How to Validate a Startup Idea Before You Build

To validate a startup idea, find out whether real people have the problem, feel it strongly, already spend time or money on it, and will pay or commit to your solution. Do that with customer interviews, then a cheap demand test such as a landing page or preorder, before you build anything substantial.

By the Revolutionary Startups editors · Reviewed 11 November 2026 · 15 min read · 3,243 words

What it means to validate a startup idea

Validation means collecting evidence that a specific group of people has a specific problem and will take a specific action to solve it. It is not the same as getting compliments. Friends, family and strangers at a networking event will tell you your idea sounds great. That costs them nothing and tells you almost nothing.

The only evidence that counts is behavior and commitment. Someone describes the last time they hit the problem and what they did about it. Someone gives you their email address, their calendar time, a deposit, or a signed letter saying they would buy. The more it costs the other person, the stronger the signal.

What you are actually trying to learn

Almost every idea rests on four beliefs. Validation is the work of testing each one:

• The problem is real. People run into it regularly, not once in a blue moon.
• The problem is painful enough. They already spend money, time or effort working around it.
• You can reach them. There is a place or channel where these people can be found without a giant advertising budget.
• Your solution is something they would pay for. Not just a problem worth solving in the abstract, but your particular approach at a price that works for you.

Validation is not a single event

You do not validate once and then relax. You reduce uncertainty in stages. First you test the problem, then the solution, then the price, then whether people come back. Each stage is cheap compared with the next one. The goal is to spend a few weeks and a few hundred dollars finding out that something will not work, instead of a year and your savings. The leading cause of failure is building something nobody wants badly enough, which you can read more about in common reasons why startups fail.

Validation also does not guarantee success. It lowers the odds of an expensive mistake. Plenty of validated ideas still fail on execution, timing or money. But an unvalidated idea is a guess, and you can do better than guess.

Step 1: Write down your assumptions first

Before talking to anyone, put your idea on one page. This feels like homework, but it keeps you from hearing what you want to hear later.

The one-page idea statement

Fill in these blanks as plainly as you can:

1. Who: the specific type of person or business. "Small businesses" is too wide. "Dental offices with two to five dentists" is a place to start.
2. Problem: what they struggle with, in their words, not yours.
3. Current workaround: what they do today, including doing nothing, spreadsheets, hiring someone, or buying a competing product.
4. Your solution: one or two sentences.
5. Why now: what changed that makes this possible or urgent.
6. How they would pay: a rough guess at the model and price.

Rank your riskiest assumptions

Now list everything that has to be true for the idea to work, and circle the two or three that would kill it if false. For example: "dentists will give up their current scheduling tool," or "parents will pay monthly, not once." Test those first. If the riskiest belief fails, nothing else matters, so do not spend time polishing the logo.

A useful trick is to write each assumption as something you could be wrong about, with a number: "At least 6 of 10 people I interview will say they have hit this problem in the last month." That gives you a pass or fail line before the data arrives, which protects you from moving the goalposts.

Step 2: Customer discovery interviews

Customer discovery is the plain practice of talking to potential customers to learn how they live and work. The aim is to learn about their world, not to pitch. A good interview feels more like a curious conversation than a sales call.

Who to talk to and how to find them

Aim for 10 to 20 conversations with people who match your "who" line. Start with people one step removed from you: friends of friends, former colleagues, members of online communities and forums where your target group gathers, local groups, or professionals you can reach on a professional networking site with a short, honest message. Say that you are researching a problem, not selling, and ask for 20 minutes. Many people will say yes if you are respectful and specific.

Avoid relying on family and close friends. They want to be supportive, and they may not be in your target group anyway. If you only talk to people who like you, you will learn about your social circle, not your market.

How to run the conversation

• Ask about the past, not the future. "Tell me about the last time you had to do this" produces facts. "Would you use something that did this?" produces polite guesses.
• Let them talk. Aim for them to speak about 80 percent of the time. Silence is fine. Wait a beat before filling it.
• Dig into specifics. When they mention a pain, ask what happened next, how long it took, and what it cost them.
• Do not describe your idea until late, if at all. The moment you pitch, people start to be nice.
• Take notes, or record with permission. Write down their exact words. Patterns show up in phrases that repeat.
• End by asking for the next step. An introduction to someone else with the problem, a follow-up call, or a trial are all commitments.

What a strong signal looks like

You hear the same problem described unprompted by several people. They have already tried to fix it, even in clumsy ways. They can tell you roughly what it costs them. They ask when your product will be ready. They offer to introduce you to a colleague. Weak signals are phrases like "that's interesting," "I might use that," and "you should definitely do that." Those are compliments, not evidence.

Customer discovery interview questions: a bank of good and bad

The wording of a question changes the answer you get. Good questions are about real, recent behavior and cannot be answered with a polite yes. Bad questions invite flattery, speculation or a lie told to be kind. Use the table to swap one for the other.

