How to validate a startup idea in 7 days, for free

Most first startup ideas don’t die because someone stole them, or because the code was bad. They die because not enough people wanted them. In March 2026 CB Insights published an analysis of 431 venture-backed startups that shut down since 2023, and poor product-market fit showed up in 43% of the cases where the causes were known.

Those founders raised millions before finding out. A week is all you need. It costs nothing but a few evenings and some nerve. So here’s a seven-day plan to validate a startup idea for nothing: ten honest interviews, a look at search demand, a one-page waitlist in front of about 100 real people, and a decision rule written down before the first result lands.

What does it mean to validate a startup idea?

It means swapping opinions for evidence. You’re not trying to prove the idea is brilliant. You’re trying to find out, cheaply and fast, whether it’s wrong. Under the hood there are four guesses stacked on top of each other: this group has this problem, it hurts enough to matter, they’d pay to fix it, and you can find them.

Where all this is heading is product-market fit, the point at which a clear group of customers wants your product enough to pay for it and tell their friends about it. Validation comes first. It sits before your MVP (minimum viable product, the simplest version that really works) and long before investors.

Not sure your idea is a startup at all? Read what a startup is first. And if you’ve just read about Italian startups that started small, here’s how to give your own idea the same kind of modest first test.

Why bother? What the failure data says

CB Insights went through public post-mortems, founder interviews and shutdown announcements from 431 VC-backed companies that closed since 2023. Together they had raised $17.5 billion. The median company had raised $11 million. Money didn’t save them.

Running out of capital was cited in 70% of cases, but CB Insights itself calls it almost always the final cause of death, not the root problem. Poor product-market fit came next at 43%. Behind it sat bad timing or macro conditions (29%) and unit economics that never worked (19%). Most companies named several causes, hence the total above 100.

Here’s the line that matters for you. Two-thirds of the product-market fit failures were early-stage companies that never found a market. A week of validation won’t make you immune, but it’s the cheapest test you’ll ever run against that risk. If you want the full list of causes, our guide to why startups fail has it.

The 7-day plan to validate a startup idea

Seven days will do, even if all you have is the gaps between lectures or the hours after work. Keep the order: talk first, build last.

1. Day 1: the one-page brief. Put on one page who has the problem, when it bites, what they do about it today and what that costs them in euros or hours. “Students” is too vague to test; “first-year engineering students in Bologna who have failed a maths exam once” is not.
2. Days 2 and 3: ten problem interviews. Twenty to thirty minutes each, with people who match your page. Not your flatmates, not your parents. You’re not there to sell. Just ask when the problem last came up, and jot down the words they use, exactly as they say them.
3. Day 4: demand and competitors. Put the problem into Google Trends (the problem, not your product’s name), pick Italy and the last five years, and watch which way the curve is heading. After that, pull monthly search estimates from Google Keyword Planner. It’s free, though Google makes you finish a Google Ads account setup first. Round off the day by listing five existing solutions and what they charge.
4. Day 5: build a one-page site. You need a headline about the problem, three lines on your solution, a price and one button, either join the waitlist or pre-order. Free tools are plenty for this. Tally’s free plan, for instance, includes unlimited forms and submissions and can collect payments.
5. Day 6: the threshold, then the page. Do things in that order. Before a single person sees the page, write down three numbers: the one that means go, the one that means stop, and the grey zone between them. Then show it to about 100 people from your target group, in the places they already hang out: course WhatsApp groups, a LinkedIn post, a neighbourhood Facebook group, a subreddit.
6. Day 7: count and decide. Hold the results up against what you wrote on day 6, not against what you were hoping for. Above the line, build the smallest version. In the grey zone, change one thing (price, audience or headline) and repeat days 5 to 7. Below the line, reread your notes: the real problem may be next door, and a pivot, a change of direction based on what you learned, costs almost nothing at this stage.

Why fix the threshold in advance? Because once results arrive, everyone finds a reason why nine sign-ups is actually encouraging. There’s no reliable universal conversion rate for a test page, so don’t borrow one. Work backwards from your own numbers: how many customers, paying how much, would make this worth a year of your life?

How to run customer interviews: the Mom Test rules

Rob Fitzpatrick named his book after the most dangerous person to pitch: your mum. She loves you, so she’ll tell you the idea is great. The subtitle sums up the whole method: how to talk to customers and learn if your business is a good idea when everyone is lying to you.

It comes down to three rules. You talk about their life, not your idea. You ask what actually happened last month rather than what they might do next year. And you keep your mouth shut more than feels natural, because every instinct pushes you to start explaining the idea.

Instead of asking	Ask
Would you use an app that does this?	When did this last happen to you? What did you do?
How much would you pay for it?	What have you spent on this problem in the last year?
Do you think it’s a good idea?	What have you already tried, and why did you stop?
Would you tell your friends about it?	Who else do you know who has this problem?

Compliments are noise. “I’d definitely buy that” is only a prediction, and people hand out predictions for free. What you’re after is a commitment, meaning the other person parts with something they value. That could be a second meeting, an introduction to a colleague with the same headache, or a small deposit.

A worked example: a revision course in Bologna

Let’s invent a student called Luca, in his third year of engineering at Bologna. His younger flatmates keep failing Analisi 1, the first-year maths exam, and he wants to sell small-group revision sessions built on past exam papers.

In his interviews, six of the ten students had paid for private lessons in the past year. Almost all of them said the same thing in different words: they didn’t need more theory, they needed someone to correct their exercises. That’s the product. Luca had planned to record video lectures. So the video lectures go in the bin.

Before he shares the page (four sessions, groups of five, €40, with a refundable €10 deposit to book a place), Luca writes down his rule: eight deposits out of about 100 students is a go, fewer than three is a stop. He posts it in three course chats. Around 120 people open it, 19 leave an email and 5 pay the deposit.

Grey zone. Emails are cheap, deposits aren’t, and five is not eight. So he changes one thing, moving the sessions into the two weeks before the exam, and runs the test again with a new group. No code, no company, nothing spent except his evenings.

Real companies have started the same way. Joel Gascoigne tested Buffer, a tool for queueing up tweets, with a two-page website: he tweeted the link and asked people what they thought. Then he added a pricing page in the middle to see whether people were comfortable paying. Idea to paying customers took seven weeks.

Common mistakes when you validate an idea

• Asking friends and family. They want you to succeed, which is lovely but makes what they say useless as data.
• Building before talking. Two months of evenings spent on an app is a very expensive way to learn that nobody needed it.
• Moving the goalposts. Set the threshold after you’ve seen the results and any outcome can be made to look like a win.
• Hiding the price. A waitlist without a price measures curiosity. Show a number, even a rough one.
• Forgetting privacy. Collect emails in the EU and the GDPR comes with them. Tell people who you are, why you want their data and how long you’ll keep it, and put that in a short privacy notice on the form.

Your validation checklist

• The one-page brief is done: customer, problem, workaround, cost.
• Ten strangers who match that page, booked for interviews.
• At least three specific past stories, noted in their own words.
• Google Trends and Keyword Planner checked, and five competitors listed with their prices.
• A landing page showing a price and a single button.
• Your go and stop numbers on paper before about 100 target people see the page.
• A verdict on day 7: build, rerun the test or rethink.

Numbers said go? Money is usually the next worry, so start with how to start a startup with no money and then build an MVP without coding. Keep the beginner’s map from zero to first round open to see where you are. And when the time comes to raise capital in Italy, this week’s notes will be your first evidence.