How many customer interviews do you need to validate a startup idea?

Aim for 20–30 conversations in the first few weeks. This is the minimum to see patterns rather than anecdotes. After 5–10 interviews you may start hearing recurring themes, but 20+ gives you enough signal to make confident decisions about whether the problem is real, who has it most acutely, and whether your assumptions need revision.

What is a smoke test for idea validation?

A smoke test (or demand test) checks whether people will take a meaningful action — sign up, pay, or request access — before a product is built. Common formats include a landing page with a payment CTA, a fake-door test (a button that leads to a waiting list), or a pre-sale. The goal is to measure revealed preference, not stated intent.

What questions should you ask in customer discovery interviews?

Focus on past behavior and specific situations: 'Tell me about the last time you faced this problem,' 'How are you currently solving it?', 'What does it cost you — in time, money, or frustration?', 'Have you tried other solutions? What happened?' Avoid asking 'Would you use X?' — hypothetical questions generate polite but unreliable answers.

How do you know when a startup idea is validated?

An idea is validated when real people take a concrete, costly action that demonstrates genuine demand — paying money, signing a letter of intent, or repeatedly returning to use an early product. Validation is not verbal enthusiasm ('that sounds great!'), email signups from friends, or surveys. Set a specific numeric bar before running experiments and hold yourself to it.

How to Validate Your Startup Idea Before Building

Why Validation Matters

Every failed startup has one thing in common: they built something nobody wanted.

Not because the founders were incompetent — many were brilliant engineers, designers, and marketers. They failed because they skipped the most important step: confirming that a real problem exists, that real people have it, and that they’d pay for a solution.

Validation is not a bureaucratic checkbox. It’s the fastest path to learning whether to continue, pivot, or stop — before you’ve spent a year and $200,000 finding out the hard way.

Step 1: Write Down Your Assumptions

Before talking to anyone, capture every assumption embedded in your idea:

Customer assumption: Who exactly has this problem? (job title, company size, demographics, behavior)
Problem assumption: What pain do they feel? How severe is it? How often?
Solution assumption: Does your specific approach solve it better than alternatives?
Willingness-to-pay assumption: Would they pay for a solution? How much?

List them all — then prioritize by which assumption, if wrong, would kill your idea fastest. That’s what you validate first.

Step 2: Get Out of the Building

Talk to potential customers. Not your friends. Not your family. Not other startup founders (unless they’re your target customer).

Aim for 20–30 conversations in the first few weeks. This sounds like a lot; it’s not. It’s the minimum to see patterns.

How to Find People to Talk To

LinkedIn: search by job title + industry
Reddit, Facebook groups, Slack communities for your target user
Relevant conferences, meetups, events
Your existing professional network (first-degree contacts)
Twitter/X: search for people complaining about the problem

The Right Ask

Don’t pitch. Ask for a 20-minute conversation to understand how they do X:

“Hi [Name], I’m researching how [job title]s handle [problem area]. I’m not selling anything — I want to understand the challenges you face. Would you have 20 minutes this week?”

Response rates of 10–20% on cold LinkedIn are realistic with a genuine, concise ask.

Step 3: Run Discovery Interviews

Interviews are not surveys. You’re looking for stories and behaviors, not opinions.

The Mom Test (Core Rules)

Rob Fitzpatrick’s The Mom Test distills interview best practices into three rules:

Talk about their life, not your idea
Ask about specifics in the past, not generics or hypotheticals
Listen more than you talk

Good vs. Bad Questions

Bad (Biased)	Good (Discovery)
“Would you use a tool that did X?"	"How do you currently handle X?"
"Is this a big problem for you?"	"How much time/money does this cost you monthly?"
"Would you pay $50/month for this?"	"What have you already tried to solve this? How much did that cost?”

What to Listen For

Complaints about the current workflow (unprompted = gold)
Workarounds they’ve built themselves (strong signal of unmet need)
Budget: have they paid for similar solutions? How much?
Urgency: is this a burning problem or a nice-to-have?

Step 4: Analyze Patterns

After 10 interviews, look for recurring themes:

What phrases keep coming up?
What workarounds do multiple people have?
Which assumptions were confirmed? Which were wrong?

A shared problem mentioned independently by 7 out of 10 people in similar roles is a signal. One enthusiastic person who “loves the idea” is noise.

Step 5: Test Demand Without Building

If discovery interviews confirm the problem, test whether people will take action — not just say they would.

Demand Validation Techniques

Landing Page Test Build a one-page site explaining the product. Drive 200–500 targeted visitors. Measure email sign-ups or “request access” clicks. A 10–20% conversion rate indicates genuine interest.

Pre-sales Ask people to pay (or put down a deposit) before the product exists. Even $1 tests commitment. “I’d use that” costs nothing; pulling out a credit card costs something.

Wizard of Oz Manually deliver the service before automating it. If customers pay for a manually-delivered version, the automated version has validated demand.

Concierge MVP Personally do the work for 5–10 customers. Learn what matters, what doesn’t, and what they’d actually pay for.

Step 6: Define Your Validation Bar

Before running experiments, decide what “validated” means:

“5 people sign up for the waitlist” is not validated
“10 non-friends pay us $50 in advance” is a much stronger signal
“3 companies agree to pilot and sign a letter of intent” is even better for B2B

Set the bar before you run the test — otherwise you’ll rationalize weak signals.

Red Flags to Watch For

Everyone says it’s a great idea but nobody will pay
You can only find enthusiastic users through personal connections
The “customer” is excited but someone else holds the budget
The problem exists but there’s already a dominant, entrenched solution

Key Takeaway

Validation is not about proving your idea right. It’s about learning as fast as possible whether it’s worth pursuing. The best founders treat their startup idea as a hypothesis — and run experiments until they have enough evidence to bet on it or kill it and move on.

The goal of validation is not to feel good about your idea. It’s to know.