Give your test
a clear decision.
Turn a customer problem into a brief your team can review before building. Capture the evidence, assumptions and conditions that would make you ship, revise or stop.
Start with what you know.
You can leave a section blank while you investigate. Keep unknowns visible so your team knows which decisions still need evidence. Copy or download your notes before leaving; this page does not save them.
Loading the example replaces your current notes. Copy or download them first if you want to keep them.
Fictional example loaded. Replace the assumptions with evidence from your own work.
Who is trying to do what, and where do they get stuck?
Record observations and source links. What is still unknown? An Evidence Pack from the A/B Test Library is a useful starting point.
Describe the treatment, comparison and why you expect behavior to change. Include a competing explanation.
Name one outcome, its denominator and measurement window. What is the smallest change worth acting on?
Paste the GrowthLayer planner summary: baseline, effect size, eligible traffic, allocation, sample target, duration and stopping rule. Record the randomization unit and analysis method.
What must not get worse? Define thresholds, instrumentation checks, traffic-allocation checks and rollback conditions.
What evidence would lead you to ship, revise, rerun or stop? Include an inconclusive result and implementation cost.
Who owns implementation, measurement and the decision? What must be resolved before launch?
Review and take it to your team
Blank sections are marked “Unanswered” in the draft. A filled section may still need evidence or a team decision.
Read the text before you share it
Based on the Behavioral Audit framework and the evidence briefs in the A/B Test Library.