Good questions to open with

• Tell me about your role and what a typical week looks like.
• What is the hardest part of [the area you are studying]?
• Tell me about the last time that happened. What led up to it?

Good questions to find real pain

• What did you do about it?
• What have you tried so far? What did you like and dislike?
• How long does that take, and how often?
• Who else is involved or affected?
• If you could change one thing about how this works, what would it be?
• Why is that a problem for you, and what happens if it does not get fixed?

Good questions to test commitment

• Are you currently paying for anything to solve this? How much?
• Who else would need to agree before you bought something like this?
• Where do you look for solutions when you have a problem like this?
• Would you be willing to try an early version and give feedback in two weeks?
• Who else should I talk to about this?
• Is there anything I should have asked and did not?

Questions to be careful with

If you must ask about price, ask about a concrete past purchase or budget, not a hypothetical. And be careful with "why" questions that sound like an interrogation. "What led you to that?" reads softer than "why did you do that?".

Step 3: How to test demand for a product without building it

Interviews tell you about the problem. Demand tests tell you whether people will act. Pick the cheapest test that fits your idea. The sequence below goes from lowest to highest effort.

Landing page test

A landing page test is a simple one-page website that describes your offer as if it already existed, with a single call to action such as "Join the waitlist" or "Reserve your spot." You send a small amount of targeted traffic to it and count how many visitors take the action.

• State the problem and the promised result in the headline, in the customer's language from your interviews.
• Explain how it works in three short points.
• Show a price or a price range if you can, since a free signup proves less than a real price.
• Ask for an email address, and ideally one more answer, such as "what is your biggest frustration with this today?"
• Drive traffic from places your target group already is: a community post, a newsletter, a small paid campaign, or direct outreach.

Decide your success number in advance. For example, if 100 targeted visitors arrive and you hope for at least 10 signups, say so before you look at results. The right number depends on your traffic source and audience, so treat any rule of thumb you read online with caution. What matters is a consistent comparison: does a version of your message get a clearly better response than another? Be honest about the limits too. A signup is a weak signal because it costs nothing, and the people who visit may not be typical. Treat the result as one piece of evidence, not a verdict.

Preorders and deposits

Asking for money is the strongest test. If you can take a preorder, a refundable deposit, or a signed letter of intent from a business, you learn much more than you would from any survey. Be clear about what the person is buying and when it will arrive, and do not take money for something you do not intend to deliver. Check the consumer protection rules that apply to preorders where you sell, and confirm refund practices with a lawyer if you are unsure.

Concierge and "Wizard of Oz" tests

In a concierge test you deliver the service by hand to a handful of customers, with no software. In a Wizard of Oz test the customer sees what looks like a finished product, but you are doing the work behind the scenes. Both teach you what people really need, how they behave, and what they would pay, before you write much code. They do not scale, and that is the point. A clear next step is described in how to build an MVP.

Fake door and pilot tests

A fake door test places a button or menu item for a feature that does not exist yet and measures clicks. A pilot is a short, paid trial with one or two real customers. For business software, one paying pilot customer is worth more than fifty survey replies.

Look at what competitors already prove

Existing competitors are not always bad news. They can show that people pay to solve the problem. Study their reviews, pricing and complaints, and look for a gap you can serve better. A market with no competitors can mean no demand, so ask why nobody is there.

A worked example

Suppose you think busy parents would pay for a weekly meal plan that adapts to what is already in their fridge. This is a hypothetical, with round numbers.

Assumptions written down. Who: parents of children under 10 who cook most nights. Problem: deciding what to cook. Riskiest assumptions: they feel this strongly enough to pay, and they will enter fridge contents regularly.

Interviews. You speak to 15 parents. Say 11 describe the nightly "what's for dinner" stress, but only 3 have ever paid for anything to help, such as a meal kit, and most gave up because it cost too much. Almost nobody would type in their fridge contents each week. That already changes the idea: the pain is real, but the data entry step looks like a dealbreaker.

Adjusted idea. A plan built around a short list of staples, with a shopping list, at a lower price than a meal kit.

Landing page test. You build a one-page site with a price of, for example, a small monthly fee, and share it in two parent communities. You decide in advance that 5 percent of visitors joining a paid-founder waitlist is a green light and under 1 percent means rethink. You get your numbers back and compare.

Concierge test. You manually create plans for five families for a month and charge a small amount. You learn what they actually use, what they ignore, and whether they renew.

Every step took days or weeks, cost very little, and changed the product. That is validation working: the idea you end with is better than the one you started with, or you learn to stop.

Startup idea validation checklist

Use this list as a gut check before you invest real money. If you cannot tick most of these, keep testing.

