Validating a startup idea means testing whether real people want what you're planning to build before you spend months and money building it. The fastest, cheapest way to do that is to look for evidence of demand — signups, interviews, or actual payment — using something far smaller than a finished product: a landing page, a few conversations, a fake door. Most failed startups didn't fail because the team couldn't build; they failed because they built something nobody needed. Validation is how you find that out for a few hundred dollars and a couple of weeks, instead of finding out a year later with nothing left in the bank.
Why validation is the cheapest insurance you can buy
Building the wrong product is the single most expensive mistake a founder can make — not because the code is expensive, but because of everything that goes with it: the months of runway burned, the team hired around a false assumption, the pivot that comes too late to matter. Founders tend to fall in love with the solution before confirming the problem is even worth solving, and by the time the product ships, they're too invested to hear the market's real answer clearly. Validation flips the order of operations. Instead of building first and hoping someone shows up, you find out who shows up first, then build for them specifically. It doesn't guarantee success, but it kills the ideas that were never going to work before they cost you anything real.
Start with a landing page, not a product
A single landing page that describes your product, states the price, and ends in an email signup or waitlist button is enough to test real interest. You don't need a designer or a working backend — a plain page with a headline, three bullet points on what the product does, and a form is enough. Drive a small amount of traffic to it — a few hundred dollars of ads, a post in a relevant community, outreach to your own network — and watch the conversion rate. A page that converts at 2-5% of visitors into signups is a real signal; a page nobody signs up for, despite reasonable traffic, is telling you something too. The point isn't the number in isolation, it's that people took an action, not just said something nice in a survey.
Talk to customers, but ask the right questions
Interviews are the cheapest validation tool that exists, and also the easiest to get wrong. People are polite by default — ask "would you use this?" and almost everyone says yes. Ask about their past instead of their future: how do they solve this problem today, what did they last pay for something like it, what's the actual cost — in time, money or frustration — of not having a solution. Specific, past-tense answers reveal truth. Vague, future-tense enthusiasm reveals nothing except that the person likes you.
Run a fake-door test before you build the real thing
A fake door is a button, feature or offer that appears real but isn't built yet — click it and you get a "coming soon, join the waitlist" message instead of the actual feature. It's the fastest way to measure real demand for a specific feature or product before writing a line of production code. The same logic applies to pricing: put a "Buy now" or "Get started" button in front of people and see who actually reaches for their card, even if the checkout leads to a page asking for their email instead of taking payment.
Do the manual version before you automate anything
Before building software, deliver the outcome by hand. If your idea is a scheduling tool, manually match a handful of customers by email or spreadsheet. If it's an analytics dashboard, send the first ten customers a manual report you put together yourself, even if it takes you an evening each time. This is the "concierge MVP" — slow, unscalable, and exactly what you want at this stage, because it tells you whether the underlying problem is real and whether people will pay for the outcome, before you spend a cent automating a process nobody wanted. Plenty of founders skip this step because it feels beneath them; it's usually the fastest few weeks of learning they'll get in the entire life of the company.
Read your signals honestly — don't fool yourself
The easiest person to convince that your idea works is yourself. Compliments, "this is cool" reactions, and vague interest from friends and family are not validation — they cost the other person nothing to give. Real signals cost something: a real email address, a spot on a waitlist, a card number, time spent in an interview, a repeat visit. Weigh signals by what they cost the person giving them, not by how many of them you collected.
When you've validated enough to start building
You don't need certainty, you need a repeatable signal: a landing page converting consistently across more than one traffic source, a handful of interviews independently naming the same problem in their own words, or a few people willing to pay — even a deposit — for something that doesn't exist yet. One good week of data is luck. The same result across a few different tests is a pattern worth building for.
Frequently asked questions
How much traffic do I need to trust a landing page test?
A few hundred visitors from more than one source is usually enough to see a real pattern. If your only traffic is from friends and family, the results won't tell you much about strangers — try to include at least one channel of people who don't know you.
Is a fake door test dishonest to potential customers?
It's honest as long as you tell people what happens next — a "coming soon" message and a way to join a waitlist. What makes it unethical is charging for something you have no intention of building; asking for interest is fine, taking payment for a product that doesn't exist is not.
How many customer interviews are enough to validate an idea?
Somewhere between eight and twelve focused interviews is usually enough to see whether a pattern repeats. If the same specific problem and the same workaround come up unprompted from several people, you have a signal; if every answer is different, you likely haven't found the real problem yet.
What if my landing page gets signups but no one will pay?
That's a real result, not a failure — it usually means the problem is real but the price, framing, or audience is off. Go back to interviews, ask what they'd expect to pay and why, and test a different price or a narrower audience before concluding the idea itself is dead.