A creative test should investigate why someone responds. Ten cosmetic variations of the same weak idea may produce a ranking without teaching the team much about the customer.
Start with different hypotheses about the buying reason. Then develop enough execution quality to give each idea a fair chance.
Separate concept from treatment
The concept is the argument: the problem, desired outcome, or reason to choose the product. The treatment is how that argument is presented through words, images, demonstration, or format.
For a fictional operations product, one concept could focus on time lost to manual coordination and another on missed work caused by unclear ownership. Changing only the background color does not test that distinction.
Write a brief for each concept
Record the audience, buying situation, promise, evidence, objection, and next action. Check that the product can support the promise. Use customer language where you have evidence for it rather than inventing quotes.
Keep a small number of concepts in the first round. Production and interpretation capacity should determine volume.
Use a two-stage learning process
First investigate which arguments produce relevant attention and progression. Then refine the execution of promising concepts. Avoid declaring a concept dead because one rushed treatment failed.
| Layer | Example question |
|---|---|
| Audience | Are the intended buyers seeing and responding? |
| Argument | Which problem or outcome creates relevance? |
| Evidence | What makes the claim believable? |
| Execution | Is the message clear in the available format? |
| Journey | Does interest lead to an appropriate next step? |
Evaluate beyond engagement
A provocative idea can attract attention while creating poor-fit enquiries. Keep qualified outcomes and customer progress in the review. Where volume is small, use qualitative evidence to explain uncertainty rather than manufacturing a precise winner.
Preserve what was learned
Record the concept, treatment, audience, result, and interpretation. Link to the actual asset so future reviews are not based on memory. Separate observed performance from the team’s explanation of why it happened.
The experiment brief provides the decision structure. A creative library should accumulate usable knowledge about customers, not simply become an archive of files.
Co-founder and CEO of Stackmatix, startup advisor, and former Head of Sales at MightyHive. Connect with Matt on LinkedIn ↗ · More about Matt →