiply by the number of variants for the suggested test budget, and compare it with what your daily budget can deliver across the test window. If the window cannot fund it, the test is underpowered — the result will look like a winner but really be noise.
If the test is powered, run it and call a winner only once a variant clears the conversion threshold at or below target cost. If it is underpowered, you have three honest options: extend the window, add budget, or test fewer variants. The last is usually the cheapest — and the creative volume calculator helps you decide how many to carry. For short-form video, narrow the test to openings with the hook testing calculator.
This calculator plans a test before it runs. ElenIQ’s Loki watches it while it runs — tracking each concept and variant, flagging tests that are underpowered, and surfacing which creatives are genuinely winning rather than riding variance, so testing budget turns into reliable decisions.
Enough for every variant to reach a meaningful number of conversions. The rule of thumb is minimum conversions per variant × target cost per conversion, multiplied by the number of variants. Spread budget too thin and the data is too noisy to trust a winner.
A common minimum is around 30–50 conversions per variant before you read a result, though higher is better. Below that, normal variance can make a weak creative look strong or hide a genuine winner, so the test is effectively underpowered.
A concept is a distinct creative idea; a variant is a version of it. You usually test several variants per concept so that one weak execution does not bury a strong idea. The total cells in your test matrix — concepts × variants — drives the budget you need.
An underpowered test spreads too little budget across too many variants, so no single variant gathers enough conversions in the test window. The result looks like a winner but is really noise. The fix is more budget, a longer window, or fewer variants.
ElenIQ’s Loki tracks how each concept and variant performs, flags tests that are underpowered, and surfaces which creatives are genuinely winning rather than riding variance — so testing budget turns into reliable decisions.
Use Loki to track every concept and variant, flag underpowered tests, and surface the creatives that are genuinely winning.
Explore Loki