You should already know the original objective, what happened and where the path appears weakest.
Campaign Test & Experiment Planner | Run One Controlled Marketing Test
Find Your Brand · Free Apply + Tool · Controlled optimization
Change One Thing So You Can Learn Something.
A campaign test is useful only when the change is clear enough to connect new evidence back to a decision. This planner takes a TEST or ADAPT decision from Campaign Review and turns it into one controlled marketing experiment.
Use it after completing Campaign Review & Optimization. If you do not yet have a diagnosis, go back and review the evidence first.
Prerequisites
Your Campaign Review should end with TEST or ADAPT. KEEP, STOP and WAIT do not automatically require an experiment.
You need a reason for the proposed change. “Maybe this will work better” is not enough direction.
Core rule
HYPOTHESIS → ONE PRIMARY VARIABLE → CONSTANTS → EVIDENCE → DECISION. If you cannot explain what changed and what stayed the same, you are running another campaign—not a useful controlled test.
The test-planning sequence
Restate the weak point you believe the evidence exposed.
State what you believe will improve and why.
Choose the one primary thing you intend to change.
List the important elements that should remain unchanged.
Define when the test begins, ends or reaches a meaningful decision point.
Choose the primary evidence tied to the campaign objective.
Define conditions that require stopping, holding or escalating.
Define what result would justify KEEP, ADAPT, TEST AGAIN, STOP or WAIT.
Choose one primary variable
Change who receives or sees the campaign while keeping the message, offer and destination as stable as practical.
Change the framing, headline, hook, visual or creative treatment while keeping the audience, offer and destination stable.
Change the value proposition, incentive, package or requested action while preserving the rest of the path where possible.
Change how people move from attention to the next step without simultaneously redesigning the whole campaign.
Change the landing page, product page, form or checkout element when evidence points to friction there.
Change channel, placement, frequency or timing when the problem appears to be reach or delivery rather than the offer itself.
What counts as a constant?
A constant is any meaningful campaign condition you deliberately avoid changing so the result remains interpretable. You may not be able to hold everything perfectly constant in real marketing. Record the differences you cannot control instead of pretending they do not exist.
Write a useful hypothesis
Use this structure:
Because we observed: ________
We believe the likely problem is: ________
Therefore we will change: ________
We expect this to improve: ________
We will judge that primarily by: ________
A useful hypothesis connects evidence to a decision. It does not need to sound scientific. It does need to be specific enough that the result can teach you something.
Complete the Campaign Test Record
Campaign / test name: ________
Original campaign objective: ________
Campaign Review decision — TEST / ADAPT: ________
Observed weak point: ________
Diagnosis: ________
Hypothesis: ________
Primary variable changing: ________
Exact change being made: ________
Important constants: ________
Known conditions we cannot fully control: ________
Test start: ________
Test end / decision point: ________
Primary objective evidence: ________
Supporting behavior evidence: ________
Attention evidence worth noting: ________
Minimum evidence needed before deciding: ________
Stop / hold conditions: ________
Decision rule: ________
Owner / approver: ________
Before you launch the test
Can you describe the main change in one sentence? If not, the test is probably too broad.
Have you accidentally changed the audience, offer, creative, destination and distribution at the same time?
Is the primary evidence connected to the original business objective rather than just the easiest metric to access?
Is the evidence window long enough to be meaningful for the campaign and action being measured?
Could the change create brand, legal, privacy, permission, customer-experience or client-approval risk?
Do you know what evidence would make you keep, adapt, test again, stop or wait before seeing the result?
After the test
Return to Campaign Review & Optimization. Compare the new evidence to the original objective and to the hypothesis. Do not declare a winner simply because one visible metric improved.
Completion and success criteria
You are finished when one planned test is specific enough that another competent person could run it without inventing the missing logic.
- The test begins with a reviewed campaign and evidence-based diagnosis.
- The hypothesis is explicit.
- One primary variable is identified.
- Important constants are documented.
- Uncontrolled differences are acknowledged.
- The evidence window and minimum evidence requirement are defined.
- Stop or hold conditions are explicit.
- A decision rule exists before the result is known.
- The test has an accountable owner or approver.
When the problem becomes bigger than one test
This free planner is enough to design one controlled experiment. When you are managing multiple campaigns, a backlog of possible tests, several owners or clients, and learning that needs to carry forward, move into the paid Campaign Optimization Operating System.
That Build layer adds a Campaign Portfolio Register, Experiment Backlog, Active Test Queue, Decision & Learning Log, cross-campaign pattern review and cycle-based optimization summary. It extends this free workflow rather than replacing it.
Primary road: Find Your Brand · Secondary: Find Your Voice · Free Apply + Tool. Controlled testing can improve the quality of decisions; it does not guarantee causality, performance or a winning campaign.
