The Mom Test: How to Talk to Customers Without Fooling Yourself

You describe your idea to a room of friendly people. Their faces light up. “Wow, that sounds great,” they say. “I’d totally use that.” You walk away glowing. Months later you launch, and nobody buys.

You weren’t lied to by liars. You were lied to by polite people who didn’t even know they were doing it. The fix is a small, famous book by Rob Fitzpatrick called The Mom Test, and it can save you from building something nobody wants.

Why this matters

Every founder is told the same thing: “Talk to customers first.” So they do. Then they hear nice things, feel validated, and build the wrong product anyway.

The problem isn’t that you skipped customer conversations. It’s that you ran them in a way that guaranteed flattering, useless answers. Bad conversations don’t just waste time. They actively mislead you, because false hope feels exactly like real validation.

A customer conversation is just a talk where you try to understand someone’s problems before you’ve built or sold anything. Get good at it and you stop guessing. You learn what people actually do, what it costs them, and whether your idea solves a problem they already care about. That’s the difference between launching to silence and launching to a line of people waiting to pay.

The reason everyone lies to you

The book is named after a simple truth. If you ask your mom, “Do you think my business idea is good?” she will say yes. She loves you. She doesn’t want to hurt your feelings. She has no clue whether it’s actually good, but she’ll cheer anyway.

Here’s the uncomfortable part: everyone does this, not just your mom. Strangers do it too. When you pitch, people sense you’re emotionally invested, and the kind thing to do is encourage you. So they say “I’d use that,” and then never do.

Think of it like asking, “Do I look nice today?” Almost nobody answers “No, you look terrible.” The polite default is yes, so a yes tells you nothing.

Fitzpatrick’s insight is hidden in the title. A good set of questions should pass the mom test, meaning your questions are so well-built that even your own mom couldn’t lie to you, because you never ask her opinion of your idea. You only ask about facts from her life.

The three rules

The whole book boils down to three rules. Write them on a sticky note before every call.

1. Talk about their life, not your idea

The moment you describe your product, you’ve poisoned the conversation. Now they’re reacting to you and being polite to you. Keep the spotlight on their world: their problems, their day, their frustrations.

2. Ask about the past, not the future

People are wildly over-optimistic about the future. “Would you use this?” gets a hopeful guess. “What did you do the last time this happened?” gets a fact. The past actually happened. The future is a daydream.

3. Talk less, listen more

The more you talk, the more you steer them and the less you learn. Your job is to be curious and quiet. A good gauge: they should talk far more than you do.

Good data versus bad data

“Data” here just means the things people tell you. Some of it is gold and some is garbage, and they sound nearly identical in the moment.

Bad data feels good but means nothing:

• Compliments. “That’s a great idea!” Fitzpatrick calls these “the fluff.” Free candy: sweet, zero nutrition.
• Opinions and hypotheticals. Anything starting with “I would,” “I could,” or “I might.” A guess about an imaginary future.
• Promises about the future. “I’d definitely buy that.” “Send it when it’s ready.” A promise costs nothing to make, so it’s worth nothing.
• Wishlists. “It should also do X and Y.” People are bad at designing solutions. Treat feature requests as clues to a problem, not orders.

Good data is the truth, even when it’s quiet:

• Facts about the past. What actually happened, step by step.
• Past behavior. What they’ve already tried, used, or built to cope.
• Money already spent. Tools they pay for, people they’ve hired, workarounds they bought.
• Time already spent. Hours wasted, hacks built in spreadsheets, late nights.

Stories are facts. Opinions are guesses. If someone has already spent real time or real money fighting a problem, that problem is real. If they only say they care, you’ve learned nothing yet.

Questions you can steal word for word

Notice the pattern below. Bad questions ask for a verdict on your idea. Good questions ask for a true story from someone’s past. You can’t fake “the last time it happened.” It either happened or it didn’t.

