Discovery by Silence
Stop pitching, ask an open question about their whole life, then say nothing until the roadmap falls out
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 7
- Confidence
- 95%
Weinstein's method for extracting the burning problem from a customer: never pitch, never anchor them to your product, ask a deliberately broad question about their day or their inbox, and then sit in silence long enough that they volunteer the thing they most want off their plate. