Readout process guide

How we sequence experiment readout work with product teams: frame the decision, lock metrics, prove data health, then log the call.

From hypothesis to logged decision

Each step produces something reusable—templates and checklists your team keeps after we leave.

  1. Name the decision

    Ship, iterate, or stop. Write the question the readout must answer before charts appear.

  2. Assign metric roles

    Primary for success, secondaries for learning, guardrails for safety—with pre-agreed thresholds.

  3. Prove sample health

    Show assignment balance, exposure counts, and peeking notes near the top of every view.

  4. Structure the first screen

    Hypothesis, primary lift with uncertainty, guardrail status, and a recommendation field—nothing else competes.

  5. Log the call

    Capture the decision, owner, and follow-ups in the same place the numbers live so newcomers do not need a private briefing.

Browse services Book a readout walkthrough

Notebook and laptop used while planning experiment reviews

Want this applied to your experiment program?

Start with Readout Dashboard Design or ask for a facilitated process workshop.

See dashboard design