What counts as real validation for a startup idea?

The short answer

Real validation is a stranger taking a costly action: paying, preordering, signing up with a real email, booking a call, or returning to use the thing twice. Compliments, likes, survey answers, and encouragement from friends are not validation, because they cost the other person nothing.

Founders systematically grade their own validation too generously, because the alternative is admitting the idea has no evidence behind it yet. A simple rule fixes the grading: evidence is only as strong as what it cost the person who gave it.

The hierarchy of evidence

At the top: money. A payment or preorder from a stranger is the strongest signal that exists at this stage. Below that: identity and time, such as a real email address, a booked call, or a completed onboarding. Below that: attention, like waitlist signups from people you do not know. And at the bottom, worth almost nothing: opinions. Survey answers, social media likes, and any sentence from a friend containing the words 'great idea.'

Why friends and surveys mislead

People are polite, and hypothetical questions get hypothetical answers. 'Would you pay for this?' costs nothing to answer with a yes, which is why validation built on surveys collapses on launch day. The test that matters forces a real trade-off today: a price on a page, a signup form, a calendar link. This is also the difference between tools that score your idea and systems that test it, an argument laid out in GRILLR versus Validator AI.

Practically: put a landing page live, drive a modest amount of real traffic to it, and count costly actions from strangers. The page half of that is covered in how to smoke test a startup idea, the conversation half in customer interview questions for startup validation, and GRILLR will grade whether you actually did it rather than whether people were nice to you.