Write stop and review rules before a marketing test starts

Start with test validity

List what must remain true for the test to mean anything: the intended audience can reach the offer, the service is available and the contact route works. Identify the failures that would make further acquisition unhelpful.

Write the observable evidence for each failure. A received test notification can confirm routing; a low click count does not prove the form is broken. Keep diagnosis tied to the actual condition.

Define immediate action rules

Choose actions for invalid conditions: pause promotion, fix the destination, correct the offer or restore response coverage. Assign a named role and a place to record the decision. If someone else runs the account, agree on how urgent changes reach them.

Avoid a stop rule based on emotion alone. A disappointing day is different from a broken customer journey, an exhausted capacity limit or a cost setting outside the approved test boundary.

Set a normal review boundary

Specify the planned duration and maximum total effort or spend. Write the evidence you will inspect, including suitable requests, known decisions and recurring questions. Check the platform's current budget behavior when relevant.

An average daily advertising budget may produce uneven daily costs. Keep account controls and reporting understood rather than treating a handwritten daily target as a guaranteed spending limit.

Use a short daily validity check

Review the few conditions capable of invalidating the test. Check outstanding requests, current availability and any routing alerts. Leave ordinary optimization decisions for the planned review unless new evidence requires a repair.

A routine should be small enough to perform. A long checklist skipped every day gives less protection than a few specific checks with an owner and a recorded outcome.

Document interruptions and changes

When a stop occurs, record the time, affected route, action and return condition. Separate the interrupted period from valid observations. If the offer or audience changes substantially, label it as a new test condition.

Preserve the reason for a restart. Otherwise later reports may combine broken and working periods and suggest that customers rejected an offer they could not use.

End with a defensible next decision

At the normal review, summarize the observed evidence and the remaining uncertainty. Choose to continue, repair, retest or stop. Explain why the evidence supports that action and what new information a further test would add.

Do not escalate spending solely because the first attempt did not reach an arbitrary sales target. Make the next commitment concrete and bounded.

Illustrative rules for a leaflet test

A hypothetical service business distributes a small leaflet batch pointing to one offer page. Immediate stops are an incorrect service price, failed form delivery or no response coverage. The normal review occurs after two weeks with a fixed printing and handling budget.

A test request fails during the first day. Distribution pauses, the route is repaired and the owner records the interruption before resuming. The failure is treated as a routing problem, not evidence that the leaflet audience lacked interest. The example illustrates process rules rather than expected campaign performance.

Questions about this task