How many ad creatives should you test? A practical framework

There is no magic number of creatives. The right number is the smallest one that can teach you something with the budget you have. Too few and you learn nothing; too many and each ad gets so little spend that the platform decides at random. The framework below keeps the count small and the questions sharp.

Test in layers

1. Angle: the reason to buy, such as problem–solution, unboxing, objection handling or a before and after. Different angles reach different buyers.
2. Hook: the first one to three seconds of the winning angle. The same angle with a better hook can change results sharply.
3. Presenter and voice: who says it. Worth testing once you have a winning script.
4. Format details: caption style, length, aspect ratio and end card. Small gains, test last.

How many in each round

A first round of three to five angles, each with two hooks, gives six to ten ads: enough to see a pattern, few enough to fund properly. Divide your test budget by the number of ads and check that each one can reach a meaningful number of results; if it cannot, cut angles rather than spreading thinner. In later rounds, keep one or two winning angles and test four to six new hooks on them.

How to judge a test

• Hook rate: the share of viewers still watching after the first seconds. It tells you whether the opening works.
• Hold rate: how many watch most of the ad. It tells you whether the script keeps its promise.
• Cost per result: purchases, installs or leads. This is the number that decides.
• Ignore likes and comments as success signals; they reward entertainment, not sales.

Refreshing before fatigue

Even a winning ad wears out as the same audience sees it again. When its cost per result starts to climb steadily, keep the angle and write new hooks, new presenters or a new length rather than a brand-new concept. Producing those variations quickly is what keeps a winning angle alive.

Doing it with AdFace

AdFace is built for this kind of testing: one product brief produces scripts across several angles, you edit hooks before rendering, and each script and presenter combination becomes its own ad. Review the batch estimate before starting: ten 15-second variations do not have a universal credit price. Ad style, model and resolution determine the cost. Choose one campaign format: 9:16, 1:1 or 4:5.