Validation

Validation is gathering real evidence that people want what you're building, before you sink months into building it. The goal is to be wrong cheaply: test the riskiest assumption, usually 'will anyone actually pay for this', with the smallest experiment that gives an honest answer.

The strongest validation involves money or commitment, a pre-order, a paid pilot, a signed letter of intent, because talk is free and people are kind. Weaker signals, survey results, waitlist sign-ups, enthusiastic interviews, are useful but inflate easily; nobody's wallet is on the line. The hierarchy runs from cheap-and-noisy to expensive-and-true, and you want to climb it before committing serious effort.

A few ways this looks in practice:

• Say you're testing a new SaaS idea and want a real demand signal: stand up a one-page offer in Unbounce describing the product as if it exists, run a little traffic at it, and count how many people click 'Get early access' or actually enter a card. A dead page is a cheap no.
• Say a handful of prospects sound keen: instead of trusting the enthusiasm, put a real paid-pilot call on the calendar with Cal.com. Someone who books a slot and pays a small fee is worth more than ten people who said 'love it'.
• Say you've shipped a rough first version to early users: watch what they really do with session recordings and heatmaps in Hotjar, so you see where they drop off rather than relying on what they tell you in interviews.

For a lean founder this is risk management. Most startup deaths trace back to building something nobody wanted, the single most expensive mistake there is. Validation doesn't guarantee success, but it kills bad ideas early, while they're cheap to kill. Design each test so a clear no is possible, otherwise you're not validating, you're collecting reassurance.