The Mom Test: Why Most User Interviews Fail

Reading time: 8 min · Level: Beginner · Author: PulseCheck Team

Most founders and product managers think they're doing user research. They schedule calls, ask questions, take notes. But here's the uncomfortable truth: 90% of user interviews are fundamentally broken.

The problem isn't effort—it's approach. And it all comes down to a simple book called "The Mom Test" by Rob Fitzpatrick.

What is The Mom Test?

The Mom Test is a set of rules for crafting questions that even your mom couldn't lie to you about. It's named after a simple insight: if you ask your mom "Do you think my business idea is good?", she'll say yes. Not because it is, but because she loves you.

The same thing happens in user interviews. People are polite. They want to help. They'll tell you what you want to hear.

The Mom Test flips this on its head with three core rules:

Talk about their life, not your idea
Ask about specifics in the past, not generics about the future
Talk less, listen more

Why Most User Interviews Fail

Mistake #1: Pitching Instead of Learning

Bad: "We're building an app that helps you track your expenses automatically. Would you use it?"

Good: "Walk me through the last time you tried to figure out where your money went last month."

The first question is a pitch disguised as research. You're asking them to validate your solution before understanding their problem. The second question uncovers real behavior.

Mistake #2: Asking About the Future

Humans are terrible at predicting their own behavior. "Would you pay for this?" is meaningless. "Have you paid for something similar?" is gold.

Bad: "Would you use an AI tool to conduct user interviews?"

Good: "How many user interviews did you do last quarter? What stopped you from doing more?"

Mistake #3: Accepting Compliments as Data

When someone says "That sounds really useful!", you've learned nothing. Compliments are social lubrication, not validation.

What to do instead: Push for commitment. "Would you be willing to try a beta next week?" or "Can you introduce me to two colleagues who have this problem?"

The Three Types of Bad Data

Rob Fitzpatrick identifies three types of bad data that pollute your research:

1. Compliments

"That's a great idea!" · "I'd definitely use that!" · "You should build this!"

Why it's bad: People are being polite, not truthful.

2. Fluff

Generics, hypotheticals, and future promises. "I usually..." · "I would never..." · "I always try to..."

Why it's bad: There's no specificity. You can't verify or act on it.

3. Ideas

"You know what you should add?" · "It would be cool if..." · "Have you thought about..."

Why it's bad: Feature requests without understanding the underlying problem lead to bloated products.

The Mom Test in Practice

Here's how to transform bad questions into good ones:

| Bad Question | Good Question | | --- | --- | | Do you think this is a good idea? | What are you currently doing to solve this? | | Would you pay $30 for this? | What's your budget for tools like this? What have you paid for? | | How often would you use this? | How often did you do X last month? Show me. | | What features would you want? | Walk me through the last time this was painful. | | Who else has this problem? | Who else should I talk to about this? |

Five Questions That Always Work

These questions pass The Mom Test and consistently uncover valuable insights:

"What's the hardest part about [doing thing]?"

Opens up pain points without leading.
"Tell me about the last time that happened."

Forces specificity and real stories.
"Why was that hard?"

Digs into root causes.
"What have you tried to fix it?"

Reveals if the problem is painful enough to act on.
"What don't you love about your current solution?"

Uncovers switching triggers.

How PulseCheck Applies The Mom Test

PulseCheck's AI interviewer is trained on Mom Test principles:

It never asks leading questions — The AI probes based on what users actually say, not what you hope they'll say.
It digs into specifics — When someone gives a vague answer, the AI follows up: "Can you give me a specific example?"
It ignores compliments — Positive statements don't increase pain point scores. Only concrete problems and behaviors matter.
It tracks past behavior — The AI is trained to ask "When did this last happen?" rather than "Would you...?"

Key Takeaways

Remember: The goal of user research isn't validation—it's learning. If you're not occasionally discovering that your assumptions were wrong, you're not doing it right.

Stop asking about your idea. Ask about their life.
Stop asking about the future. Ask about the past.
Stop accepting compliments. Push for commitments.
Stop talking. Listen more.
Stop collecting opinions. Collect facts and stories.