Mom Test Questions: The Complete Question Bank

You have a conversation scheduled with a potential customer. You know you should ask good questions. And then the call starts and you realize you do not actually know what to say.

Rob Fitzpatrick’s The Mom Test defines what makes customer questions useful: talk about their life, not your idea; ask about the past, not the future; and listen more than you talk. Those three rules are the foundation. But rules are not a script.

This is the script.

Below is a complete question bank for a Mom Test customer conversation, organized by phase: opening, problem exploration, behavior, pain intensity, and commitment signals. Each section includes example questions, the logic behind them, and what to listen for in the answers.

If you want the full framework first, the Mom Test summary and framework explanation covers the three rules and the reasoning behind them. This article is for when you understand the principles and are ready to run an actual conversation.

Two things to cover before the question bank: what separates questions that generate real signal from questions that generate polite noise, and which question patterns to remove from your script entirely.

Understanding both will make every question in the bank below land differently. According to CB Insights’ analysis of startup and product post-mortems, no market need is the leading cause cited in product failures — ahead of running out of money, team problems, or bad timing. These conversations are how you find out whether that market need exists before you build anything.

What Makes a Mom Test Question Work?

Two characteristics separate questions that generate real information from questions that generate polite noise.

Past versus future orientation. “How do you currently handle X?” asks someone to report a fact. “Would you use a tool that does X?” asks them to predict their own behavior. People systematically overestimate how likely they are to adopt new tools, pay for solutions, or change habits around problems they currently tolerate. Past behavior is a fact. Future behavior is optimism. The questions that work ask for the facts.

Your idea stays invisible. If the other person can infer what answer you are hoping for, they will tend to give it to you — not because they are dishonest, but because they are polite. Questions that make your product concept visible invite social support instead of honest evaluation.

The practical test: can you ask this question without mentioning your idea at all? If the answer is no, the framing is probably pointing them toward a “yes.”

Questions to Remove From Your Script

Most founder conversations are full of patterns that produce false positives. Recognizing them matters as much as knowing the right questions.

Every question in this table shares one characteristic: it signals what answer you want and invites the other person to tell it back to you. Delete them before your next conversation.

The Core Question Bank by Phase

A structured Mom Test conversation moves through roughly five phases. Adapt based on what you hear — not every question gets asked in every conversation.

Phase 1: Opening Questions

These come before you mention your idea or any specific problem. They establish context: who you are talking to, how they spend their time, and what their current workflow looks like.

• “Can you walk me through a typical week in your [role / business]?”
• “What does [relevant activity] look like for you right now?”
• “What tools or processes are you currently using for [area you care about]?”
• “What are you spending the most time on these days?”
• “Where do most of your [problems / clients / tasks] tend to come from?”

What to listen for: the gap between what they say is a priority and where they actually spend time. Stated priorities and actual behavior diverge constantly, and that gap is often where the real constraint lives.

Phase 2: Problem Exploration Questions

These establish whether the problem you are investigating shows up in their life — without you naming it first.

• “What is the most time-consuming part of [relevant workflow]?”
• “Where do things tend to break down or slow down for you?”
• “What do you find yourself doing manually that feels like it should not take this long?”
• “What is the thing you are least satisfied with in how you currently handle [area]?”
• “Is there anything you have been meaning to fix or improve and just have not gotten to?”

What to listen for: whether they name your problem unprompted. If you have to narrow in through several questions before they arrive at it, note that. They may have the problem but consider it low priority.

Phase 3: Behavior Questions

Past behavior is the most predictive signal in a customer conversation. These questions surface what someone has actually done — not what they think they would do.

• “When did this last come up for you, and what did you do about it?”
• “Have you looked for a solution to this before? What did you try?”
• “Is this something you have worked around, or do you just deal with it?”
• “How long have you been handling it this way?”
• “Have you ever paid for anything to help with this — even partially?”

What to listen for: active problem-solving behavior. Fitzpatrick’s central point in The Mom Test is that the best signal is someone already trying to solve the problem, however imperfectly. A person who has paid for a workaround, built a spreadsheet to manage it, or hired someone to deal with it part-time is already spending resources on the problem. Someone who says “that would be great” but has never attempted to fix it is telling you the problem is tolerable, not acute.

Phase 4: Pain Intensity Questions

These calibrate whether the problem is severe enough to warrant a paid solution.

• “How much time does this add up to in a week?”
• “Has this ever cost you money, a client, or a missed deadline?”
• “If you could fix only one thing about this process, what would it be?”
• “How often does this come up?”
• “If this were solved tomorrow, what would actually change for you?”

What to listen for: specificity. Vague answers (“it is kind of annoying”) suggest low intensity. Specific answers (“I lose half my Friday dealing with this”) suggest a real constraint. The more specific and unprompted the description, the more the person has actually confronted the problem as a genuine obstacle — not just an inconvenience they have learned to live with.

Phase 5: Commitment Signal Questions

These are not “would you pay” questions. They reveal what the person is already willing to do — which is far more reliable than stated future intentions.

• “Is this something you are actively trying to solve right now?”
• “Have you looked at or budgeted for any paid tools in this space?”
• “Would you be open to a follow-up conversation once I have something to show you?”
• “Do you know anyone else I should talk to about this?”
• “If I put together a simple version of this, would you be willing to try it before I go further?”

