Back to the on-screen lesson ·
Questions about a specific past occasion produce checkable evidence; hypotheticals, leading questions and fishing for praise produce polite answers, and only a commitment that costs something says anything about the future.
Paper packet. Every task here also exists on screen, where it is checked automatically; answers written on paper are not assessed by Nydus. When you are back at a device, enter your answers there.
You will choose questions about a specific past choice rather than opinions or hypotheticals, put an interview in an order that earns the customer's story, and separate the facts in the notes from compliments, promises and real commitments.
Observation tells you what customers did. It rarely tells you what they were trying to do, what else they tried, or what it cost them. For that you have to ask — and asking well is harder than it looks, because people are kind to someone with an idea, and kindness sounds exactly like demand.
| Term | What it means |
|---|---|
| Past-behavior question | A question about a specific, real occasion that already happened. |
| Hypothetical | A question about an imagined occasion: would you, if. |
| Leading question | A question that contains the answer the asker hopes for. |
| Compliment | Praise for the idea, which weighs nothing as evidence. |
| Promise | A statement about future behavior that costs nothing to make. |
| Commitment | Giving something up now: money, time, a signature, an introduction. |
People are generous to someone who is excited about their idea. Ask 'would you use a service like ours?' and most will say yes, and mean it kindly. Ask 'how much would you pay?' and they will name a figure they have never had to hand over. Neither answer can be checked, and both are poor guides to what the same people will do when money is on the table.
The question that produces evidence is about a specific, recent occasion: 'Tell me about the last time you had your home cleaned by someone else. How did you find them? What did you pay? What went wrong?' The answer is a story, and a story has facts in it — where they looked, who they asked, what they paid, what annoyed them — that do not depend on being polite to the owner.
Three kinds of question to drop: the hypothetical ('would you…'), the leading ('don't you think…'), and the request for praise ('do you like my…'). All three are easy to answer kindly and impossible to check.
Another way: table
Rewriting three weak questions.
| Weak question | Rewritten about the past |
|---|---|
| Would you use a cleaning service? | Who last cleaned your home, and how did you find them? |
| How much would you pay for lunch? | What did the last team lunch cost, and who approved it? |
| Do you like my glaze? | When you last replaced cups, what did you buy and why? |
Another way: steps
An interview works in a fixed order. Set the context — you are trying to understand something, not selling. Get the story of the last time. Ask what they tried and what it cost, in money and in time. Leave anything about your own offer until the end, if at all: the moment you describe it, the rest of the conversation is about being nice to you.
Afterwards, sort your notes into three piles. Facts about what they did. Compliments ('that sounds amazing'), which are pleasant and weigh nothing. Promises about the future ('I'd order every week'), which cost nothing to make. The one kind of future statement worth recording is a commitment: a deposit paid, a trial booked, an introduction made.
Write down what you need to learn. Not 'will they buy?' but 'how do offices arrange lunch now, and what goes wrong?'
Turn each need into a past-behavior question. 'Walk me through the last time…' is the most useful opening in the method.
Follow the money and the time. What did it cost? How long did it take? What did they do when it went wrong? Costs show how much the problem matters to them.
Ask for a commitment, not a prediction. A deposit, a trial date, an introduction to a colleague.
Sort and count. Facts, compliments, promises, commitments; then work out how many promises became commitments.
Check the answer. If most of your notes are compliments and promises, the questions were weak, whatever the people said. A commitment rate can never be higher than a hundred percent of the promises. And a fact should be checkable: if the person said they ordered a platter three times last month, the receipt would say so.
What someone did is fixed and specific. What they say they will do is a prediction about themselves, made in front of a person they would like to please, and it costs them nothing if it turns out wrong. Researchers who compare stated intentions with later purchases find again and again that intentions run far ahead of behavior.
A commitment changes that, because it has a cost. Paying fifty dollars, booking an afternoon, introducing a colleague: each is a small piece of behavior, and behavior is what predicts behavior.
Maya interviewed café owners about tableware. Twenty said they would definitely buy her cups. When she offered a trial set for a 50-dollar deposit, five paid: a quarter of the promises. That quarter is the first real measure of demand she has, and it is a useful one. It tells her to plan for roughly one buyer for every four enthusiastic conversations, not for twenty sales.
The arithmetic is a division and a multiplication — commitments over promises, times a hundred — but the habit behind it is the important part: always ask for something small and real before counting anyone as a customer.
