What Customer Discovery Interviews Actually Are

Customer discovery interviews are structured conversations designed to understand a specific person’s behavior, frustrations, and priorities — not to pitch your idea or validate that you were right.

This is the distinction most founders miss. You are not running these interviews to find people who agree with you. You are running them to understand reality well enough to build something people actually need.

Discovery Interviews vs. Other Research Methods

Method	What it tests	Best for	Limitation
Discovery interview	Behavior, context, pain	Formative research	Time-intensive; n=small
Survey	Stated preferences at scale	Measuring after you know what to ask	No follow-up; easy to lie
Focus group	Group reactions to stimuli	Concept testing	Social dynamics distort answers
A/B test	Behavioral response to a specific change	Optimization	Requires existing traffic/product

Discovery interviews belong at the beginning — before surveys, before A/B tests, before a product. Their job is to tell you what questions to ask later, and whether there is even a problem worth solving.

The Mom Test Principles

Rob Fitzpatrick’s The Mom Test is the most useful 100 pages written on customer interviews. The core principle: ask about their life, not your idea.

Your mom (and most people) will tell you your idea is great to be kind. To avoid this, follow three rules:

Ask about the past, not the future. “Have you ever…” beats “Would you ever…”. People lie about future behavior. They can’t lie about what they already did.
Ask about specifics, not generalities. “How often does this actually happen?” beats “Is this a problem for you?” People exaggerate both pain and frequency in general terms.
Listen for actions, not opinions. Someone who says “that’s a great idea” is giving you nothing. Someone who says “I spent $3,000 trying to solve this with a consultant last year” is giving you signal.

Who to Interview and How to Recruit Them

Who to Target

Define your hypothetical customer persona as specifically as possible before recruiting:

Job title + industry + company size (for B2B)
Demographic + specific behavior pattern (for B2C)

Example (too broad): “Marketing managers” Example (specific enough): “Marketing managers at B2B SaaS companies with 10–100 employees who currently run paid search campaigns”

The more specific your recruit, the more useful the interview. Broad personas produce muddled, contradictory insights.

How to Recruit

Warm intros (conversion rate: 60–80%): Tell every person you know what you’re researching and ask if they know someone who fits. “I’m researching how [X type of person] deals with [Y problem]. Do you know anyone I should talk to?” Works faster and produces more candid conversations than any other method.

LinkedIn cold outreach (conversion rate: 5–15%): Search for exact job titles. Send a short, honest note — no pitch, no product mention:

Hi [Name], I’m doing research on how [their role] teams handle [specific workflow]. I’m not selling anything — I’m trying to understand the problem before building anything. Would you be open to a 30-minute call? Happy to share what I learn.

Online communities (conversion rate: 10–30%): Post in relevant Slack groups, subreddits, forums, or LinkedIn groups. Be specific about who you’re looking for and be transparent that it’s unpaid research.

Typeform/Screener: For B2C, run a screener survey to filter for the right profile. Ask 3–5 qualifying questions; offer the interview to those who qualify.

How Many to Interview

Interview at least 10 people before drawing conclusions. The first 5 interviews orient you — they surface the vocabulary, the workflow, and the landscape. Interviews 6–10 are where you start hearing repetition, which is where confidence comes from.

Stop when you’re hearing the same things repeatedly and new interviews aren’t adding new dimensions. For most problems, this happens between 15 and 25 interviews.

Structuring the Interview

A well-structured discovery interview runs 30–45 minutes and follows this arc:

Opening (5 minutes)

Set context and lower their guard:

“Thanks for making time. Before we start, a few things: this is not a sales call, I won’t be pitching anything. I’m here to understand your experience with [problem area]. There are no right or wrong answers — I’m genuinely trying to learn. Is it okay if I record this for my notes? The recording won’t be shared.”

Then: “Can you tell me a little about your role and what you work on day to day?”

This last question warms them up and gives you context for everything that follows.

Core Questions (25–30 minutes)

Move from general to specific. Sample structure:

Understand the workflow:

“Walk me through how you currently handle [the thing your product might do].”
“What does a typical week look like when it comes to [this area]?”

Identify frequency and recency:

“When did you last deal with this? What happened?”
“How often does this come up — roughly?”

