Identify the uncertainty
Before planning a minimum viable product, explain what you do not yet know. Will customers use the service? Can the team deliver it reliably? Will users pay? Different uncertainties require different tests.
Choose the minimum evidence
A clickable prototype can test comprehension but cannot establish repeat use. A manually delivered service can reveal demand while leaving technical scalability untested. Match the test to the question rather than assuming every experiment needs production code.
Define success before you start
Write down the audience, intended behaviour and decision you will make from the result. Decide what would count as a promising signal and what would cause you to change direction. A test without a decision can become an endless demonstration.
Protect the essential experience
Small does not mean unreliable. If the experiment asks customers to trust a service, basic accuracy, clarity and support may be necessary. Reduce scope while preserving the experience needed to learn something useful.
Choose a focused experiment that tests your most important assumption.
