How do you decide whether this is a discovery call or a sales call (they're different)?

Discovery calls and sales calls have opposite goals. In a discovery call you're trying to learn the truth about a prospect's problem — you should not pitch, demo, or persuade. In a sales call you've validated the problem and you're trying to advance the deal. Most founders blur the two and end up doing both badly: learning too little and selling too soon.

How do you spend 25 of 30 minutes asking questions?

Your job is to be a scientist, not a lawyer. Don't show the demo. Don't pitch the vision. Don't list features. Ask questions, then ask follow-up questions, then ask follow-up questions to those. The most useful discovery calls feel like the prospect is teaching you their job for 25 minutes.

How do you ask about past behavior, not future intentions?

Humans are terrible at predicting their future behavior — even our own. 'Would you use this?' is a useless question because the honest answer is 'I have no idea.' Ask about what they've already done instead: 'How did you handle this problem last quarter?' / 'What did you try before that?' / 'Why didn't that work?' Past behavior is data. Future intentions are speculation.

How do you listen for active spend, not abstract pain?

The strongest signal that a problem is real is that the prospect already spends money to address it. Money spent on a partial solution, a workaround, a consultant, a competitor — any of these tells you a budget exists. It's far easier to redirect spend than to create new spend. If the prospect has never spent a dollar on this problem, the problem isn't real to them yet.

How do you distinguish real demand from polite interest?

Most founders mistake politeness for demand. 'That sounds great, keep me posted' is a polite rejection. Real demand sounds completely different — the prospect interrupts to ask when they can get access, starts scheduling follow-ups with their colleagues, asks about pricing unprompted, or wants to see the product right now. If you're squinting to see demand, you don't have it.

How do you always end with a closing test?

Even in a pure discovery call, end with an explicit ask that forces a real answer: 'Based on what you've told me, would you be willing to sign an LOI?' / 'If we had this built in 4 weeks, would you commit to a 6-month pilot at $X?' The point isn't to literally close — it's to surface what's blocking the buy decision (integration needs, wrong buyer, no budget, timing).

How do you use graduated-fidelity tests to validate demand cheaply?

Don't build the product to test demand. Validate at the cheapest fidelity that gives you signal, then increase fidelity only when the signal stays positive. Start with questions (zero cost). Then a landing page (one day of work). Then a Figma mockup (a week). Then a working prototype (a month). Only invest more after the previous step generated real interest, not just polite nods.

What goes wrong: You ask leading questions to confirm your idea?

The most common discovery mistake. You spent 6 months on this idea — of course you want it to be right. You ask 'wouldn't it be great if there was a tool that did X?' and the prospect politely says 'yeah, sure.' You file that as a positive data point and keep building. Be a scientist, not a lawyer.

What goes wrong: You mistake politeness for demand?

Salespeople especially give wildly positive false signals — they're trained to be encouraging. 15 sales leaders saying they want your product can produce 0 customers if none of them click the signup link. Watch what they do, not what they say.

What goes wrong: You pitch when you should be listening?

It's hard to resist when the conversation lulls. But every minute you spend pitching is a minute you're not learning. Pre-PMF, the call is for learning. Save the pitch for the second call.

What goes wrong: You ask about future behavior instead of past behavior?

'Would you use this?' is a useless question because the honest answer is 'I don't know.' 'How did you handle this last quarter?' is a useful question because the answer is fact, not prediction.

What goes wrong: You confuse enthusiasm for authority?

The eager IC who loves your product can't buy it. The skeptical VP who controls budget can. Always ask about the buying process: who approves, what's the cycle, when's the budget reviewed.

What goes wrong: You don't end with a closing test?

Without an explicit ask, you walk away with vague vibes. Force a real answer. 'Would you sign an LOI?' / 'Would you commit to a 6-month pilot at $X?' The reaction to that ask is more useful than 25 minutes of discovery on its own.

