The Lean Startup, described by Eric Ries in 2011, formalised a loop for building under uncertainty: build the smallest version of an idea that can produce real evidence, measure the specific outcome you said you would measure, and let that result, not momentum or sunk cost, decide whether you continue, pivot, or stop. It borrowed the same instinct behind the Toyota Production System’s small batches, that the most expensive mistake in any project is the one nobody notices until it is too large and too expensive to reverse.
Applied outside software, the discipline still depends on deciding the stop-or-go metric and the review date before building anything. Skip that step and build, measure, learn becomes build, then explain, where the metric that matters is invented retroactively to fit whatever result already occurred.
The method’s real output isn’t a shipped product. It’s a validated answer to whether the idea was worth building at all, which is exactly as useful when that answer is no.