What to listen for: movement versus passive interest. “Sure, I’d check it out” costs the other person nothing. “I want to be on the list for early access” — especially when paired with behavioral evidence from earlier in the conversation — is a different category of signal.

Finished your first conversations and trying to figure out what the signal actually means? The Idea Validation Scorecard turns your conversation findings into a structured go/wait/kill recommendation. Free. About 15 minutes.

How to Read the Answers

The answers themselves are half the information. What is absent from the answers matters equally.

Unsolicited mentions carry more weight than prompted ones. If you ask an open question about their workflow and they immediately name the exact problem you are investigating — before you bring it up — that is disproportionately strong evidence. If you have to steer through four phases to get them there, they may have the problem but not consider it a priority worth solving.

Emotional intensity is more informative than frequency. Someone who describes a problem as genuinely frustrating, who becomes animated when discussing it, who says “this drives me crazy every single week” — this is often a better signal than someone who encounters the problem daily but describes it with resignation. Frustrated urgency is more likely to become a paying customer than resigned tolerance.

Look at what they have already tried. A person who has tried two tools, built a spreadsheet workaround, and hired a freelancer to deal with the problem is already spending resources on it. Your job is to figure out whether you can solve it better than their current workaround — and at what cost threshold they would switch.

Compliments are not data. When a potential customer says “that sounds really interesting” mid-conversation, they are typically being polite, not confirming demand. Fitzpatrick is direct about this: compliments are the conversational equivalent of your mom saying your idea is great. They feel encouraging. They are not data. What you want are behavioral evidence and specific past facts — not encouragement.

Adapting Questions by Product Type

The question bank above applies broadly, but the framing shifts depending on what you are evaluating.

For SaaS and software tools: Focus phases 2 and 3 on current tool usage. “What tools do you currently use for this?” and “What made you choose that tool?” reveal existing alternatives and switching costs. “What are you paying for it right now?” establishes a price reference point without requiring you to ask directly about willingness to pay for your product.

For digital products (courses, templates, playbooks): Focus phase 3 on self-directed learning behavior. “Have you ever paid for a course or resource to learn this?” is more diagnostic than “would you buy a course on this.” What someone has already spent on education in this area predicts future spending more reliably than what they say they might spend.

For services and coaching: Focus phases 4 and 5 on outcome specificity. Buyers of services purchase outcomes, not deliverables. “What would success look like to you, specifically?” and “What would you need to see in the first 30 days to know this was working?” reveal whether a potential client has clear enough expectations to become a satisfied customer — or whether the engagement will be perpetually underdefined.

The customer interview scripts article extends this question bank into full conversation outlines for each product type, including how to handle the most common paths a conversation takes.

Frequently Asked Questions

How many questions should I ask in one Mom Test conversation?

Aim for roughly 8 to 12 questions per conversation, not counting natural follow-ups. Fewer and you will not have enough to identify clear patterns. More than 15 and the conversation shifts from discussion to interrogation. Most useful conversations run 20 to 30 minutes. Depth on each question matters more than breadth across questions — one detailed answer about past behavior tells you more than five vague opinions about what might happen.

What if someone says they would definitely pay for this?

Treat it as weak signal. Stating a purchase intent during a conversation costs nothing — the threshold for saying “I’d buy that” is zero. What you want is behavioral evidence: whether the person has already tried to solve the problem, spent money on related solutions, or taken meaningful action to deal with it. A stated price point with no behavioral evidence behind it should be noted and weighted low. The Mom Test addresses this directly: commitments that cost the other person nothing are not commitments.

What if the person I am talking to does not have the problem I am solving?

That is useful information. Thank them, ask if they know anyone who does face this problem, and log their profile as outside your target customer. If this happens in several consecutive conversations — you keep finding people who either lack the problem or do not consider it significant — treat that pattern as a signal worth investigating before you build anything. Most products that fail to find an audience share a common root: the founder assumed demand existed and never discovered through early conversations that the people they were talking to did not actually face the problem acutely. The common mistakes article covers this pattern in detail.

How many conversations do I need before the signal is reliable?

There is no universal number, but a commonly cited threshold in lean startup practice is roughly 5 to 10 conversations where you start hearing consistent, unprompted repetition of the same problems, the same language, and the same past behaviors. Convergent signal across independent conversations — where you did not lead people to the answer — is more meaningful than one enthusiastic response. If you are still hearing widely varied problems and language after 10 conversations, your target customer definition may be too broad. The lean canvas framework covers how to tighten that definition before continuing.

Start With the Scorecard

Customer conversations give you raw material. The Idea Validation Scorecard turns that material into a structured decision.

It scores your idea across ten evaluation criteria — including the quality of signals from customer conversations — and produces a go/wait/kill output you can act on.

Run the Idea Validation Scorecard →

Free. About 15 minutes. No email required.

Keep Reading

• The Mom Test Summary: What the Book Gets Right (And How to Use It)
• Customer Interview Scripts: What to Ask and What to Ignore
• Why Solo Products Fail: 5 Core Patterns