Playbook · Customer discovery

Run customer discovery without poisoning your data

The Mom Test in playbook form. Most founders run discovery calls that produce useless data — they ask leading questions, mistake politeness for demand, and build the wrong product. Here's how to do it right.

Who this is for

Pre-PMF B2B founders. Anyone running customer interviews, user research, sales discovery, or design-partner recruiting conversations.

Time to ship

Diagnostic in the next call. Useful pattern after 15-20 calls.

What you’ll do

You'll decide whether you're running a discovery call or a sales call (they have opposite goals). Spend 25 of 30 minutes asking questions, not pitching. Ask about past behavior, not future intentions. Listen for active spend on the problem. Distinguish real demand from polite interest. End every call with an explicit closing test. And use graduated-fidelity tests so you validate at the cheapest level that gives signal.

The steps

01
Decide whether this is a discovery call or a sales call (they're different)
Before booking
Discovery calls and sales calls have opposite goals. In a discovery call you're trying to learn the truth about a prospect's problem — you should not pitch, demo, or persuade. In a sales call you've validated the problem and you're trying to advance the deal. Most founders blur the two and end up doing both badly: learning too little and selling too soon.
- Pre-PMF: almost every call should be a discovery call. You're scientists, not lawyers.
- Post-PMF: discovery becomes the first 5-10 minutes of a sales call, not a separate call.
- If you're not sure which kind of call this is, default to discovery. The cost of learning more is low; the cost of selling too soon is a poisoned data point.
↳ Reference: Discovery call architecture
02
Spend 25 of 30 minutes asking questions
Per call
Your job is to be a scientist, not a lawyer. Don't show the demo. Don't pitch the vision. Don't list features. Ask questions, then ask follow-up questions, then ask follow-up questions to those. The most useful discovery calls feel like the prospect is teaching you their job for 25 minutes.
- Open with their world, not your product: 'Walk me through what your week looks like.' / 'What part of your job annoys you the most?'
- Get them to share their screen. Watch them work. Bad demo = you show them yours. Good discovery = they show you theirs.
- If you find yourself talking for more than 3 sentences in a row, you've slipped into pitch mode. Stop and ask a question.
↳ Reference: The 70/30 listen-talk rule
03
Ask about past behavior, not future intentions
Per question
Humans are terrible at predicting their future behavior — even our own. 'Would you use this?' is a useless question because the honest answer is 'I have no idea.' Ask about what they've already done instead: 'How did you handle this problem last quarter?' / 'What did you try before that?' / 'Why didn't that work?' Past behavior is data. Future intentions are speculation.
- Bad: 'Would you pay for a tool that does X?' Good: 'What have you spent money on to address this problem?'
- Bad: 'Would you switch from your current tool?' Good: 'When was the last time you switched tools for this? What triggered the switch?'
- Bad: 'How important is this problem?' Good: 'What happens if this problem isn't solved this quarter?'
↳ Reference: Discovery question taxonomy
04
Listen for active spend, not abstract pain
Per call
The strongest signal that a problem is real is that the prospect already spends money to address it. Money spent on a partial solution, a workaround, a consultant, a competitor — any of these tells you a budget exists. It's far easier to redirect spend than to create new spend. If the prospect has never spent a dollar on this problem, the problem isn't real to them yet.
- Probe specifically: 'How much do you spend on X today?' / 'Who pays for that?' / 'How was the budget approved?'
- Active workarounds (a custom internal tool, a recurring contractor engagement, a manual spreadsheet someone updates every Monday) are spend signals even when no money changes hands.
- If the prospect says 'we don't really spend on this' and you can't find a workaround, the problem isn't burning enough to fund.
↳ Reference: ICP hypothesis testing
05
Distinguish real demand from polite interest
End of every call
Most founders mistake politeness for demand. 'That sounds great, keep me posted' is a polite rejection. Real demand sounds completely different — the prospect interrupts to ask when they can get access, starts scheduling follow-ups with their colleagues, asks about pricing unprompted, or wants to see the product right now. If you're squinting to see demand, you don't have it.
- Polite interest: 'sounds interesting,' 'keep me posted,' 'send me more info,' 'would you like to do a follow-up?'
- Real demand: 'can I get access today?' / 'who internally needs to be looped in?' / 'what does pricing look like?' / 'I'd push my CEO to look at this'
- The 'screen share test': if they won't show you their current workflow on screen share, they don't care enough.
- The 'click test': if they say they want it but don't click the signup link, they don't want it. Watch what they do, not what they say.
↳ Reference: Reply triage and signal classification
06
Always end with a closing test
Last 5 min of every call
Even in a pure discovery call, end with an explicit ask that forces a real answer: 'Based on what you've told me, would you be willing to sign an LOI?' / 'If we had this built in 4 weeks, would you commit to a 6-month pilot at $X?' The point isn't to literally close — it's to surface what's blocking the buy decision (integration needs, wrong buyer, no budget, timing).
- The reaction to the closing ask tells you more than 25 minutes of discovery. 'Yes, send me the LOI' is data. 'Let me check with my CTO' is data. 'That feels premature' is data.
- If they deflect: dig into the deflection. 'What would have to change for you to be willing to sign?' / 'Who else needs to be part of this decision?'
- Track every closing test in your notes. The pattern across 10-20 calls tells you whether you have demand or polite enthusiasm.
↳ Reference: Qualification frameworks
07
Use graduated-fidelity tests to validate demand cheaply
Across the discovery program
Don't build the product to test demand. Validate at the cheapest fidelity that gives you signal, then increase fidelity only when the signal stays positive. Start with questions (zero cost). Then a landing page (one day of work). Then a Figma mockup (a week). Then a working prototype (a month). Only invest more after the previous step generated real interest, not just polite nods.
- Landing page test: 'Are people landing here and giving you their email?' If 0/100 visitors convert, the offer is wrong before you build anything.
- Figma mockup test: 'Can you walk me through how you'd use this?' If they don't reach for it, the workflow is wrong.
- Prototype test: get them to use it on real data for 1-2 weeks. Did they keep coming back? Did they tell anyone? Did they ask when it'll be ready?
↳ Reference: ICP hypothesis testing

