
How to validate a startup idea: what counts as proof
Check the problem is real, worth solving and big enough, then ask for money, time or a signature. A scorecard, what counts as proof and a worked example.
A startup idea is validated when people who have the problem give something up to get it fixed: money, time, a pilot, a signed letter of intent. Compliments, likes, waitlist sign-ups and "I'd use that" don't count, however many you collect. They cost the person nothing, so they can't tell you what that person will do once it costs something.
Having the idea is the easy part. Most ideas are crap, and you still have to validate yours. Here is how to validate a startup idea in four steps, a scorecard for deciding whether a problem is worth solving, and a made-up example of two founders working through it before launch.
What counts as validation, and what doesn't
| Signal | Counts? | Why |
|---|---|---|
| "Great idea" or "I'd definitely use that" | No | It's free to say, and people are kind to someone excited about a project. |
| Likes, shares and comments on a post | No | The topic caught their eye. Nobody decided to buy anything. |
| Waitlist or newsletter sign-ups | No | An email address takes ten seconds to give. It's a list of people to call, not proof. |
| Survey answers saying they would pay | No | A guess about the future, made with no money on the table. |
| A free pilot with no end date and no price | Weak | Easy to say yes to, and just as easy to forget. |
| A second meeting with the person who signs off on spending | Yes | They're spending their time, and their boss's. |
| Access to their real work so you can do the job by hand | Yes | Handing over their real orders, diary or files costs them effort and trust. |
| A signed letter of intent with a price and a start date | Yes | A signature from the person who holds the budget. |
| A paid pilot, a pre-order or a deposit | Yes, the strongest | Money moved. |
The first five rows are interest. The last four are commitment, and only commitment is validation.
How to validate a startup idea in four steps
- Check the problem is real. Talk to people who have it.
- Check it's worth solving. Score what you heard.
- Check the market is big enough. A rough count, from the bottom up.
- Ask for a commitment. Money, a signature or real time, against a pass mark you set beforehand.
Each step costs more than the one before it, and each one can tell you to stop. Run them in order.
Step 1: check the problem is real
Talk to people who have the problem, about the last time it happened and what they did about it. Leave your idea out of it. Customer interview questions for startups has the questions, where to find people, how to run a 20-minute call and a note template for each one.
On how many: my rule is to keep going until you stop hearing anything new. That page breaks down what that means for a consumer app versus a business with few possible customers.
You come out of this step with a stack of notes, each with a real story in it. If most people can't remember the last time the problem happened, stop here. It isn't the problem you thought it was.
Step 2: is the problem worth solving?
A real problem isn't always one worth building a company around. Everyone has problems they would never pay to fix.
Two things decide it: how often the problem happens, and how much it hurts each time. Frequent and painful is worth solving. Rare and mild isn't, however real it is. In between, one of the two has to be strong: a small cost that hits every day, or a big one that hits a few times a year.
Score what you heard against six tests:
| Test | Weak sign | Strong sign |
|---|---|---|
| How often | "Now and then." Nobody can name the last time. | A date and a count: "Twice last week." |
| What it costs | A mild annoyance. | Money, hours or customers lost, and they can put a number on it. |
| What they do today | Nothing. A shrug. | A workaround: a spreadsheet, someone fixing it by hand, a tool they pay for and complain about. |
| What they've tried | They never looked for a fix. | They tried something, dropped it, and can say why. |
| Where it ranks | Below the other things on their list this month. | In the top two or three. |
| Who pays | The person with the problem has no budget and doesn't know who does. | They hold the budget, or will introduce you to whoever does. |
Listen for their own words too. "Hate" and "impossible" point at real pain; "interesting" doesn't.
Mark each test strong or weak for each conversation, then count across all of them. You want most of the people who fit your customer to show strong signs on the first three rows, and the third, "What they do today", matters most. If people already spend money or hours on a workaround, the problem hurts enough for them to act. That workaround is also the competition you'll have to beat, and market research for startups shows how to lay it out next to the other options.
If the strong signs show up in only two or three conversations, look at what those people have in common. The problem may be real for a narrower group than you thought, and that group is who you talk to next. If they have nothing in common, the problem probably isn't there.
Step 3: check the market is big enough
A problem can be real, painful and frequent and still be too small for the company you have in mind. Do a rough count from the bottom up: how many customers have this problem and could buy, times what one of them would pay in a year. How to calculate market size for a startup works through one from start to finish.
The number rarely kills an idea outright. It tells you what kind of company you're building. When I look at whether a founder should raise money, the test is scale and size. Investors look for companies that can grow fast and become massive. If you are a corner shop, you should not raise money. A smaller market can still be a good business, paid for by its customers. Bootstrapping vs fundraising helps you pick which road.
Step 4: ask for a commitment
This is the step founders are most tempted to skip, because it's the one where people can say no. Go back to the people from Step 1 with the strongest stories. Make a clear, small offer: what you'll do for them, by when, for how much. Then ask them to commit.
Write down your pass mark first. What result means you build, what result means you change something, what result means you stop. Decide before you ask, because afterwards it's very easy to talk yourself into whatever came back.