1. I can describe my target customer in one sentence, narrow enough to find them.
2. I have talked to at least 10 people in that group who are not friends or family.
3. Several of them described the problem in their own words without my prompting.
4. Most of them already spend time or money on a workaround.
5. I know roughly what the problem costs them each month.
6. At least a few agreed to something that costs them: a follow-up, an introduction, a trial, a deposit or a preorder.
7. I know where to find more people like them and have a rough idea of what that costs.
8. I have looked at competitors and alternatives, including the option of doing nothing.
9. I ran at least one demand test with a pass or fail number set in advance.
10. I have a rough sense of pricing, and the math could plausibly work. See business startup budget for how to check costs.
11. I can describe the smallest version I could build or deliver by hand to test the core promise.

Common validation mistakes

• Asking friends and family. They are kind, which is not the same as honest.
• Pitching instead of listening. If you spend most of the call explaining, you learned nothing.
• Counting compliments as demand. "Great idea" does not predict a purchase.
• Surveys with leading questions. A survey can confirm anything you want it to if you write it that way.
• Interviewing too few, or too similar, people. Three conversations are anecdotes. Ten to twenty with the right mix start to show patterns.
• Moving the goalposts. If your test fails, resist the urge to say "the test was wrong" without a real reason.
• Building first. Months of code is the most expensive way to ask a question.
• Ignoring who pays. The user and the buyer may be different people, and both need to say yes.
• Giving up after one no. One bad conversation is not evidence either way. Look for the pattern across people.

Also be careful about what you share. If your idea involves something you may want to protect, read about trade secrets and NDAs first, though in practice most early ideas are not harmed by customer research and many people will not sign an NDA for a casual interview.

What to do this week

1. Monday: write the one-page idea statement and your three riskiest assumptions.
2. Tuesday: make a list of 30 people in your target group and where to reach them. Send 15 short, honest messages asking for 20 minutes.
3. Wednesday to Friday: run your first interviews using the question bank above. Write notes within an hour of each call.
4. Weekend: tally patterns. Which problem do people describe most? What did they already try? Rewrite your idea based on what you heard.
5. Next week: pick one demand test, set a pass or fail number, and run it.

When the evidence is positive, move on to building the smallest useful version and finding your first customers. The wider path is in how to start a startup step by step, and later you will want to check for product market fit. If the evidence is negative, that is a result too. You saved time and money, and you can pivot the idea or move on to a better one.

Frequently asked questions

How do I validate a startup idea with no money?

Talk to potential customers. Interviews cost nothing but time. Then run a free or very cheap test, such as a simple landing page, a post in a relevant online community, or delivering the service by hand to a few people. Money only becomes necessary when you want to run paid traffic or build software.

How many customer interviews are enough?

There is no magic number, and it depends on the market. Many founders find that 10 to 20 interviews with the right kind of people begin to show clear patterns. If you keep hearing new things, keep going. If every conversation sounds the same, you probably have what you need.

What is a landing page test?

A landing page test is a single web page that describes your product as if it already existed and asks visitors to take one action, such as joining a waitlist or preordering. You send targeted visitors to it and measure how many act. It tests interest in your message and offer before you build anything.

Should I sign an NDA before talking to customers about my idea?

Usually not. Most people will not sign one for a casual conversation, and early customer research rarely exposes something valuable. Ideas are less valuable than execution. If your idea involves a specific invention or secret, speak to a lawyer first about what is worth protecting.

What is the difference between validation and market research?

Market research looks at the size and shape of a market, usually from reports and data. Validation tests whether specific people will actually behave a certain way toward your specific offer. You often need both, but validation gives you direct evidence from real customers rather than general information.

What if people say my idea is great but nobody signs up?

Believe the behavior, not the compliments. Low signups usually mean the problem is not painful enough, your message is unclear, or you are reaching the wrong audience. Go back to interviews, ask what they do today, and test a clearer offer before deciding to give up or change direction.

Key terms on this page

• Term Sheet
• NDA
• Pivot

Keep reading

• How to Build an MVP: A Step-by-Step Guide GuideHow to build an MVP step by step: define the one problem, cut features, pick no-code or code, set a timeline, and test with real users before you spend more.
• How to Get First Customers: 10, then 100, Without a Budget GuideHow to get first customers for your startup: find early adopters, run cold outreach, do things that don't scale, and reach 10 then 100 with little or no money.
• How to Find Product Market Fit: Signs and Tests GuideHow to find product market fit and know when you have it: signs to watch, a simple survey, how long it takes, and what to do if the signs are not there yet.
• How to Start a Startup: A Step-by-Step Roadmap GuideHow to start a startup step by step: validate the idea, pick a structure, build an MVP, handle legal basics, fund it and get customers, even with a day job.
• Why Startups Fail: Common Reasons and How to Avoid Them GuideWhy startups fail: no market need, running out of cash, team conflict, poor timing and weak sales. Learn the warning signs and how to avoid startup failure.

Part of Starting a Startup · All guides · Glossary

General information, not legal, tax or financial advice. Reviewed 11 November 2026.