How Many Ad Creatives Should You Test?

"Test more creative" is the advice everyone gives. It's true, but incomplete. Fifty ads with a small budget won't tell you much, and five near-identical ads won't either. The useful question is: what do you want to learn, and how much creative do you need to learn it?

Concepts vs variations

It helps to separate two levels:

• Concepts are different ideas: a different angle, story, format or audience pain point. "Saves time" vs "saves money" vs "founder story".
• Variations are changes within a concept: different hooks, lengths, captions, music or CTA.

New concepts find big wins. Variations optimise a winner. Most accounts need both, in different proportions at different times.

Match creative volume to budget

Each ad needs enough spend and impressions to show a meaningful signal. If you split a small budget across many ads, each one gets too little delivery to judge. A practical rule: only launch as many new ads as your budget can give a fair chance in your testing window. Your media buyer can estimate this from your typical cost per result.

Smaller budgets

Test fewer, more different concepts. You want big swings, not subtle tweaks.

Larger budgets

Test more concepts and more variations, and refresh more often. You'll also hit creative fatigue faster, so you need a steady pipeline.

What to vary first

1. The angle or concept. The biggest lever.
2. The hook. Often the biggest lever within a concept.
3. The format. UGC vs founder vs demo vs text-led.
4. Length. Short and punchy vs longer and story-driven.
5. The CTA or offer. Changes conversion more than attention.

Smaller details like music or caption style matter less, and are better tested once you have a winner.

A simple testing rhythm

Here's a structure many teams use, adjusted to their budget:

1. Each cycle, launch a few new concepts, each with a few hooks.
2. After the testing window, pause clear losers and note what the winners have in common.
3. Iterate on winners with new hooks, lengths and formats.
4. Brief the next cycle using what you learned.

Weekly or fortnightly cycles are common. The important part is consistency, so you're always learning.

How editing makes testing efficient

A well-planned edit produces many variations from one shoot:

• Shoot several hook lines and openings on purpose.
• Build a modular body that works with any hook.
• Record two or three CTAs.
• Export each combination in the ratios you need.

That turns one shoot into a testing plan instead of a single ad.

Naming and tracking

Test results are only useful if you can read them. Use a naming convention that encodes the concept, hook, length and ratio, like "SaveTime_HookB_Short_916". Keep a simple sheet listing each ad and what it tested. When you look back in three months, you'll know exactly what worked.

Creative fatigue

Winners don't last forever. As frequency rises, performance usually drops. Signs include rising costs and falling click-through on an ad that used to perform. Refresh by swapping hooks, changing the opening shot, or bringing in a new creator with the same script. Plan for refreshes as part of your monthly creative volume.

Common mistakes

• Testing many tiny variations of one idea and calling it a test.
• Judging ads too early, before they've had enough delivery.
• Changing several things at once, so you can't tell what made the difference.
• Not documenting results, so the same mistakes repeat.
• Stopping testing once something works, then scrambling when it fatigues.

Where to start

If you're new to structured testing, start small: a handful of genuinely different concepts, a couple of hooks each, in the ratios you need. Run them, learn, repeat. Increase volume as your budget and confidence grow. Your editor and media buyer should be talking regularly, ideally sharing the same results sheet.

Briefing a testing batch

A good testing brief is short and specific. For each concept: the angle in one sentence, the audience, the key benefit, the hooks to try, lengths and ratios, and what you want to learn. "We want to know whether time-saving or cost-saving gets more clicks from small agency owners" is a far better brief than "make some ads". Your editor can then build variations that actually answer the question.

Reading results together

Hold a short review after each cycle with whoever runs the ads and whoever edits them. Look at attention metrics like hook rate and hold, as well as conversions. A hook that stops people but doesn't convert tells you something different from a hook nobody stops for. Write down the lesson in one line, and start the next brief from it.

The short version

Test genuinely different ideas, give each enough budget to show a signal, iterate on what wins, and write down what you learn. The right number of creatives is whatever lets you keep that loop running every week without wasting spend on ads nobody can judge.

Related guides

Part of our ad creative editing series. See also editing for media buyers, Meta ad editing and video ad editing costs. Browse our work or start an enquiry.