Surface the pain:

“What’s the most frustrating part of the way you handle this today?”
“What have you tried to fix this? How did that go?”

Find existing solutions and their failure points:

“What tools or processes do you use for this right now?”
“What do you wish those tools did differently?”

Uncover the stakes:

“What happens if this doesn’t get solved? What’s the cost?”
“Has this ever caused a real problem — missed a deadline, lost a deal, cost you money?”

Closing (5 minutes)

“Is there anything about this area that I haven’t asked about that you think I should understand?”
“Who else do you think deals with this problem? Could you introduce me?”

Always ask for referrals at the close. One satisfied interviewee can unlock 3–5 more.

Question Types to Avoid

Hypothetical Questions

“Would you use a tool that did X?” — This tells you nothing. Humans are terrible at predicting their own future behavior. Replace with: “Have you ever tried to do X? What happened?”

Leading Questions

“Don’t you find it frustrating when Y?” — This suggests the answer. Replace with: “How do you feel about Y?”

Binary Questions

“Do you have this problem?” — Yes/no shuts down the conversation. Open questions open it: “Tell me about the last time you dealt with [Y].”

Opinion Questions About Your Idea

“What do you think of my idea?” — Do not share the idea during discovery. Once you pitch, they stop telling you the truth and start managing your feelings.

The 10 Discovery Questions Every Interview Should Include

Adapt these to your context, but these 10 cover the essential territory:

“Walk me through how you handle [the core workflow] today.”
“When did you last do this? Can you walk me through exactly what happened?”
“What’s the hardest part of this for you right now?”
“What have you tried to solve this? Why did those solutions fall short?”
“How much time does this take — per week, per month?”
“What does it cost you when this goes wrong?”
“Who else on your team deals with this?”
“What would a perfect solution look like — if you could wave a magic wand?”
“Have you ever paid for something to solve this? What?”
“Who else do you think I should talk to about this?”

How to Take Notes

Record With Permission

Always ask: “Is it okay if I record this call for my personal notes?” Most people say yes. Use Otter.ai, Fathom, or Fireflies for automated transcription. Review the transcript after — you miss things when you’re focused on listening.

Note Emotion, Not Just Words

Flag moments where the interviewee sighs, laughs, gets animated, or pauses. “They paused and looked uncomfortable when I asked about their current tool” is as useful as what they said.

Use a simple note template per interview:

Date:
Interviewee: [role, company size]
Biggest pain mentioned:
Quotes worth capturing:
Current solutions they use:
Willingness to pay signals:
Surprises / things I didn't expect:
Referrals offered:

Analyzing Patterns Across Interviews

After 10 interviews, run an affinity mapping session:

Write each distinct insight, quote, or observation on a separate sticky note (digital: use Miro or FigJam)
Group similar observations together
Name each cluster with a problem statement
Count how many interviewees contributed to each cluster

Clusters that appear in 7 out of 10 interviews are signal. Clusters from 2 out of 10 are noise. Your product should start by solving the 7-out-of-10 problems.

When You Have Enough Signal

You have enough signal when:

You can accurately predict what a new interviewee will say before they say it
Three or more problems appear repeatedly and you understand their root cause
You have identified who experiences the problem most acutely (your ICP)
At least three interviewees have said something equivalent to “let me know when you build this”

Unprompted expressions of intent (“when can I buy this?”) are the highest-value signal in discovery research.

From Interviews to Next Steps

Discovery interviews should produce three outputs:

A problem statement: “[Type of person] struggles with [specific problem] because [root cause]. Today they solve it by [current workaround], which costs them [time/money/frustration].”
An ICP hypothesis: “The person who feels this problem most acutely is [specific profile], and here’s why they feel it more than others.”
A prioritized assumption list: The beliefs your product depends on, ranked by importance and uncertainty. The most important, most uncertain assumption is what you test first.

Do not start building until you can write a clear problem statement. If you can’t, you need more interviews.

Key Takeaway

Customer discovery interviews are the cheapest research you can do, and the most commonly skipped. Ten 30-minute conversations will tell you more about whether your product has a market than six months of building. Recruit specifically, ask about the past not the future, take notes on emotion as well as content, and stop when you’re hearing the same things repeatedly. The goal is not to validate your idea — it is to understand the problem well enough that the right solution becomes obvious.