// chapter 00
Introduction
Talking to customers is essential — but almost everyone does it wrong, collecting false positives instead of truth. This book teaches you to ask questions that produce reliable data about whether your idea actually solves a real problem.
The archaeological dig
Learning from customer conversations is like excavating a delicate archaeological site. The truth is buried down there, but it's fragile. Most founders use one of two broken approaches:
- The bulldozer — forcing people to say something nice with heavy-handed questions like "do you think it's a good idea?" This shatters the truth.
- The toothbrush — being too timid to dig deep, flinching from uncomfortable questions, never finding out if anything of value is there.
The goal is the truth. Questions are your tools — but every question carries the risk of biasing the person and ruining the whole conversation.
Why conversations go wrong
We all know we ought to talk to customers. Many of us even do. Yet we still build things nobody buys. The problem isn't the talking — it's how we talk. Fitzpatrick learned this the hard way at his first company, after months of full-time customer conversations that turned out to be worthless.
Bad customer conversations aren't just useless — they're harmful. They hand you false positives that convince you you're on the right path, so you over-invest cash, time, and team into something that doesn't work.
Key takeaways
- The truth about your business is fragile — bad questions destroy it before you find it.
- Talking to customers is necessary but insufficient. How you talk decides whether you get truth or lies.
- False positives are more dangerous than not talking at all — they make you over-invest in the wrong direction.
- You don't need to be a natural salesperson. The good rules are learnable.
"Bad customer conversations aren't just useless. Worse, they convince you that you're on the right path."
// chapter 01
The Mom Test
You shouldn't ask anyone whether your business is a good idea. Talk about their life — problems, workflows, past behavior — and decide for yourself whether your idea is good, based on the facts you gather.
The core problem
People say you shouldn't ask your mom whether your business is a good idea. True — but it misses the point. You shouldn't ask anyone. Your mom lies the most (she loves you), but it's a bad question that invites everyone to lie at least a little. It's not their job to show you the truth; it's yours to find it, by asking good questions.
Failing vs. passing the test
In the failing version, the son pitches a cookbook app directly: "Would you buy this?" Mom gives lukewarm signals, he ignores them, piles on features, and she finally caves: "Oh, that sounds amazing, honey." He walks away thinking he's validated the idea.
In the passing version he never mentions the idea. He asks about her life: how she uses the iPad, the last thing she did on it, where the new cookbooks on the shelf came from. She can't lie because she's only describing real behavior — and the conversation yields gold (gift markets, younger cooks, where she discovers apps).
- Talk about their life instead of your idea.
- Ask about specifics in the past instead of generics or the future.
- Talk less and listen more.
Good questions vs. bad questions
| Bad question | Why it fails |
|---|---|
| "Do you think it's a good idea?" | Only the market can tell you. Everything else is opinion. |
| "Would you buy a product that did X?" | Hypothetical + wanting to please you = always "yes". |
| "How much would you pay for X?" | The number feels rigorous, but it's still a lie. |
| Good question | Why it works |
|---|---|
| "Why do you bother?" | Gets from the perceived problem to the real one. |
| "Talk me through the last time that happened." | Show, not tell — concrete past behavior. |
| "What else have you tried?" | Reveals current solutions, price anchors, and whether they care. |
| "Who else should I talk to?" | Multiplies your leads. |
You aren't allowed to tell them what their problem is, and in return they aren't allowed to tell you what to build. Gather facts about their life; the visionary leap to a solution is your job.
Key takeaways
- Never ask "is this a good idea?" — only the market can answer.
- Replace hypotheticals about the future with concrete questions about the past.
- If someone hasn't already looked for a solution, they won't buy yours.
- Opinions are worthless; behavior is everything.
- Always end with "who else should I talk to?"
Rules of thumb
- Opinions are worthless. Only the market decides if your idea is good.
- Anything involving the future is an over-optimistic lie. Ask about the past.
- People lie if they think it's what you want to hear — so don't reveal what you want to hear.
- If they haven't tried to solve it already, they won't buy your solution.
"You aren't allowed to tell them what their problem is, and in return, they aren't allowed to tell you what to build."
// chapter 02
Avoiding Bad Data
Three types of bad data sabotage your learning: compliments, fluff, and ideas. Deflect compliments, anchor fluff to concrete past behavior, and dig beneath ideas to find the real motivation.
1 · Deflecting compliments
Most meetings end with a compliment, and it feels great — but they're almost certainly being polite. Even genuine praise is worthless data. The best defense is not mentioning your idea at all; if praise slips out, deflect and redirect to facts.
Bad: "Thanks! We're predicting 35% cost savings." (accept praise, keep pitching)
Good: "Whoops — I got excited and started pitching. Mind if I ask how you handle this stuff at the moment?"
