How to validate a business idea with real evidence before you build

Validation comes down to one discipline. Collect evidence that people have the problem, will change what they do to solve it and will pay, before you spend money on code. This guide shows how to validate a business idea with the cheapest evidence first. Start with conversations, then run a smoke test, then ask for a commitment such as a pre-order or a paid pilot.

The stakes are real. A March 2026 CB Insights analysis looked at 385 venture-backed startups that shut down from 2023 onward with identifiable reasons. Poor product-market fit was one of the causes for 43% of them. Below are the market validation process we use with founders, the go, pivot or stop signals we read, and market validation examples you can adapt. At BrainStormers, Roger Grekos, our co-founder for product, projects and scale, leads this work.

What is market validation?

Market validation is the process of checking that a market exists for your idea before you build it. That means enough people you can reach have the problem, care about it and will pay to solve it. It looks at the buyers, not at your product, so it comes first in any plan to validate a business idea.

Market validation answers three questions: who has the problem, how they deal with it today and what they already spend on it. Testing your particular solution comes after that. A strong product cannot rescue a problem nobody will pay to fix. That is why the five steps below test the problem and the buyer first.

What validation proves, and what it does not

Validation does not prove your idea will succeed. It reduces four risks. The problem: people have it often enough to care. The buyer: someone with a budget will pay to solve it. The price: they will pay enough to build a business on. The channel: you can reach them at a sensible cost. The same steps answer how to validate a product idea inside an existing company and how to validate a startup idea from scratch.

Test the four risks in order

Test the problem before the buyer, the buyer before the price, and the price before you spend on the channel. Each answer changes the next question, so founders who validate business ideas out of order end up spending money on the wrong audience.

Evidence also comes in grades. Opinions are the weakest: people are polite, and "I would use that" costs them nothing. Past behaviour is stronger: what they did the last time the problem came up, and what they spent on it. Commitments are strongest: time, introductions, deposits and signed orders. Plan your tests so each one buys a higher grade of evidence than the last.

How to validate a business idea in five steps

1. Write down what must be true

Turn the idea into statements you can test. For example: "Office managers at clinics with five to twenty staff lose hours each week chasing no-shows, and they already pay for a reminder tool they dislike." Then rank the statements by how likely they are to be wrong, and set your go, pivot and stop thresholds now, before any data arrives. Thresholds written after the results come in tend to move.

Deliverable: an assumptions sheet with the thresholds that will decide the outcome.

2. Talk to people who have the problem

Interview people in your target segment about their past behaviour, not about your idea. Ask about the last time the problem happened, what they did, what it cost them and what they tried before. Listen for repeated phrases, workarounds they built themselves and money they already spend. Save your pitch for the end, if you use it at all.

How many conversations is enough? Research on qualitative interviews gives a useful anchor. In a 2006 study of 60 in-depth interviews, new themes had mostly stopped appearing by the twelfth interview. The basic elements were visible after six, in a fairly uniform group. Plan for about a dozen conversations per customer segment and keep going while you still hear new things. For interview technique in more depth, our guide that answers "What is product discovery?" covers how to run them.

Deliverable: interview notes grouped by theme, with the exact words people used.

3. Size the market from the bottom up

Skip the "1% of a huge market" slide. Count the buyers you could actually name, then multiply by a price they told you is realistic. Official statistics make this easier than most founders expect. Statistics Canada's Canadian Business Counts break businesses down by industry, employment size and province. The US Census Bureau's County Business Patterns counts establishments with paid employees by industry. For a consumer product, count the reachable people in the channel you plan to use.

Deliverable: a one-line, bottom-up estimate you can defend: buyers, times price, times the share you can realistically reach.

4. Test demand with a smoke test

A smoke test is a landing page that describes the offer, shows the price and asks for a next step, such as joining a waitlist or choosing a plan. It is one of the cheapest ways to test business ideas with people who owe you nothing. Send traffic from the channel you would really use, whether that is ads, a community or direct outreach, and measure how many visitors take the step. Traffic from friends only tells you about your friends.

Be honest about what exists. Canada's Competition Bureau explains that advertising is judged on its general impression as well as its literal meaning. A page that implies a product is ready when it is not can mislead, even if each sentence is technically true. Say plainly that the product is in development and what happens after someone clicks.

