The Mom Test: Interviewing for Truth

Why Most Customer Research Is Worthless

Here is an uncomfortable truth that will save you years of wasted effort: most of what founders call ‘customer research’ is closer to therapy than science. You describe your beautiful idea, your friend nods enthusiastically, you walk out the door with a glow of validation, and six months later you discover that nobody — including the person who said ‘I love it’ — will actually pay for it.

The problem isn’t that people are malicious. It’s that they’re polite. Humans are wired to preserve social harmony, and when you stand in front of someone with obvious enthusiasm and ask ‘Do you think this is a good idea?’, the cost of saying ‘No, this is terrible’ is far higher than the cost of saying ‘Sounds interesting!’. So they smile, they encourage, and they lie. Not to hurt you — to spare you.

This is the trap that Rob Fitzpatrick’s book The Mom Test exists to break. The title comes from a simple observation: even your own mother — the person who loves you most in the world — will lie to you about your business idea. So if your validation strategy can’t survive a conversation with your mum, it isn’t a validation strategy at all. It’s confirmation bias dressed up in a notebook.

The Mom Test is a methodology for extracting truth from prospective customers without ever giving them the chance to lie to you politely. You do this by talking about their life, not your idea. You ask about the past, not the future. You dig into actual behaviour, actual spending, and actual workarounds. And you treat every compliment as evidence of nothing.

In this lesson, you’ll learn the specific question patterns that work, the ones that don’t, and how to leave an interview with real evidence rather than warm feelings. By the end, you’ll be able to run a 20-minute conversation that generates more useful data than a hundred surveys.

The Three Rules of a Mom Test Conversation

Fitzpatrick distils the methodology into three rules that, if followed, transform a flattering chat into genuine research:

1. Talk about their life, not your idea. The moment you describe what you’re building, the conversation becomes about you. Their job changes from ‘tell me about your world’ to ‘evaluate this stranger’s pitch politely’.
2. Ask about specifics in the past, not generics or opinions about the future. ‘Would you’ questions invite fantasy. ‘Did you’ questions force memory.
3. Talk less and listen more. The interview is not your moment to shine. If you’re talking more than 20% of the time, you’re selling, not learning.

Notice the pattern: every rule is designed to keep your idea out of the room. Counter-intuitive, isn’t it? You came to validate your idea, and the first instruction is: don’t mention it. But that’s precisely the discipline. Your idea is the bias-generator. Remove it, and you can actually hear what’s true.

Good Questions vs Bad Questions

Let’s get concrete. The difference between a useful interview and a useless one usually comes down to question construction. Here are the patterns that work — and the ones that lie to you.

Bad questions (avoid these)

• ‘Would you use a tool that…?’ — Hypothetical. People are wildly optimistic about future behaviour.
• ‘Do you think this is a good idea?’ — You’re asking for a compliment. You will receive one.
• ‘How much would you pay for something like this?’ — Pure fantasy pricing. Nobody knows until their card is out.
• ‘Would you pay £X per month for it?’ — Yes-or-no future commitments are almost always answered ‘yes’ to be polite.
• ‘What features would you want?’ — Turns the customer into a product manager, which they aren’t. You’ll get a shopping list nobody will pay for.

Good questions (use these)

• ‘Walk me through the last time this happened.’ — Forces concrete memory. The answer reveals real behaviour, real context, real frustration.
• ‘What do you currently spend to deal with it?’ — Money already moving is the single best signal of pain.
• ‘What have you tried so far?’ — Reveals existing workarounds. If the answer is ‘nothing’, the problem isn’t painful enough.
• ‘Why is that hard for you?’ — Digs past surface complaints into the underlying mechanism.
• ‘What happened next?’ / ‘And then what did you do?’ — Keeps the story going. Stories are where truth hides.
• ‘Who else is affected by this?’ — Maps the wider system and uncovers economic buyers.
• ‘Last time it happened, what did it end up costing you — in money, time, or stress?’ — Quantifies the pain in their own terms.

Notice the structural difference. The bad questions ask for predictions, opinions and reactions to your idea. The good ones ask for stories about their life. Stories contain evidence. Predictions contain hope.

Digging for Evidence, Not Validation

The shift that takes founders from amateur to expert interviewer is learning to chase specifics relentlessly. When someone says ‘Yeah, that’s annoying’, you have nothing. When they say ‘Last Tuesday I spent three hours rebuilding a spreadsheet because the export from our CRM was broken, and I had to cancel a client call to do it’, you have a goldmine.

Train yourself to follow up every vague answer with one of three drills:

1. The time drill: ‘When was the last time that happened? Walk me through it.’
2. The money drill: ‘What did that end up costing you? Are you paying for anything to fix it today?’
3. The consequence drill: ‘What happened because of that? Who noticed? What did you do next?’

These three drills convert opinions into observations. If someone can’t answer them — ‘Oh, I can’t remember the last time…’ — that itself is data. It means the problem isn’t frequent or painful enough to be a business.

The Currencies of Commitment

The deepest principle in the Mom Test is that only commitment counts as validation. Compliments cost nothing, so they prove nothing. What costs something? Three currencies:

• Money — a deposit, a pre-order, a paid pilot, a signed contract. The gold standard.
• Time — agreeing to a follow-up meeting at a specific date, completing a substantial homework task, introducing you to their boss or three colleagues.
• Reputation — a public endorsement, an introduction to someone they respect, putting their name on a waiting list visible to others.

If a meeting ends with no commitment of money, time or reputation, it ended in a polite no — even if the words were enthusiastic. Get used to ending interviews with what Fitzpatrick calls an advance: a concrete next step that costs the other person something. ‘Great — can I send you a £200 paid pilot proposal this afternoon?’ tells you more in their reaction than the previous 45 minutes of conversation combined.

Running the Interview: A Practical Structure

A good Mom Test interview lasts 20-30 minutes and follows a loose arc. Don’t script it word-for-word — you’ll sound like a robot and miss the gold that emerges from genuine curiosity. But do hold this shape in your head:

1. Warm-up (2 minutes): Thank them, frame the conversation honestly: ‘I’m researching how people handle [problem area]. I’m not selling anything — I just want to learn how it actually works in your world.’ This explicit framing gives them permission to be candid.
2. Context (5 minutes): Learn their world. What’s their role, their day, the tools they use, the people they work with? You’re building a mental model of where the problem lives.
3. The last incident (10 minutes): ‘Walk me through the last time [problem] happened.’ Then shut up. Ask only follow-ups: ‘What happened next?’, ‘Why was that hard?’, ‘What did you do then?’. This is the heart of the interview.
4. Workarounds and spend (5 minutes): ‘What have you tried to fix this? What are you spending on it today — money, tools, your own time, other people’s time?’
5. The advance (3 minutes): Ask for a real commitment. An introduction to two colleagues with the same problem. A follow-up call in two weeks where you show them a prototype. A small deposit toward a paid pilot. Watch their reaction more than their words.

What to Write Down

Don’t transcribe the conversation — you’ll miss the cues. Instead, capture three things in your notes:

• Quotes verbatim when they say something painful, specific or surprising. Direct quotes are persuasive evidence for your own decision-making later.
• Numbers and timeframes — ‘3 hours’, ‘£400/month’, ‘happens weekly’. These are the bones of your business case.
• Signals and anti-signals — did they lean in? Did they bring up a problem you didn’t prompt? Did they immediately offer to introduce you to someone? Or did they keep checking their phone?

Within an hour of the interview, write a one-paragraph summary while it’s fresh. After ten interviews, you’ll start to see patterns that no single conversation revealed — and that pattern recognition is the real output of this process.

Finding the Right People to Interview

The methodology only works if you’re talking to actual prospective customers. Five interviews with the wrong people are worse than zero, because they generate false confidence. So how do you find the right 15 strangers?

• LinkedIn cold outreach: Search for the exact role you’re targeting. Send a short, honest message: ‘I’m researching how [role] handle [problem]. Not selling anything — would you give me 20 minutes? Happy to share what I learn from the wider research.’ A 5-10% response rate is normal.
• Relevant communities: Reddit subs, Slack groups, Facebook groups, industry forums. Post a respectful research request, or DM active members.
• Existing networks at one remove: Don’t interview your friend — ask your friend to introduce you to three of their colleagues who fit the profile.
• Event sniping: Industry meetups and conferences are dense with target users. A 10-minute chat over coffee at a conference is often more honest than a scheduled Zoom.
• Paying for time: For hard-to-reach roles (specialist clinicians, senior execs), offering £50-£100 for 30 minutes via a platform like Respondent.io is faster and produces more candid conversations than you’d expect.

Aim for 15 interviews minimum before drawing conclusions. Fewer than that and you’re pattern-matching on noise. Around the 10th conversation you’ll start hearing the same phrases come up unprompted — that’s the signal that you’ve found a real pattern, not just one person’s quirky take.

A Worked Example: The Meal-Prep App

Let’s ground this in the scenario from our previous lesson. You’re a fitness coach considering a £39/month done-for-you meal-prep app. Forty friends have said ‘great idea’. What would a Mom Test interview with a busy professional actually sound like?

Bad version (the founder is selling, not learning):

Founder: So I’m thinking of building an app that does your meal prep for you — sends you a weekly plan, shopping list, the lot. Would you use that?
Prospect: Yeah, sounds great actually. I’m terrible at meal prep.
Founder: Brilliant! Would you pay £39 a month for it?
Prospect: Yeah, I reckon. Sounds reasonable.

You walk out elated. They will never pay you. You have learned nothing except that this person is polite.

Good version (the founder is learning, not selling):

Founder: Walk me through what you ate yesterday — breakfast through dinner.
Prospect: Honestly? Skipped breakfast, grabbed a Pret sandwich at lunch, got a Deliveroo curry at like 9pm because I was knackered.
Founder: Is that pretty typical?
Prospect: Probably four nights a week, yeah.
Founder: What did you spend on food last week roughly?
Prospect: God… £30 on a Tesco shop I half-used, maybe £80 on Deliveroo, £40 on lunches out. So like £150.
Founder: Have you ever tried to fix this?
Prospect: I bought one of those HelloFresh boxes for two months. Cancelled it because the cooking still took 40 minutes and I was too tired.
Founder: What made you cancel exactly?
Prospect: The cooking. I want food to appear, not to chop onions at 9pm.

Now you’ve learned something real: this person spends £150/week on food, has paid for a solution before, and the actual job-to-be-done is ‘food appearing’, not ‘meal planning’. Your £39/month plan-and-shopping-list product probably won’t solve their problem — they need prepared food delivered. That’s a completely different business. One good interview just saved you £8,000 and six months.