A simple framework for creative testing

Start with one hypothesis

A useful test changes one meaningful thing, such as the hook, the offer framing or the visual style. Write the hypothesis down before launch: we believe showing the product in use will lift click through rate compared with a studio shot.

Set up a clean test

Keep audience, budget and placement the same across variants. Decide the success metric and the minimum sample before you look at results.

• One variable per test
• Equal budget split
• A pre agreed metric and duration
• No mid test edits

Read results honestly

Small differences on small samples are usually noise. Look for a clear and consistent gap, and record the learning even when the test is inconclusive. A log of learnings is often worth more than any single winner.