Liva 7 Help / Current product

Run a test

Direct guidance for the product Liva 7 is now, with the control and evidence boundary kept visible.

A useful test begins with a decision question. Liva 7 then defines who can receive each version, what result matters, how long learning may take and when safety should stop the test.

Use this guide when

A decision needs stronger evidence than observation or association can provide.

Follow this route

  1. Name the question

    Choose one idea about recommendations, a bundle, content, placement or eligible offer.

  2. Define the plan

    Set the shopper group, versions, primary result, guardrails, minimum information, duration and stop rules.

  3. Preview delivery

    Confirm exactly what each group can receive and that assignment is distinct from genuine display.

  4. Choose authority

    Start manually, ask first or use a suitable existing bounded policy.

  5. Wait for the decision state

    Review collecting, paused, insufficient, clear, no-clear-difference or guardrail-stopped states without teasing a premature winner.

Before you finish

  • One primary result is named
  • Delivery denominators are available
  • The stronger version can be applied or restored safely

Continue with the right context

ProductExplore Tests and learning InsightsRead Measurement and learning Related guideUnderstand attribution Related guideUse the learning loop