The Next-Gen Brand Lab
Meta, Google and channel choice
How many ad creatives should a launch test?
Last updated 2026-09-09
Think in concepts rather than files. Four to six distinct concepts, each with a few variations, is enough to find a winner without splitting the budget so thin that no cell reaches statistical significance. Fifteen static creatives a month is the volume we include in the fee, which supports that structure with room to refresh on fatigue. Twenty variations of one concept is not a test — it is one bet with extra steps.
What counts as a distinct concept
A different argument, not a different crop. Price-led, location-led, amenity-led, urgency-led and proof-led are five concepts. The same price message in five colourways is one concept and will win or lose as one.
The purpose of a test is to find out which argument the market responds to. Only then does it make sense to generate variations of the winner.
When to refresh rather than restructure
When frequency rises and cost per result rises with it, the creative is spent. Refresh it. When cost per result rises while frequency stays flat, something changed in the auction or the audience, and new creative will not fix it. Distinguishing these two cases correctly is most of the skill in running the account, and getting it backwards wastes both the budget and the month.
Want this answered for your project specifically?
A Phase 0 call is a questionnaire, not a pitch. You get the market read before anyone spends a rupee on advertising.
Book a Phase 0 call ↗