2 · Anchoring fluff
Fluff comes in three shapes: generic claims ("I always…"), future promises ("I would…"), and hypothetical maybes ("I might…"). The world's most deadly fluff is "I would definitely buy that." People are wildly more positive about an imagined future than the real one — Fitzpatrick's first startup lost ~$10M trusting fluffy promises.
Them: "I'm an Inbox-Zero zealot." You: "What's your inbox at right now?" → "When did it last fall apart?" → "Talk me through how you handled it." Now you're in reality, not fantasy.
3 · Digging beneath ideas
When people "flip" and start requesting features, write them down — but don't obey them. Understand the motivation first.
MTV asked for analytics, so the team built a full dashboard. Nobody used it — MTV's clients just liked receiving a nice-looking CSV every Friday. Three months wasted because no one asked "why do you want this?"
Dig into requests: "Why do you want that?" · "What would that let you do?" · "How are you coping without it?"
Dig into emotion: "Tell me more." · "That seems to really bug you." · "Why haven't you fixed it already?"
Key takeaways
- Bad data = compliments, fluff (generics/hypotheticals/future), and unexamined feature requests.
- "I would definitely buy that" is the most dangerous sentence a customer can say.
- Always anchor generic claims to specific past behavior.
- Understand feature requests; don't obey them. Ask "why?" until you hit the root cause.
- Emotional signals are goldmines — dig in, don't move on.
- The more you're talking, the worse you're doing.
Rules of thumb
- Compliments are the fool's gold of customer learning: shiny, distracting, worthless.
- Ideas and feature requests should be understood, not obeyed.
- If you've mentioned your idea, people protect your feelings. Keep your ego out.
- The more you're talking, the worse you're doing.
"The world's most deadly fluff is: 'I would definitely buy that.'"
// chapter 03
Asking Important Questions
Once you can ask unbiased questions, make sure you're asking the right ones — the scary, business-critical questions you're tempted to avoid, not the safe, trivial ones that keep things comfortable.
Find the scary questions
Run two thought experiments: imagine the company has failed — why? Imagine it's a huge success — what had to be true? Then go learn about those critical pieces.
Every time you talk to someone, ask at least one question that has the potential to completely destroy your currently imagined business.
Love bad news
Bad news early is the best outcome — it saves your time and money. Spending $5k to learn an idea is dead beats spending $50k to find out the hard way. And lukewarm responses are more informative than enthusiastic ones — "meh" is reliable data; don't pitch harder to convert it into a "wow".
Look before you zoom
Bad: "How often do you go to the gym?" → "Never." → "What's your biggest problem with the gym?" — they don't even care about fitness, yet you "validate" convenience.
Good: "What are your big goals right now?" → if fitness isn't on the list, they're not your customer. Move on.
Product risk vs. market risk
| Risk type | Questions it answers |
|---|---|
| Market risk | Do they want it? Will they pay? Are there enough of them? |
| Product risk | Can I build it? Can I grow it? Will they keep using it? |
Conversations validate market risk. Product risk (e.g. marketplaces, video games) usually can't be talked away — you have to build to find out.
Key takeaways
- Be terrified of at least one question in every conversation, or you're asking the wrong ones.
- Bad news is good news — it saves you from the wrong path.
- A "meh" is more reliable than a "wow". Don't pitch harder.
- Don't zoom into a problem before confirming they care about the area at all.
- Address every failure point, not just the most interesting one.
- Always prepare your top 3 questions beforehand.
Rules of thumb
- You should be terrified of at least one question in every conversation.
- There's more reliable information in a "meh" than a "Wow!"
- Start broad; don't zoom in until you've found a strong signal.
- You always need a list of your 3 big questions.
"You're searching for the truth, not trying to be right."
// chapter 04
Keeping It Casual
The best customer learning happens in quick, informal chats — not formal scheduled meetings. Strip the formality and you remove the bias, overhead, and awkwardness.
The problem with formal meetings
A one-hour meeting really costs about four hours once you count coordination, commuting, and review. Early on it's a big ask, and you may not have the credibility to land one anyway. The most precious startup resource is the founders' time and attention.
Instead of booking a formal meeting with a conference speaker, just ask: "How did you end up getting this gig?" You're now learning about their world — and they don't even know you're doing customer discovery.
How long should conversations be?
| Stage | Duration | Purpose |
|---|---|---|
| Early broad questions | 5 min | Is this a real problem? Does it matter? |
| Workflow & current solutions | 10–15 min | How do they solve it now? What have they tried? |
| Industry deep-dive | 60+ min | Learn the details from an expert. |
You can disprove an entire idea in five minutes of casual chat — Fitzpatrick killed an investor-dealflow idea after the investor pointed to a dozen sticky notes on the wall, their whole "system".
