Testing Plans: Define the decision before the result arrives
A test launches before the team agrees on the hypothesis, comparison, primary measure, guardrails, sample rules, or decision threshold.
The useful response is not another status layer. It is a mechanism that can define the decision before the result arrives.
What the skill produces
The Testing Plans skill turns supplied context into a precommitted test plan with hypothesis, assignment, measures, limitations, and decision rules. It should expose missing evidence instead of inventing it. This article explains the operating model. The skill is part of the premium library, so the brief does not reproduce private package contents or a public download path.
Run the mechanism
- State the decision and falsifiable hypothesis.
- Define baseline, comparison, population, duration, primary result, and guardrails.
- Record analysis and keep-change-stop rules before launch.
The accountable owner reviews the output against the source evidence before it is used. Material unknowns remain labeled and receive an owner and date when they affect the decision.
Failure check
A test becomes post-hoc storytelling when the team selects the most favorable measure after seeing the data.
Adoption test
Audit the next three tests for plan changes after launch. Record whether each conclusion stayed inside the design and evidence.
Keep the skill only if the result improves against that baseline. Revise the playbook when the same operating failure survives more than one review cycle.
Infographic