Deliverable: the share of visitors who took the next step, by channel, against the threshold you set in step 1.

5. Ask for a commitment

The strongest evidence is a commitment that costs the customer something. For consumers, that can be a pre-order or a refundable deposit. For businesses, it can be a paid pilot with agreed success criteria or a signed letter of intent that names a budget and a start date.

You do not need a product to take a payment: Stripe Payment Links let you accept payments without building a checkout. Mind the mechanics. Stripe notes that an authorization on an online card payment is usually valid for seven days. A long pre-order needs a different setup, such as saving the payment method and charging later. If you pre-sell physical goods to US customers, the FTC's Mail, Internet, or Telephone Order Merchandise Rule applies. It expects a reasonable basis for any shipping date you state, or shipment within 30 days if you state none. When you cannot ship on time, it expects you to offer a refund.

Deliverable: a count of commitments, their value and who made them.

How to validate SaaS ideas

To validate SaaS ideas, focus on the workflow the software would replace. SaaS buyers rarely start from nothing: they already use a spreadsheet, an inbox or a tool they resent. Ask to see it. Watching someone walk through their current spreadsheet tells you more than an hour of opinions.

• Run it by hand first. Deliver the outcome manually for a few customers, using spreadsheets and email. If they do not value the result when it is done for them, automating it will not change that.
• Put a real price on the page. Show plans and prices on the smoke test. Price is part of what you are validating, not a detail for later.
• Find the buyer, not just the user. In B2B, the person with the problem is often not the person who signs. Ask who approves the spend, from which budget, and what the purchasing steps are.
• Sell a paid pilot. Agree on success criteria and a conversion price before the pilot starts, so a good result turns into a contract instead of another meeting.
• Watch for nice-to-have signals. No current workaround, no budget owner and no deadline usually mean the problem is real but not urgent enough to buy a solution for.

Go, pivot or stop: reading the signals

Read the results against the thresholds you wrote in step 1, not against how you feel on the day. The table shows the patterns we look for. Your thresholds will differ by market, so set them before you start.

A pivot changes one thing, such as the user, the problem, the price or the channel, and then reruns the cheapest test that checks it. A stop is a result, not a failure: you bought the answer for the cost of a few weeks of research instead of a funding round. A go means you have earned the right to build, and the next questions move to scope and MVP testing.

Market validation examples

These market validation examples show the methods above in practice. The first is documented by the founder; the others are hypothetical patterns you can adapt.

• Buffer, 2010. Founder Joel Gascoigne described his test as a two-page MVP. One page presented the idea, and a second page let interested visitors leave an email to hear about the launch. He then added a pricing page between the two, so visitors chose a plan before leaving their email. A small number clicked on paid plans. Only then did he build the first version, and he reported the first paying customer within four days of launch.
• A B2B letter-of-intent test. For example, a founder building scheduling software for logistics teams might send a one-page proposal, with a price, to the operations leads they interviewed. Signed letters of intent naming a budget would be a go. Enthusiasm with no budget owner would point to a pivot toward a different buyer.
• A consumer price test. A team testing a meal-planning app might run small ad campaigns to two landing pages with different prices. If visitors reach the pricing step at one price and stall at the other, the team has learned something about willingness to pay before writing any code.
• A concierge pilot. A team exploring software for independent clinics might run patient reminders by hand for five clinics for a month, for a small fee. Clinics that renew are the strongest evidence that a reminder product deserves to exist.

Common ways founders fool themselves

Most validation mistakes come from wanting the idea to work. These are the ones we see founders make, and the fix for each.

• Interviewing friends and family. They want to support you, which makes their answers weak evidence. Recruit strangers who have the problem.
• Pitching instead of listening. Once you describe your solution, people react to it instead of telling you about their problem. Ask about the past first.
• Counting the wrong things. Likes, followers and compliments are not demand. Count behaviour and commitments.
• Testing the solution before the problem. A polished prototype shown to people who do not have the problem proves nothing.
• Moving the goalposts. Lowering the threshold after the results come in turns a cheap stop into a slow, expensive one.

Validation ends when the evidence says go and the build begins. After launch, the question changes from "do people want this?" to "does what we built meet that need?" A product-market fit survey helps answer it.

Frequently asked questions