Leap of Faith

Every startup begins with two leap-of-faith assumptions: a value hypothesis (people will get genuine value from the product) and a growth hypothesis (the way the product gets to more people is sustainable). Both are easy to assume and hard to verify. The Lean Startup discipline is to make both assumptions explicit, to design experiments that test them directly, and to refuse to scale before the tests have produced clear positive results.

The value hypothesis is tested with the earliest customers. Do they use the product enough to demonstrate they value it? Will they pay, refer, return? The metric is not what they say in surveys, which is unreliable. The metric is what they do, repeatedly, when given the choice.

The growth hypothesis is tested by tracing where new customers come from. If they come because existing customers told them, the growth engine is viral. If they come because the product itself generates the channel (newsletters, referrals built into the flow), the growth engine is sticky or paid. Each engine has its own metrics and its own scaling logic; conflating them produces growth strategies that look right and produce nothing.

The chapter contains the book's strongest argument against the conventional startup playbook of build-launch-scale: scaling before validation is the most reliable way to burn money. Lean teams scale only after the leap-of-faith assumptions have produced real evidence.

The two leap-of-faith assumptions deserve scrutiny because everything else rests on them: the value hypothesis asks whether the product delivers real value once people use it, and the growth hypothesis asks whether the mechanism by which it reaches new customers is sustainable. Ries borrows Randy Komisar's notion of analogs and antilogs — existing companies whose successes and failures supply evidence for or against your assumptions — as a starting point, but insists they are no substitute for direct contact with real customers. Echoing Steve Blank, he urges founders to 'get out of the building': test assumptions against living prospects through qualitative customer discovery before committing to quantitative experiments, because no amount of office analysis can substitute for watching a real person react to the actual offering. The discipline is to identify the riskiest assumption first and design the cheapest experiment that can kill it, rather than building the whole product and discovering the fatal flaw at the end. Above all, the chapter forbids scaling before both hypotheses have been validated, because pouring resources into growth on top of an unproven value proposition simply accelerates the company toward a larger, more expensive failure.