Most validation advice tells you to talk to customers. That is correct and it is not enough, because the failure mode is not skipping the conversations — it is having them and misreading what you heard.
People are polite. They will tell you an idea sounds useful, agree the problem is annoying, and never think about it again. Enthusiasm in a conversation is close to worthless as a signal, and mistaking it for validation is how founders end up six months into building something nobody wanted.
This is a guide to getting signals that actually mean something, in about two weeks.
The Only Question That Matters
Not "would you use this?" — the answer is almost always yes, because saying no to a hopeful person is awkward.
The real question is: what are they doing about this problem right now?
Everyone with a genuine, expensive problem has already built a workaround. A spreadsheet. A manual process every Monday morning. A contractor. A tool they hate but pay for anyway. Somebody who does it by hand.
If they have no workaround, the problem is not painful enough to pay to solve. That single question separates real problems from ones people merely agree are problems.
What Counts as Evidence
Ranked from weakest to strongest, because the weak signals are the ones that feel most encouraging:
Signal | What it means |
|---|---|
"That sounds useful" | Nothing. Politeness. |
"I'd definitely use that" | Nearly nothing. Hypothetical future behaviour. |
Email signup | Weak. Costs nothing to give. |
Describes their current workaround unprompted | Real. The problem exists. |
Already pays for something inadequate | Strong. Budget exists. |
Asks when they can have it | Strong. |
Pre-pays or commits payment details | Definitive. |
The gap between rows three and four is where most validation goes wrong. A landing page with 200 email signups feels like traction and frequently converts to nothing, because signing up costs a person nothing and commits them to nothing.
Past Behaviour, Not Future Intentions
The single most useful shift in how you ask questions.
Bad: "Would you pay for a tool that does X?"
Good: "Walk me through the last time you dealt with X."
The first asks someone to predict their own behaviour, which humans are reliably bad at. The second asks them to describe something that actually happened, which they can do accurately.
Follow-ups that work: How often does that happen? How long did it take? What did you try first? What did you do when that did not work? Did you look for a tool? Why did you stop?
You are listening for time, money, and frequency. A problem that costs someone four hours a month is a business. A problem that costs them ten minutes a quarter is not, however annoying it is.
A Two-Week Plan
Days 1–3: Find where the problem lives
Search Reddit, Hacker News, and niche forums for how people describe the problem in their own words. Not your product category — their frustration. Someone building expense tooling should search "chasing receipts from contractors," not "expense management software."
You are doing two things: confirming people complain about this unprompted, and collecting the vocabulary they use. That language goes into your landing page and your tagline later, and it will be better than anything you would have written yourself.
Days 4–8: Have ten conversations
Ten is enough. Patterns show up by the sixth or seventh, and if they do not, that is itself an answer.
Find people in the communities you identified. Ask for fifteen minutes to hear about how they handle the problem — not to show them anything. Do not pitch. The moment you describe your solution, they start being polite instead of informative.
Take notes on what they currently do, what it costs them, and what they have already tried and abandoned.
Days 9–11: Ask for money or a commitment
This is the step people skip, and it is the only one that produces a definitive answer.
Go back to the people whose problem sounded real and make a specific offer: you are building this, it will do exactly these three things, it will cost this much, and you would like them to be an early customer. Ask for a pre-order, a deposit, or a firm commitment with a date.
You do not have to take the money. What you need is the reaction. Someone who says "yes, send me an invoice" has validated your idea. Someone who says "definitely keep me posted" has not, and the difference between those two responses is the entire exercise.
Days 12–14: Decide honestly
Three to five genuine commitments from ten conversations is a strong signal. Zero commitments from ten conversations is also a signal, and it is worth more than the two months you just saved.
The uncomfortable middle — lots of interest, no commitments — usually means the problem is real but not expensive enough, or you are talking to the wrong people. Both are fixable, and both are cheaper to fix now.
Landing Page Tests, Honestly Assessed
The standard advice is to put up a landing page, drive traffic, and count signups. It is useful for one narrow thing: testing whether your explanation works. If people arrive and do not sign up, your positioning is unclear.
It is much weaker at testing whether the product should exist. Email signups measure momentary curiosity, and the conversion rate from "gave email to a landing page" to "paid for a product" is brutal.
If you run one, at least make it harder: ask for a small deposit, or add a question that requires actual thought. Ten people who answered a real question are worth more than three hundred email addresses.
What Not to Do
Do not ask friends and family. They are optimising for your feelings, not for accuracy.
Do not build an MVP to validate. An MVP is what you build after validation, not the instrument of it. Building takes months; the conversations above take two weeks.
Do not validate with a survey. Surveys collect hypothetical answers at scale. Ten conversations beat two hundred survey responses, every time.
Do not validate the solution. You are testing whether the problem is expensive, not whether people like your approach. The approach will change; the problem either exists or it does not.
Do not stop at "yes, that's a real problem." Everyone has real problems they will never pay to solve.
When Validation Says No
Most ideas do not survive this, which is the point.
The useful thing is that a failed validation usually hands you the next idea. In ten conversations about a problem you thought mattered, someone almost always mentions a different problem they care about more. That aside is frequently worth more than the idea you started with.
Listen for it. The best ideas tend to arrive as digressions in conversations about something else.
The Short Version
Ask what they do about the problem now, not whether they would use your thing.
Past behaviour is evidence. Stated intentions are not.
Ten conversations beat any number of survey responses.
Ask for money or a firm commitment. Interest is not validation.
Two weeks of this beats two months of building.
Once you have something worth showing, our launch checklist covers what actually breaks launches, and this guide covers getting your first users with no audience.
