Frameworks · Offer Architecture

The Mom Test: SaaS questions that actually work

Ask about the past, not the future. Six swaps that turn a conversation full of encouragement into one full of facts.

The Mom Test is a rule for customer conversations: ask questions so specific about the past that even your mother could not mislead you with a kind answer. Most founder interviews fail because they ask about the future, and everyone is generous about hypotheticals.

The three rules
  • Talk about their life, not your idea
  • Ask about specifics in the past, not opinions about the future
  • Talk less, listen more

Any question that can be answered with encouragement rather than a fact is the wrong question.

The failure this prevents is specific and expensive: ten enthusiastic conversations, a product built on them, and nobody buying.

Why "would you use this?" fails

Three reasons, and they compound.

People are kind. Told you are building something, most will find something encouraging to say, especially if they like you.

People are bad at predicting their own behaviour. "I would definitely pay for that" is a sincere guess about a future self who has different priorities and a different week.

And you are leading. The moment you describe your idea, you have told them what answer would please you, and the conversation stops being research.

The questions to ask instead

Each of these is answerable only with a fact about something that already happened.

Instead ofAsk
Would you use a tool that does X?Walk me through the last time you had to do X.
Would you pay $50 a month for this?What are you spending on this problem today?
Do you think this is a good idea?What have you already tried to fix it?
How often does this happen?When did it last happen?
Would this be useful to your team?Who else deals with this, and what do they do?
What features would you want?What did you do the last time it went wrong?

The pattern in the right-hand column is that every question points at a specific past event. Nobody can be encouraging about last Tuesday.

Marcus · GhostCoach's AI coach
"I recommend not mentioning your product until the conversation is over. The moment they know what you are building, you stop collecting information and start collecting politeness — and politeness has cost founders years."

Reading the answers

Three responses that look like validation and are not.

"That sounds great." A compliment. Compliments are the enemy of a useful interview because they feel like progress and contain nothing. Redirect to a specific past event.

"I would definitely use that." A hypothetical commitment, which costs nothing. Convert it into a real one: a pre-order, a pilot date, an introduction to whoever holds the budget.

"Yes, we have that problem." Recognition, not urgency. Ask what they have done about it. If the answer is nothing, and it has been a problem for two years, it is not painful enough to fund a purchase.

The signal that matters is evidence of past spending — money, time, or a workaround someone built. Validating with no audience covers how to find people to ask in the first place.

Running the conversation

Twenty minutes is enough. Four beats, in this order.

  1. Context. What is your role, what does your week look like?
  2. The last occurrence. Walk me through the last time [problem] came up.
  3. What they did. What did you try? What did that cost in time or money?
  4. The commitment ask. Only at the end, and only if the first three went well.

Do not take notes on what you hoped to hear. Write down the exact words they use for the problem — that vocabulary becomes your headline copy, and it is the single most reusable output of the whole exercise.

If a conversation felt great and you cannot name one specific thing you learned, it was a bad interview. Enthusiasm is the most common way a founder walks away from twenty minutes with nothing.

After the interviews

Five to eight conversations is usually enough for patterns to repeat. Look for the same trigger event appearing across people, and for consistent language about the problem.

Then convert the pattern into a job statement — Jobs to be Done for SaaS covers the format. And run the result through the validation checklist before committing to build anything.

For the interview mechanics in more depth, including how to recruit and how to handle a conversation that goes off track, see the customer interview guide.

The Mom Test for SaaS FAQ

What is the Mom Test?

A rule for customer conversations: ask questions so specific about the past that even someone who loves you could not mislead you with a kind answer. The three rules are to talk about their life rather than your idea, ask about specifics in the past rather than opinions about the future, and listen more than you talk.

What questions should I ask in a SaaS customer interview?

Questions that point at a specific past event. 'Walk me through the last time you had to do this' rather than 'would you use a tool that does this'. 'What are you spending on this today' rather than 'would you pay $50 a month'. Nobody can be encouraging about last Tuesday.

Why should I not ask if someone would use my product?

Because people are kind, they are poor at predicting their own behaviour, and describing your idea tells them which answer would please you. The result is sincere encouragement that carries no information about whether they will buy.

How many customer interviews do I need?

Five to eight is usually enough for patterns to repeat. You are looking for the same trigger event appearing across people and consistent language about the problem, both of which show up quickly if they exist at all.

What counts as a real signal in a customer interview?

Evidence of past spending — money, time, or a workaround someone actually built. Compliments, hypothetical commitments and simple recognition of the problem all feel like validation and are not. Convert enthusiasm into a pre-order, a pilot date or an introduction.

Turn your interview notes into a decision

Tell Marcus what people actually said. You get a read on whether the signal is real and what to test next.

Try GhostCoach free →

14-day free trial · cancel anytime · 30-day money-back on Lifetime