Most creative tests are called long before the data can support the decision. Put in your account’s real numbers and this works out what it would actually cost to know — and what your current budget can and can’t detect.
Pull these from the last 30 days of the ad set you'll be testing in.
What you're planning to run, and for how long.
The defaults are the academic standard. Loosen them and you'll need less spend, but you'll be wrong more often.
Testing on clicks instead of purchases costs £1,988 and takes about 13 days — a fraction of the full test.
Use it to kill obviously weak creative fast. Don’t use it to crown a winner: a hook that earns clicks and a hook that earns purchases are not the same hook, and plenty of high-CTR creative converts worse than what it replaced.
Reaching significance would take about 657 days — more than three times the window you've allowed. Either cut the number of creatives, raise the budget, or accept that you're testing for a much bigger difference than 20%.
The spreadsheet version, plus the note on reading tests that finish short of significance without fooling yourself.
Also worth a look: the ROAS you actually need to see