Customer Discovery: How to Run Interviews That Find PMF

Why Most Customer Interviews Fail

90% of founders do customer interviews wrong. They ask leading questions, pitch their solution, and hear what they want to hear.

Real customer discovery is uncomfortable. It challenges your assumptions.

This is critical because 90% of startups fail – and the #1 reason is building something nobody wants.

The Mom Test

Rob Fitzpatrick's "Mom Test" is the gold standard:

Rule 1: Talk about their life, not your idea Rule 2: Ask about specifics in the past, not hypotheticals Rule 3: Talk less, listen more

Bad Questions

"Would you use an app that does X?" (Hypothetical)
"Don't you think X is a problem?" (Leading)
"Would you pay $10/month for this?" (Hypothetical + leading)

Good Questions

"Tell me about the last time you dealt with X"
"What have you tried to solve this?"
"How much did that cost you?"

The 5 Essential Questions

1. "What's the hardest part about [problem area]?"

Let them define the problem. Don't assume.

2. "Tell me about the last time that happened"

Get specific stories, not generalizations.

3. "Why was that hard?"

Understand the root cause.

4. "What have you done to solve it?"

If they haven't tried anything, it's not a real problem.

5. "What don't you love about your current solution?"

Find gaps you can fill.

Interview Structure

Before (5 minutes)

Thank them for their time
Explain you're researching, not selling
Ask permission to take notes

During (20-30 minutes)

Start broad, go specific
Follow interesting threads
Ask "why" repeatedly
Get specific numbers and dates

After (5 minutes)

Ask who else you should talk to
Thank them again
Send a brief follow-up

Reading Between the Lines

Strong Signals

They've already tried to solve it
They can quantify the cost
They get emotional describing the problem
They ask when your solution launches

These are early signs of product-market fit potential.

Weak Signals

"That would be nice to have"
"I could see someone using that"
"Sounds interesting"
They won't commit to a follow-up

Understanding these dynamics is crucial. Learn more about why early customers often mislead you – not intentionally, but through politeness.

How Many Interviews?

Minimum: 20-30 for initial validation Pattern recognition: After 10-15, you'll see themes Segmentation: Interview different customer types separately

Use these interviews to define your ICP – the specific customer segment that feels the problem most acutely.

Common Mistakes

1. Pitching Instead of Learning

You're there to learn, not sell. Save the pitch.

2. Confirmation Bias

You'll hear what you want to hear. Actively seek disconfirming evidence.

3. Not Recording Specifics

"Many people said..." is useless. Track exact quotes and numbers.

4. Interviewing the Wrong People

Talk to decision-makers who feel the pain and control the budget.

5. Stopping Too Early

First 5 interviews are warm-up. Keep going.

From Interviews to Action

After 20+ interviews:

Identify patterns – What problems came up repeatedly?
Rank by intensity – Which problems caused the most pain?
Check willingness to pay – Did anyone offer to pay for a solution?
Define your ICP – Who had the problem most acutely?

Then you're ready to build your MVP.

When Interviews Suggest a Pivot

Sometimes customer discovery reveals:

The problem isn't painful enough
Your target market doesn't have budget
Competitors already solve it well enough

This is valuable information. It might be time to pivot.

Take Action

Customer discovery never really ends. Even after PMF, keep talking to customers.

Ready to validate your findings? Our PMF Assessment helps you measure where you stand.

Check your validation strength →