An experiment is a structured test of a specific hypothesis with predefined success criteria. In the context of product development, this means: before an important assumption becomes the basis for an investment decision, it is made testable. The guiding question is: What is the fastest and cheapest way to learn whether our assumption is correct?
The range of possible experiment formats is broad. An A/B test compares two variants quantitatively. A smoke test measures interest via a landing page before the product exists. A Concierge MVP tests the service manually, a Wizard of Oz MVP simulates automation behind the scenes. What matters is not the format but the structure: every experiment needs a clear hypothesis, a measurable metric, and a threshold above which the hypothesis counts as confirmed or refuted. Without establishing these criteria beforehand, the risk of reinterpreting results after the fact is high.
The concept connects the scientific method with the Lean Startup methodology. Experiments are especially valuable under high uncertainty — precisely where traditional planning reaches its limits.
Apply methods, don't just know them. View trainings and workshops