What to ask for, strongest first:
- Money. A pre-order, a deposit or a paid pilot. For business customers a paid pilot works well: a fixed length, a real price and one result you promise to deliver. Use a real price. A token price gets a polite yes and tells you very little.
- A signature. A short letter of intent from the person who holds the budget, saying they plan to buy, at what price, once what is ready. No money moves, so it counts for less than a payment. It still counts for far more than a yes on a call.
- Time. A second meeting with whoever signs off on spending. Access to their real work so you can do the job by hand for a few weeks. Introductions to people like them.
Always aim for payment from day one. A free pilot with no date for payment is easy to agree to and easy to let drift, and when it ends you've learned almost nothing. If a customer wants a free pilot, agree on the date they start paying before it begins. If they ask for a discount because it's a pilot, say yes and extend the contract: half price, for example, locked in for a year. How to get your first customers covers both.
For a consumer product: a pre-order, a deposit, or doing the service by hand for a handful of people and charging them for it. Sell what you can do for them now, not what you plan to build next year.
When people won't commit, ask why, with an open mind. "It costs too much", "we have something for that" and "not right now" point at three different things: the price, the competition and the urgency. Each needs a different fix.
Where a landing page fits
A landing page that describes the problem and the result you promise, and asks for an email, is a cheap way to see whether the problem catches attention. It also shows you which words make people stop. The startup landing page covers how to build one as a test.
But it tests interest. Never confuse a demand test with validation: waitlist sign-ups are not validation. There's no standard conversion rate to aim for either, so set your own pass mark before the page goes live. Then treat every sign-up as a name to call, and take them through Step 4.
A made-up example: the physio waitlist startup, before launch
This example is invented, the company and every number included. It's the same made-up startup as in the pitch deck outline for a 10-minute pitch, a few months earlier, before they had built the real product.
Two founders want to fill the appointments independent physio clinics lose to late cancellations. One ran the front desk for a group of clinics. The other built booking-system integrations at a health software company.
Step 1. They talked to about 20 clinic owners and receptionists. After about 15, the stories stopped changing. A patient cancels the evening before or that morning. The receptionist phones the waitlist one name at a time. By the time someone picks up, the slot has gone.
Step 2. Their scorecard, across those conversations:
| Test | What they heard |
|---|---|
| How often | Most clinics had several late cancellations a week. |
| What it costs | Each empty slot is a lost appointment, about €60. |
| What they do today | Phoning the waitlist by hand, the booking system's built-in email waitlist, or nothing. |
| What they've tried | Several had switched on the email waitlist and stopped relying on it. Nobody reads an email in time to take a slot two hours away. |
| Where it ranks | High for small clinics with one receptionist. Low for clinics where the physios fill their own diaries. |
| Who pays | The clinic owner, who usually also treats patients and is easy to reach. |
Strong on the first three rows for most of the clinics they spoke to, and clearest for small clinics with one receptionist. That became their customer.
Step 3. A rough count: say 8,000 physio clinics in Ireland and the UK, at €100 a month, or €1,200 a year each. About €9.6 million a year to start with, and every clinic that runs on appointments has the same problem.
Step 4. Before asking anyone, they wrote down their pass mark: at least three clinics pay for a two-month pilot. Fewer than that, and they would go back and rethink who the customer is.
By then they had also put up a landing page. A post one founder wrote about late cancellations got hundreds of likes from physios, and 150 people joined its waitlist.
The second founder put together a rough version in two weeks. The receptionist types in the free slot, it texts the waitlist, and the receptionist updates the booking system by hand. They went back to the 12 clinics with the strongest stories and offered a two-month pilot at €100 a month.
What came back:
- Three clinics paid for the pilot.
- Two more signed a letter of intent to start at €100 a month once it worked inside their booking system.
- Seven said some version of "let us know when it's ready".
Only the three payments and the two signatures counted. The pass mark was met, just. The seven "let us know" answers went in the same pile as the likes and the sign-ups: names to call later, not evidence.
It also changed what they built first. Both letters of intent depended on the link to the booking system, so that came before anything else. And the three pilot clinics became their first paying customers at launch.
When you have enough evidence to start building
You're ready to build when you can tick all five:
- Most of the people you spoke to who fit your customer told a recent, specific story about the same problem.
- They already spend money or time on a workaround.
- The rough market count fits the company you want to build, raising money or not.
- People have paid, signed or given real time, and the result beat the pass mark you wrote down beforehand.
- You can name your first customers. They're the people who committed.
Then build the smallest thing that serves the people who committed. How to build an MVP picks up from there. Put what you found on one page as well. Ash Maurya's Lean Canvas is made for that, and a Lean Canvas example, filled in box by box shows one finished.
If you can't tick them all, change one thing and run the steps again: a narrower customer, a different problem from your notes, a different offer. Or stop. Stopping after a few weeks of conversations costs far less than stopping after a year of building. Building before you have this evidence is the first of the startup mistakes to avoid.
Ready to Move Faster?
Get actionable feedback and insights on your challenges and goals. 30 minutes focused on helping you move forward.
No sales pitch. No obligations. Just helpful advice.