What goes wrong

The failure modes that catch most founders.

You ask leading questions to confirm your idea
The most common discovery mistake. You spent 6 months on this idea — of course you want it to be right. You ask 'wouldn't it be great if there was a tool that did X?' and the prospect politely says 'yeah, sure.' You file that as a positive data point and keep building. Be a scientist, not a lawyer.
You mistake politeness for demand
Salespeople especially give wildly positive false signals — they're trained to be encouraging. 15 sales leaders saying they want your product can produce 0 customers if none of them click the signup link. Watch what they do, not what they say.
You pitch when you should be listening
It's hard to resist when the conversation lulls. But every minute you spend pitching is a minute you're not learning. Pre-PMF, the call is for learning. Save the pitch for the second call.
You ask about future behavior instead of past behavior
'Would you use this?' is a useless question because the honest answer is 'I don't know.' 'How did you handle this last quarter?' is a useful question because the answer is fact, not prediction.
You confuse enthusiasm for authority
The eager IC who loves your product can't buy it. The skeptical VP who controls budget can. Always ask about the buying process: who approves, what's the cycle, when's the budget reviewed.
You don't end with a closing test
Without an explicit ask, you walk away with vague vibes. Force a real answer. 'Would you sign an LOI?' / 'Would you commit to a 6-month pilot at $X?' The reaction to that ask is more useful than 25 minutes of discovery on its own.

Want the technical depth?

The chapters with the full reference detail.

We can help you build the discovery program.

If you want a second set of eyes on your discovery process — call recordings reviewed, interview script audited, signal-vs-noise feedback — we'll do a 60-min review and give you written notes. Useful before you write the next 30 calls.

How we run this for you →Book a call →

← All playbooks Library home