Key takeaways
- The best early conversations are 5–15 minute casual chats, not hour-long meetings.
- With your top 3 questions ready, any encounter becomes an opportunity.
- Formality kills learning — people switch into performance mode.
- Problem discovery never needs a meeting. Save those for when you have something to show.
- If it feels like they're doing you a favour, you're being too formal.
Rules of thumb
- Learning about a customer works better as a quick, casual chat than a long, formal meeting.
- If it feels like they're doing you a favour by talking to you, it's probably too formal.
- Give as little information about your idea as possible while still nudging the discussion usefully.
"If it feels like they're doing you a favour by talking to you, it's probably too formal."
// chapter 05
Commitment & Advancement
Once you've designed a solution, the only way to know it's real is to push for commitment — something they give up that has value (time, money, reputation). Compliments cost nothing and prove nothing. Commitment is the truth serum.
| Concept | Definition |
|---|---|
| Commitment | They show they're serious by giving up time, reputation, or money. |
| Advancement | They move to the next step of your real funnel, closer to a sale. |
The zombie-lead problem
When you don't push for advancement, you collect zombie leads — people who keep taking meetings, say nice things, but never commit. Your startup has been friend-zoned. There's no such thing as a meeting that just "went well": every meeting either succeeds (commitment to advance) or fails (a compliment or stall).
Time: a next meeting with clear goals, feedback on wireframes, a real trial. Reputation: intro to the boss/team, a public testimonial. Money: letter of intent, pre-order, deposit.
| Meeting ending | Verdict |
|---|---|
| "That's so cool, I love it!" | Bad — pure compliment. |
| "Let me know when it launches." | Bad — a stall, not a pre-order. |
| "I would definitely buy that." | Bad — dangerous false positive. |
| "Can I buy the prototype?" | Great — money for something unfinished. |
| "Come talk to the rest of the team." | Good — a reputation commitment. |
Earlyvangelists
Your first customers are "crazy" in the best way: they have the problem, know they have it, have a budget, and have already cobbled together their own workaround. When you see deep emotion, keep that person close — they're your first sale.
Key takeaways
- After learning, push for commitment to separate real customers from zombie leads.
- A meeting that ends with only a compliment has failed.
- Commitment = giving up time, reputation, or money. Compliments = nothing.
- "I would definitely buy that" is bad; "Can I buy the prototype?" is great.
- Not asking for commitment is worse than being rejected.
- Find earlyvangelists — emotional people who've already built workarounds.
Rules of thumb
- Friendly "customers" who never buy are a dangerous source of mixed signals.
- If you don't know what happens next after a meeting, it was pointless.
- The more they give up, the more seriously you can take their kind words.
- In early-stage sales, the real goal is learning; revenue is a side-effect.
"There's no such thing as a meeting which just 'went well'. Every meeting either succeeds or fails."
// chapter 06
Finding Conversations
The hardest conversations are the first ones. Cold outreach gets you started, but the goal is to convert it into warm intros as fast as possible — by immersing yourself in communities, teaching, organizing events, and framing meetings properly.
Going to them & bringing them to you
Cold calls are fine — if 98 of 100 hang up, you still have two conversations in play. But going to people puts you on the back foot. Better to plant a flag they come to: organize your own meetup ("the most unfair trick for rapid customer learning"), teach, blog, or host a panel and absorb the credibility of your guests.
Framing the meeting — V / F / W / P / A
- Vision — half a sentence on how you're making the world better.
- Framing — what stage you're at; you're not selling.
- Weakness — the specific problem you need help with.
- Pedestal — why they in particular can help.
- Ask — ask for help.
"Hey Pete, I'm trying to make desk rental less painful for new businesses (vision). We're just starting and have nothing to sell (framing). I only understand the tenant's side and I'm lost on the landlord's (weakness). You've rented desks for years and could help me cut through the fog (pedestal). Got time in the next couple weeks? (ask)"
The advisory flip
Don't go looking for customers — it creates a needy vibe. Go looking for advisors. It gives you a reason to be there, puts you in control, and changes the dynamic even if the topics are identical. Keep talking to a focused segment until you stop hearing new things — often just 3–5 conversations.
Key takeaways
- The goal of cold outreach is to stop doing it — convert to warm intros fast.
- Organizing your own meetup is the most unfair trick for rapid learning and instant credibility.
- Teaching makes customers come to you and take you seriously.
- Frame formal meetings with V/F/W/P/A — never let it default to a sales pitch.
- Start in person; phone/video loses too much signal early on.
- Look for advisors, not customers. 3–5 focused conversations can answer your core questions.
Rules of thumb
- If it's not a formal meeting, you don't need excuses — just have a good conversation.
