Experiment Tracker

What this is for

Learn whether an assumption is true before spending a full delivery cycle on it.

When this helps

• Teams call launches experiments without defining what would count as learning.
• Results are scattered across analytics, docs, and conversations.
• A failed test becomes an argument instead of reusable evidence.

How it works

Each experiment has a hypothesis, method, audience, success measure, status, and evidence. The tracker preserves the plan and the result so future decisions can use what was learned.

Before you start

• A falsifiable hypothesis.
• A method such as prototype test, interview, survey, or rollout.
• A metric or decision rule and the linked assumption.

What you should see

• A planned experiment with a clear owner and status.
• Recorded evidence and result.
• A validated, invalidated, or unresolved assumption linked to strategy.

Try it now

1. Open Experiments and create an experiment before building the full feature.
2. Write the hypothesis as “If we do X, Y will change because Z.”
3. Choose the smallest method that can change your decision.
4. Record the result, link evidence, and mark the assumption as supported, rejected, or still unknown.

A real example

If a shorter invite form increases completion by at least 10%, the team will ship it. A five-user prototype test reveals confusion before engineering effort is committed, so the form is revised first.

Your next move

Use Outcome Intelligence after launch, or follow the outcome tracking recipe.