Plan an advertising test around one question

Name the difference you want to understand

Describe two proposed versions and circle the deliberate difference. Perhaps the headline explains the use case in one version and the product feature in the other. Keep the remaining offer and destination consistent where the chosen method allows. List any unavoidable differences as limitations. This is a planning exercise; check the current platform’s supported testing setup before assuming that two independently launched adverts form a controlled experiment.

Choose an outcome the business can explain

Decide what evidence would answer the question and how it connects to the business goal. A readable headline, a qualified enquiry and a completed sale are different outcomes. Record which one the test can actually observe. Agree a spending boundary and review point with the person authorised to launch. Do not keep increasing the budget because an early result looks encouraging or because an inconclusive result feels disappointing.

Protect the record from hindsight

Save the original versions and the plan before launch. During a later review, note changes in availability, destination or measurement that could affect interpretation. If the comparison cannot answer the question, record that outcome rather than inventing a winner. A small test may suggest another question without establishing a universal rule. Business decisions still need truthful offers, suitable permissions and a realistic view of what the evidence can support.

A practical check before you finish

Ask a colleague to restate the test question from the plan. If they describe several different questions, narrow it before launch.

• Identify the intended difference between the versions.
• Connect the chosen outcome to an observable event.
• Agree the spending boundary before collecting results.
• Keep unexpected changes with the original plan.