You don't need hundreds of users to know if an idea works. You need a small group who stay in contact after the first conversation, keep using the thing, and tell you when it's wrong. A founder who recently turned down a $15,000 acquisition offer for his solo-built tool had 20 active users and just 10 people in what he called a "close feedback loop." That second number, not the first, is what actually validated the product. Signups measure curiosity. A close feedback loop measures whether anyone cares enough to stay.
Twenty signups and ten people who still answer your emails are two different companies.
Why does user count feel like the wrong metric to trust?
Because it is. A signup costs the user nothing. Clicking a button, typing an email, trying a free tier for four minutes, none of it requires belief. It requires curiosity, which is a much cheaper thing to fake, including to yourself. A founder who recently walked away from a $15,000 offer for his solo-built developer tool had around 20 active users. He didn't lead with that number when explaining why the offer was too low. He led with a smaller one: about 10 users in what he called a close feedback loop, the people still filing bugs, still replying, still around months in.
What is a close feedback loop, and why does it matter more than signups?
A close feedback loop is the subset of your users who keep showing up without being asked. Not the ones who said the idea sounded great. Not the ones who signed up and never opened the app again. The ones who complain, request a fix, come back a second time, and eventually pay without a discount code attached. That group is almost always smaller than your total user count, often by a wide margin. It is also the only group whose behavior tells you anything, because their attention has a cost and they keep paying it.
Can a stalled launch or a quiet Kickstarter tell you the idea was bad?
Not by itself. One founder posted publicly this week asking whether his stalled Kickstarter meant his idea was bad, or whether the problem was distribution and who he'd actually reached. Those are two different failures wearing the same outcome. An idea nobody wants and an idea nobody saw look identical from the founder's chair. The only way to tell them apart is to get the idea in front of people who are not already in your network, and watch what a small number of them actually do, not what a large number of them say when you ask nicely.
What does it mean when users describe your product in words you never used?
It means you got a free spec update. A founder building a news-story visualization tool discovered that almost nobody who tried it talked about the visualization. They kept saying "catch me up" and "help me understand what changed." He'd built a way to see a story evolve. They experienced it as a way to stop missing the story. That gap between the builder's language and the user's language is not noise. It is usually the most accurate description of the product you have, because nobody coaches a stranger on what to call the thing that actually helped them.
How do you tell a polite yes from a real signal?
You stop asking people what they think and start watching what a small number of them do when nobody is prompting them. A friend telling you it's a great idea costs them nothing and protects the friendship. A stranger who keeps coming back, keeps complaining when something breaks, and eventually pays, is spending something real: attention, trust, sometimes money. One is free. The other has a price. Count the ones who paid a price.
Key takeaways
- Total signups measure curiosity, not conviction. A close feedback loop of returning, complaining, paying users is the number that actually validates an idea.
- A stalled launch can mean the idea was bad or that distribution failed to reach the right people. Those are different problems with different fixes.
- When users describe your product in language you never used, that language is usually a more accurate spec than the one you started with.
- Friends and casual signups give free, low-cost opinions. A real signal costs the person something: attention, time, or money.
- You do not need a large sample to know if you're right. You need a small, honest one.
FAQ
How many real users do I need before I trust the signal?
There's no fixed number, but the pattern that shows up again and again is small: somewhere around 10 to 20 people who stay engaged past the first interaction is usually enough to tell you whether you're onto something, long before you have hundreds of signups.
What's the difference between a signup and a validated user?
A signup clicked a button. A validated user came back without being asked, told you something specific was wrong or missing, and eventually did something that cost them, usually paying.
If my launch got no traction, does that mean the idea is bad?
Not necessarily. It might mean the idea reached the wrong 200 people instead of the right 20. Distribution failure and demand failure look the same from the outside and require different diagnoses.
Should I trust it when users say they love the idea?
Trust it less than what they do next. Saying they love it costs nothing. Coming back, using it unprompted, or paying costs something. Weight your evidence accordingly.
What if users describe my product completely differently than I intended?
Pay close attention. That's usually not confusion, it's the market handing you a more accurate description of the value than the one you wrote yourself.