Interviewers steer without meaning to. Nodding at the answers you hoped for, filling silences, finishing people's sentences: each teaches the person what you want to hear. Ask, then wait. Follow up on what surprised you, not on what confirmed your idea. And write down the person's words, not your summary of them: 'we just grab sandwiches, honestly it's fine' is very different evidence from 'unhappy with current lunch options'.
Interview the people whose situation the business serves, not friends and family who love the owner. For Bright Home Cleaning that means tenants who moved out in the last few months, found through the property managers who referred them, not neighbors who have never hired a cleaner.
Small numbers go a long way. Five to ten conversations with people in one segment usually bring the same few stories back again and again: the same worries, the same alternatives, the same moment of decision. When the last two or three interviews tell you nothing new, stop and count what you heard. More interviews in the same segment rarely change the picture; interviews in a second segment often do.
Write during the conversation, briefly, and fill in the notes straight afterwards while the words are fresh. Keep the customer's exact phrases in quotation marks: 'honestly, we just grab sandwiches' is evidence, and your summary of it is an interpretation. Put each note on its own line so it can be sorted later into facts, compliments, promises and commitments.
If you want to record audio, ask first and explain what the recording is for. Several states require the consent of everyone on a call before it is recorded, so the simple rule is to ask every time. Most people agree to a short conversation being noted; fewer are comfortable being recorded, and the notes are usually enough.
After a round of interviews, write a one-page summary before deciding anything. List the facts that came up more than once, with the number of people who said each. List every commitment, with what was given: a deposit, a booked trial, an introduction. Leave the compliments off the page entirely.
Then ask what the summary supports. Eight of ten office managers ordering supermarket platters, and three of ten booking a paid trial lunch, supports a test offer for that segment. Ten people saying 'what a lovely idea' and nobody booking anything supports nothing, however good the conversations felt.
Keep the summaries. Six months later, when the business is deciding whether to add a service, the facts and commitments recorded then are evidence; the memory of how enthusiastic everyone seemed is not.
An interview tells you what people did and what worried them; it cannot tell you for certain what they will do next. That is the job of a small test, the subject of lesson 10: an offer put in front of real customers, with a number decided in advance that will settle the question. The best sequence is interviews first, to find the segment, the job and the alternatives, and then a test, to see whether the offer built from them actually sells. Commitments gathered in interviews, such as deposits and booked trials, are the bridge between the two.
Two friends in Phoenix planned a meal-kit service for night-shift nurses: healthy dinners ready to eat at 3 a.m. They asked 60 nurses whether they would use it, and 48 said yes, many enthusiastically. On that evidence they nearly signed a lease on a commercial kitchen.
A mentor asked them to try again with different questions. This time they asked nurses to walk them through what they ate on their last night shift, where it came from and what it cost. The stories were revealing: most nurses brought food from home, prepared on their days off, at about 4 dollars a meal, and the main complaint was the hospital microwave line, not the food.
Then they offered a two-week trial at 9 dollars a meal for a 30-dollar deposit. Six nurses paid: 6 of 48 promises, 12.5 percent. The founders did not sign the lease. They started with six customers from a shared kitchen rented by the hour, and kept asking about last shifts rather than future ones. The deposits saved them from planning for 48 customers who were never going to pay.
Rob Fitzpatrick's short book The Mom Test, widely recommended to new founders, is built on the rule in this lesson: ask about people's lives and past behavior rather than your idea, because even your mother will lie to you kindly about whether she would buy. The book's test of a good conversation is whether it ends with facts and a commitment, not compliments.
Enthusiasm is demand. People are kind to people with ideas; kindness is not a purchase.
Ask what they would pay. A price someone has never had to pay is a guess, usually a generous one.
Explain the offer first so they understand. After that, the conversation is about your offer, not their life.
Record the promises; they're the strongest signal. The strongest signal is a commitment that cost the customer something.
Read the owner's first draft.
$\text{would you use us, and what would you pay?}$
A hypothetical and an invented price.
Rewrite the opening about the past.
$\text{the last time someone else cleaned your home}$
A real occasion, so the answer is a story with facts.
Add follow-ups about money and time.
$\text{what it cost; what went wrong; what you did}$
Costs show how much the problem matters.
Move the offer to the end.
$\text{could I call you when bookings open?}$
A small commitment, not a prediction.
Plan how to sort the notes.
