Insight · CreativeTOFU
A creative testing framework that actually teaches you something
The short answer
Most creative testing produces results without producing learning. Brands launch a dozen variations, keep whichever has the lowest cost per purchase, and start again next month with no clearer idea of why it won. A framework fixes that by testing ideas, not assets.
Concepts before variations
- A concept is an argument for buying — price, problem, proof, comparison, story.
- A variation is the same argument with a different colour, crop or wording.
- Test concepts first. Variations only matter once you know which argument wins.
The framework
- Hypothesis — "Customers care more about results than ingredients."
- Concepts — two or three ads built to test the hypothesis.
- Hooks — several openings for each concept.
- Launch in a dedicated testing campaign with fixed budget.
- Read early signals — hook rate and hold rate within days.
- Read outcomes — cost per purchase once there's enough data.
- Record the learning — what was true, what wasn't.
- Brief the next round based on the learning.
What to measure, in order
- Hook rate — does the opening stop people?
- Hold rate — do they keep watching?
- Click-through rate — does it create intent?
- Conversion rate — does the landing experience deliver?
- Cost per result — the final judge
The takeaway
Keep a creative learning log. After a few months, it becomes the most valuable document in your marketing team.
An example hypothesis log
- Hypothesis — "Cold audiences respond more to price clarity than lifestyle imagery."
- Test — two concepts: price-led hook vs lifestyle hook, same product and offer.
- Result — price-led concept had a higher hook rate and lower cost per purchase.
- Learning — this audience is value-driven; lead with price in prospecting.
- Next test — price-led hook with three different proof points.
Common testing mistakes
- Changing several variables at once
- Declaring winners after a day
- Testing only colours and crops
- No written record of what was learned
A simple creative testing calendar
| Week | Activity |
|---|---|
| 1 | Insight review and concept list |
| 2 | Scripts, shoots and edits |
| 3 | Launch 3–6 new concepts in a test structure |
| 4 | Read results, iterate winners, log learnings |
What to do next
- Start a hypothesis log
- Brief three to six distinct concepts
- Set spend and time limits per test
- Review results weekly
Frequently asked questions
How much budget does a creative test need?
Enough for each concept to get a meaningful number of conversions or a clear read on leading indicators — usually several days of spend at your target CPA per concept.
Should we use Meta's A/B test tool?
It's useful for clean comparisons of important variables. For routine creative testing, a consistent structure with separate ad sets or dynamic creative is often enough.
How many creatives should be live at once?
Three to six per ad set is a common range — enough to test without starving each ad of delivery.
How many creatives should we test at once?
Enough distinct concepts to learn — often three to six per test — without spreading budget too thinly.
How long should a creative test run?
Until results are meaningful, usually one to two weeks depending on spend.
Keep reading
Services behind this
Terms used on this page
Talk to a strategist
Thirty minutes. Your numbers. A straight answer.
Book a strategy call with the people who'd actually run your account. We'll tell you what we'd do — or that you don't need us yet.