Give each creative test a question.
Key takeaways
A useful creative test starts with a decision the team wants to make. Write the question, identify the meaningful difference between versions and agree on the evidence needed to review it. Keep the offer, destination and account context visible so the result is interpreted within its limits. The outcome should produce a specific next action or a clearly documented uncertainty, rather than a blanket judgment about an entire format.
“Test new creative” describes an activity but leaves the decision open. A clearer question might ask whether a demonstration explains a product feature more effectively than a spoken description. That question helps the team decide what to make, what should stay consistent and what to examine when the work has run.
Write the decision before the versions
State what you would do differently depending on the result. If the answer would not change a creative or media decision, consider whether the test is the most useful next task. The question should connect to an actual uncertainty in the message, evidence, opening or format.
For a hypothetical storage product, the team could ask whether showing the item under a bed makes the dimensions easier to understand than listing them in text. Both versions need accurate product information. The difference should be intentional enough that the result can inform the next production decision.
| Broad task | Specific question | Decision it could inform |
|---|---|---|
| Try more video | Does a demonstration explain the fit better than text? | Whether to produce more demonstration footage |
| Improve the hook | Should the size concern appear before the product reveal? | How to open the next version of this concept |
| Refresh creative | Is there a different customer concern worth addressing? | Which distinct concept to brief next |
Explain what each version changes
Record the concept, opening, evidence, format, offer and destination. A comparison can change more than one thing, but its interpretation needs to acknowledge that. If a video uses a different promise, offer and page from a static, the result cannot isolate the effect of format alone.
Plan the review before launch
Agree on the relevant business outcome and supporting observations with the person responsible for media buying. The appropriate evaluation depends on the objective, account conditions and available evidence. This framework does not prescribe a universal spend threshold, timeline or bidding rule.
- The question: what uncertainty should this work help resolve?
- The difference: what has changed between the versions?
- The context: what offer, destination and conditions accompany them?
- The decision: what action would each plausible result support?
- Question
State the uncertainty and the decision it affects.
- Versions
Document the intended difference and shared context.
- Evidence
Review the relevant outcome within its limits.
- Next brief
Choose an action and preserve the reasoning.
Keep the brief and version notes close to the performance review. A filename rarely explains the hypothesis. The reviewer should be able to see the actual ad, the intended job and the conditions relevant to the result without rebuilding the history from messages.
Separate what happened from what it proves
Describe the observed result first, then the interpretation. An execution receiving limited delivery leaves a different uncertainty from an execution with enough relevant evidence for a practical decision. A promising result also needs context; it does not establish that every future ad using the same visual treatment will work.
Choose a next action with a reason
The next step could be a clearer demonstration, a deeper variation of the same argument or a genuinely different concept. It could also be a destination or measurement check. Write why that action follows from the evidence and what would change your view.
Keep the useful knowledge attached to the concept. When someone returns to the idea later, they should see which executions were tried, which questions remain open and which conclusions were limited to a particular context. The test becomes useful when it changes the next brief, not merely when a report contains a winner label.
Putting it into practice
- Write the question and decision before commissioning versions.
- Document the differences and the surrounding account context.
- Preserve the uncertainty as carefully as the conclusion.
Frequently asked questions
Should a creative test change only one thing?
A narrow comparison can make interpretation clearer, but practical creative work sometimes changes several things. Document the differences and avoid claiming that the result isolates a single variable when it does not.
What if an ad receives very little delivery?
Treat the available evidence as limited and review the media context with the responsible owner. Limited delivery alone does not answer every question about the underlying message or concept.
How long should a creative test run?
There is no universal duration in this framework. Set the review approach around the business objective, account conditions and evidence needed for the decision, with the person responsible for media buying.
