Design lightweight experiments to validate assumptions. Use when you have a hypothesis to test but don't yet have enough evidence to commit to building.
