How do you know when an idea is actually validated?

Question

We've run 20+ interviews and people say nice things, but I can't tell if that is a real sign or just politeness.

No clear pattern, no obvious "yes this is a must have".   Starting to wonder if I'm asking the wrong questions or just hearing what I want to hear.

For those who've validated (or killed) an idea this way.  How did you know which side you were on?

CEO, New York, NY · Answer

I would probably read through them multiple times and try to find a pattern.  AI can be great for this.  Then I would also research about how to structure these questions.

Question types vary a lot by stage of your company.   But quantity is good here.

As to how to get them: first ppl in your network then cold dm everyone in the industry.

Founder & CEO, Long Branch, NJ · Answer

I think people don't know what they want until they see it, and don't know if they would use something until they have it. I swear every time I get home from the hardware store, I only end up actually implementing one in two home improvement that seemed like great ideas in the store. I think I would use it =/= I would actually use it.

The best way to validate an idea is to build fast and ship. Building a software prototype is easier than ever, but it's still not required. You can draw your idea out on a sheet of paper and see how the interviewee interacts with it.

Henry Waltz · Answer

20 interviews and no conclusion usually means the questions were too soft.  "Would you try this?" gets polite yeses.

"What did you try last time this came up, and what did it cost you?"  gets the truth.   Re-run a few with sharper questions.

Joe Berrizbeitia · Answer

Look for behavior, not enthusiasm.  Did anyone ask for early access, offer to pay, or intro you to someone?   People are nice in interviews, action is the only real signal.  If nobody leaned in, that's your answer.

Andres Gomez · Answer

Stop counting interviews, start counting patterns.  Read all 20 and tag the recurring pain in their words.  If the same problem shows up 12 times, that's your wedge.  If it's 20 different problems, the idea is too broad.

Juan Francisco Verhook · Answer

The trap: you're validating the idea instead of the problem.   Forget your solution for a sec.

Do thee people clearly have a painful, expensive problem?  If yes, build. 
If you can't even prove the pain, the idea doesn't matter yet.

Founder & CMO, Coral Gables, FL · Answer

People said nice things about ours too.  Real validation: they ask you for updates, new features.  They try to refer customers.  They ask when it ships.  If you have to chase them after the interview, it's politeness.  If they're chasing you, it's real.

Founder & CEO, New York, NY · Answer

Stop interviews, start pre-sales.  Tell 20 people "I'm building this, it's $500/year, want in?  If five 5 say yes and 3 pay, it's validated.  If zero pay and they all say "Let me know when it's live", you need to do some more work.

Student & Aspiring Founder, Washington, DC · Answer

Validation is when someone's willing to lose something for your solution.  Time, money, credibility, effort, money.  If they're not losing anything to say yes, they haven't helped you validate anything.  They're just being polite.

Founder & GP, New York, NY · Answer

The tell: they solve it themselves while you're talking.  They interrupt you mid pitch with "wait can you do X?"  That means they're imagining using it.  No interruptions = no mental model of it working.

Angel & Advisor, Coral Gables, FL · Answer

Listen for specificity.  Real validation: I'd use this for X, but it needs to do Y.   Fake validation: This is so cool, I'd definitely use it.  Specific objections beat generic praise every time.  They show someone actually thought about it.

CPO, New York, NY · Answer

Watch what they're currently doing to solve the problem.  If they'd have to change their entire workflow to use yours, they won't.   Real validation is to solve it with their existing workflow remaining intact.  If they're not willing to change how they work, the problem wasn't urgent enough.

How do you know when an idea is actually validated?

Decision context

Decision graph

Operator judgments (12)

Operator judgment

Operator judgment

Operator judgment

Operator judgment

Operator judgment

Related decisions