Bad question (gets lies)	Good question (gets truth)
“Do you think this is a good idea?"	"Talk me through the last time you ran into this."
"Would you buy a product that did X?"	"How are you dealing with this problem today?"
"Would you pay $50/month for this?"	"What have you already tried? What did it cost you?"
"Do you like this feature?"	"What happens if you don’t solve this? Why does it matter?"
"How often do you have this problem?"	"When did it last happen? Walk me through that day.”

Three questions will carry most of your calls:

1. “Talk me through the last time that happened.”
2. “How do you deal with it now?”
3. “What have you tried, and how much did it cost you in time or money?”

Lead with these, then just keep asking “why?” and “tell me more.”

Common misconceptions

“I’d definitely buy that!” means I’ve validated my idea. No. This is the most dangerous sentence in customer research because it feels exactly like a win. Saying yes cost them nothing: no money, no commitment, no risk. It’s pure hot air. The only thing that proves real interest is a commitment, which means giving up something of value right now. That could be money (a deposit or pre-order), time (a real next meeting with a goal), reputation (introducing you to their boss), or specific personal data. No commitment means it was just a compliment in disguise.

A compliment is a signal. A compliment is the absence of a signal. When one lands, don’t bask in it. Bounce off it and dig.

Good interviewers are smooth talkers. The opposite. The best customer conversations come from people who stay curious and quiet. If you dread selling, this is freeing. You don’t have to charm anyone. You just have to listen well.

Feature requests tell me what to build. They tell you where a problem might be hiding. Ask what the feature would let them do that they can’t do today, then chase that.

How to use this

1. Before the call, write the three rules on a sticky note. Their life not your idea, past not future, listen more than you talk. Glance at it when you feel the urge to pitch.
2. Open with their current behavior. “How do you handle this today?” gets a factual answer and sets an honest tone.
3. Chase the last specific instance. “When did it last happen? Walk me through that day.” Specifics can’t be faked.
4. Follow the money and the time. Ask what their current workaround costs in dollars and hours. Real spending marks a real problem.
5. Deflect every compliment by digging. “Thanks! Is this something you actually run into? When did it last bite you?” Turn fluff into a story.
6. Catch yourself pitching. If you’re talking more than they are, saying “so basically my product…”, or feeling the urge to defend your idea, stop and ask another question about their life.
7. Ask for a commitment to test for truth. A deposit, a scheduled follow-up with a clear goal, or an introduction. Watch what they do, not what they say.
8. After the call, write down only facts and commitments. If your notes are all “they loved it,” you ran a bad call. If they have real stories, costs, and a concrete next step, you ran a great one.

See the difference in one conversation

Same founder, two versions.

The bad version (pitching):

You: “I’m building an app that auto-schedules your print jobs and saves tons of time. Pretty cool, right?” Them: “Oh wow, that sounds really useful!” You: “Would you use something like that?” Them: “Definitely, send it when it’s ready!”

You leave thrilled and learned nothing. Three compliments, one hypothetical, one empty promise.

The good version (Mom Test style):

You: “How do you handle scheduling your print jobs right now?” Them: “Honestly? A whiteboard and a messy spreadsheet.” You: “Talk me through the last time that went wrong.” Them: “Last Tuesday. Two big orders collided, we missed a deadline, and I refunded a client $400 and apologized for an hour.” You: “How often does that happen?” Them: “Couple times a month. I’ve even paid a freelancer to babysit the schedule, about $600 a month, and it still slips.”

No pitch. No compliments. You learned the problem is real, painful, frequent, and already costing $600 a month plus refunds. That is a business signal.

Conclusion

Here’s the one thing to carry out the door: stop asking people what they think of your idea, and start asking what they’ve already done about their problem. The truth lives in actions and spending, never in applause.

Now notice something. The good conversation above ended with the customer already paying $600 a month for a worse solution. That’s not just validation. That’s a price signal sitting in plain sight. The next question becomes how you turn a painful, expensive problem into an offer someone commits to before you’ve written a single line of code, which is exactly where customer development stops being research and starts becoming sales.