- If it's a topic you both care about, find an excuse to talk about it.
- Keep having conversations until you stop hearing new stuff.
"The goal of cold conversations is to stop having them."
// chapter 07
Choosing Your Customers
Startups don't starve, they drown — in too many options and customer types. If feedback is inconsistent, your segment is too broad. Slice it down to a specific who-where pair you can actually find and serve.
Start narrow, expand later
| Company | Started with | Expanded to |
|---|---|---|
| PhD students finding obscure code | Everyone, anything | |
| eBay | Collectors buying Pez dispensers | Everyone buying/selling |
| Evernote | Moms saving recipes | Everyone saving anything |
Before you can serve everyone, you have to serve someone.
If you talk to 10+ people and results are still all over the map, your segment is too fuzzy. You're having one conversation each with twenty different customer types, not twenty with your customer.
Customer slicing
Drill into ever-more-specific groups: within this group, who wants it most, and why? What are they already doing about it, and where can you find them? Then pick the group that is profitable, easy to reach, and rewarding to serve. A good segment is a who-where pair — if you don't know where to find them, keep slicing.
Key takeaways
- Inconsistent feedback means your segment is too broad — not your questions.
- Before serving everyone, serve someone. Start with the narrowest viable segment.
- Keep asking "within this group, who wants it most?" until the group is findable.
- A good segment is a who-where pair.
- Choose your first segment for profitability, reachability, and personal reward.
- In multi-sided businesses, talk to every customer type, not just the payer.
Rules of thumb
- If you aren't finding consistent problems and goals, your segment isn't specific enough.
- Good customer segments are a who-where pair. If you don't know where to find them, keep slicing.
"Startups don't starve, they drown. You never have too few options — you have too many."
// chapter 08
Running the Process
Perfect questions are wasted without a process around them: prep before, take good notes during, review with your team after. The biggest failure isn't bad questions — it's becoming a learning bottleneck who hoards insights in their head.
The learning bottleneck
The classic anti-pattern: one "business guy" attends every meeting then tells the team what to do, with "the customer said so" as the trump card. Fitzpatrick bottlenecked so hard his CTO quit. Avoid it with three habits: prepping, reviewing, and good notes.
Prep with your whole team around your current list of 3 big questions. Two people per meeting is ideal — one talks, one takes notes. Founders must be in the room; you can't outsource customer learning.
Note-taking signal symbols
| Symbol | Meaning |
|---|---|
| :) / :( / :| | Excited / angry / embarrassed |
| ⚡ | Pain or problem |
| ⊓ / ☐ / ⤴ | Goal / obstacle / workaround |
| ☑ / $ | Feature request / money & budget |
| ☆ | Follow-up task |
Notes are useless if you don't look at them. Whatever medium you use, keep it lightweight enough that you'll actually review it.
Key takeaways
- Never become the learning bottleneck — share insights with exact quotes.
- Prep with your team: know your 3 big questions and target commitment.
- Two people per meeting: one talks, one takes notes.
- Use signal symbols to keep notes scannable and sortable.
- Review every batch: quotes, takeaways, what to improve.
- Founders must be in the room. Spend a week, not months.
Rules of thumb
- If you don't know what you're trying to learn, don't bother having the conversation.
- Notes are useless if you don't look at them.
- Go build your dang company already — the process should make you faster, not slower.
"Telling the rest of the team 'what I learned' is functionally equivalent to telling them 'what you'll do.'"
// chapter 09
Conclusion & Cheatsheet
Startups are more craft than science — you learn by doing. The tools in this book help, but you have to get out in the field. Make it happen.
The Mom Test (3 rules)
- Talk about their life instead of your idea.
- Ask about specifics in the past instead of generics or opinions about the future.
- Talk less and listen more.
Getting back on track
- Deflect compliments — they're the fool's gold of customer learning.
- Anchor fluff — bring generics, hypotheticals, and promises back to specific past behavior.
- Dig beneath ideas & emotions — find the real motivation behind requests.
A good meeting produces at least one of
- Facts — concrete specifics about what they do and why.
- Commitment — they give up time, reputation, or money.
- Advancement — they move to the next step of your funnel.
You're talking more than they are · they're complimenting your idea · you don't know what happens next · you have no notes (or never review them) · an unexpected answer didn't change anything · you weren't scared of any question · you're not sure which big question you're answering.
Framing the meeting (V/F/W/P/A)
Vision → Framing → Weakness → Pedestal → Ask. And the big prep question before any conversation: "What do we want to learn from these guys?"
"Go build your dang company already."
Want to learn this — and more — with me?
I'm sharing insights from every chapter as a learning-in-public series. New books are already on the way. Come follow along.