How to validate a startup idea before you build it
7 min read · Updated 13 October 2026
Most startups fail because they build something nobody wants. Validation means finding out, cheaply and quickly, whether a specific group of people has the problem badly enough to change what they do today. Here's a process you can finish in about two weeks.
1. Write the idea as a problem, not a product
Before testing anything, write one sentence: "[Specific people] struggle with [problem] when [situation], and today they [current workaround]." If you can't fill in the workaround, you may not understand the problem yet.
Example: "Freelancers in Germany struggle to prepare quarterly VAT returns, and today they pay an accountant €80 a quarter or spend a weekend with spreadsheets."
2. Talk to 10–20 people who have the problem
Don't pitch. Ask about the last time the problem happened, what they did, what it cost them and what they tried. Listen for strong emotion and money already being spent.
- "Tell me about the last time this happened."
- "What did you do about it?"
- "What have you tried? What did you dislike?"
- "How much does this cost you, in money or time?"
3. Look for demand that already exists
Competition is usually good news: it proves people pay for a solution. Search for existing products, read their reviews and note the complaints. Your opportunity is often hiding in the one-star reviews.
4. Put the idea in front of people
Share your idea on UnbuiltHub with the problem, solution and target customers clearly written. Add a one-question poll ("Would you pay $10 a month for this?") and turn on the launch waitlist. Likes, comments, votes and waitlist sign-ups are early signals; a waitlist email is a stronger one than a like.
5. Ask for a small commitment
The real test is whether people give something: an email, a pre-order, a deposit, a pilot agreement or their time. "I'd use that" is polite. "Can I pay now?" is validation.
6. Do it by hand first
Before building software, deliver the result manually for a few customers. A spreadsheet, WhatsApp and your own time are enough. You'll learn what really matters and earn your first revenue.
7. Check the numbers
Estimate how many customers there are and what they'd pay. Use the market size calculator for TAM, SAM and SOM, and the CAC and LTV calculator to check each customer is worth more than it costs to win them.
When to move on
If after 20 conversations nobody describes the problem as painful, nobody is paying for a workaround and nobody joins your waitlist, change the customer or the problem, not just the features. Pivoting early is cheap; pivoting after a year of building isn't.