Launching a clean creative test: the 5-point checklist

Launching a clean creative test: the 5-point checklist

So your test measures the creative, not the noise around it.
— 1: one variable per test. Change the hook OR the format, never both.
— 2: identical targeting + placement across variants. Advantage+ placements ON for all or OFF for all.
— 3: budget = 1x target CPA per creative per day, minimum.
— 4: minimum 4 creatives, maximum 8 — fewer and you can't separate winners, more and budget spreads thin.
— 5: run 4 full days, weekday-only if weekends skew your vertical.

Clean-test sample basis, 40 accounts: top buyers reach 95% statistical confidence in 4 days / median 7 / laggards never (too many variables).

Read: a dirty test is more expensive than no test — it teaches you the wrong lesson with conviction.