$\text{facts, compliments, promises, commitments}$
Only facts and commitments count as evidence.
Count the owners who said they would buy.
$20$
Promises, free to make.
Ask each for a 50-dollar deposit on a trial set.
$5 \text{ paid}$
Commitments.
Find the commitment rate.
$5 \div 20 \times 100 = 25$
A quarter of the promises.
Find the share of all 40 interviewed who paid.
$5 \div 40 \times 100 = 12.5$
The rate to plan outreach with.
Plan the next round of outreach for 10 paying cafés.
$10 \div 0.125 = 80$
About 80 conversations, not 20.
Check the plan against capacity.
$80 \text{ conversations in a season}$
Possible over three months of fairs and visits.
List the notes.
$12 \text{ notes}$
From one forty-minute conversation.
Count the facts.
$6$
Platters ordered, budget, who approves, what went wrong.
Count the compliments.
$3$
'Sounds lovely', 'great idea', 'love the menu'.
Count the promises.
$12 - 6 - 3 = 3$
'We'd order every week' and two like it.
Find the share of facts.
$6 \div 12 \times 100 = 50$
Half the notes are evidence.
Ask for a commitment.
$\text{a paid trial lunch next Friday}$
She books it: one commitment.
Record what the interview proved.
$\text{six facts and one booking}$
Not 'she loves us'.
Name the kind of question in 'Would you bring your broken kettle to us?'
$\text{a hypothetical}$
It asks about an imagined occasion.
Rewrite it about the past.
$\text{the last time a small appliance broke, what did you do?}$
A real occasion with facts in it.
Add a follow-up about cost.
Northside Repairs wants to learn what people do when something breaks. Which question will produce evidence rather than a polite answer?
Complete the worked solution: Maya interviewed $50$ café owners. $20$ said they would definitely buy her cups, and when she asked for a deposit on a trial set, $5$ paid. Find the share who said yes, the share of yes-sayers who paid, and the share of everyone interviewed who paid.
Find the share who said yes.
$(\text{yes-sayers}) \div 50 \times \text{a hundred} =$ y
Promises, which cost nothing.
Find the share of yes-sayers who paid.
$(\text{payers}) \div (\text{yes-sayers}) \times \text{a hundred} =$ c
How many promises became commitments.
Find the share of everyone who paid.
$(\text{payers}) \div 50 \times \text{a hundred} =$ a
The first real measure of demand.
Compare the first and last shares.
$\text{yes-sayers far exceed payers}$
Kindness is not a purchase.
Maya drafted questions for café owners about their tableware. Sort each one.
| Past behavior | Hypothetical | Leading | Fishing for praise | |
|---|---|---|---|---|
| When you last replaced broken cups, where did you buy them and how many? | ||||
| Would you buy handmade cups if they were affordable? | ||||
| You'd agree handmade looks better, right? | ||||
| Do you like my speckled glaze? |
Dev from Northside Repairs has $19$ minutes with a cyclist. Put the parts of the conversation in order.
Number the steps in order (write the number in the box):
Notes from an interview with an office manager about catered lunches. Mark every note that is evidence.
This task has no paper form; do it on a device.
After six interviews, Bright Home Cleaning's owner sorted $20$ notes: $10$ facts about what people did, $5$ compliments and the rest promises about the future. Fill in each pile's count and its share of the notes, as a percentage.
| Notes | Share, percent | |
|---|---|---|
| Facts | ||
| Compliments | ||
| Promises |
At a county fair in Vermont, $50$ visitors told a maple farm they would sign up for a monthly syrup box. The farm then asked each for a $20$-dollar deposit, and $10$ paid. What percentage of those who said they would sign up made the commitment?
Answer:
A guitar teacher is planning to interview $9$ parents about lessons for their children. Which question should come first in each interview?
Lesson test: one question per skill, one attempt each, no hints. Your answers are checked when you submit.
After her interviews, Maya offered a trial set of café cups to the $20$ owners who had said they would definitely buy, for a $50$-dollar deposit. $2$ paid it. What percentage of those who said they would buy made a commitment that cost them something?
Of the owners who said they would buy, v percent paid a deposit.
You can turn 'would you buy this?' into a question about the last time, and tell a compliment from a commitment. Tell someone why twenty promises weigh less than a few deposits. Next: what customers do instead of buying from you.
20. Your turn: Northside Repairs and the kettle, step 3
$\text{what did that cost you?}$
Money shows how much the problem mattered.