Match the experiment to the uncertainty and decide how you will interpret the result before it arrives. Use the example as a starting point, then adapt the steps to your product, team, and evidence.
WHAT YOU’LL LEAVE WITH
An experiment brief with a hypothesis, measures, guardrails, and decision rules.
Before you begin
- A problem and the riskiest assumption
- A baseline and feasible way to recruit or expose users
- Access to measurement and relevant review partners
STEP 01
Name the uncertainty
State what you need to learn. A usability test asks whether someone can complete a task. A demand test asks whether people take a meaningful action. A controlled experiment may estimate the effect of a change. These are different claims.
STEP 02
Write a falsifiable hypothesis
Specify audience, change, expected behavior, and rationale. Example: a clearer next-step prompt may help new administrators finish setup. Define what evidence would count against the idea.
STEP 03
Choose the method
Use observation, interviews, prototypes, pilots, or an experiment as appropriate. For quantitative controlled tests, work with an analyst on design, sample requirements, exposure, and analysis. Do not pick a sample size just because it sounds convenient.
STEP 04
Define measures and guardrails
Choose a primary outcome and supporting measures. Specify event definitions, observation window, and possible harms such as more errors or support demand. Check instrumentation before relying on results.
STEP 05
Agree on interpretation
Document how you will decide continue, revise, or stop. Include what happens if the evidence is inconclusive. Avoid changing the success measure after seeing results or repeatedly stopping at a favorable-looking number.
STEP 06
Review what actually happened
Compare results with the plan. Investigate missing data, exposure problems, and unexpected effects. Record the learning and next decision. A negative or inconclusive result can still prevent an expensive mistake.
MAKE THIS PROMPT YOUR OWN
Review this experiment brief. Check whether the method addresses the uncertainty, the hypothesis is testable, measures are defined, and guardrails and interpretation rules are explicit. Identify where statistical or domain expertise is needed. Do not invent a sample size or claim causal certainty.
YOUR EDITABLE WORKSHEET
Experiment brief
Define what you need to learn before choosing a test.
Download the template ↓Check your work
- Does the method match the question?
- Are outcomes and guardrails defined before the test?
- Has measurement been checked?
- Can the team handle an inconclusive result?
Further reading: related templates & resources ↗
Editorial guide · Reviewed September 27, 2026. Tool interfaces may change.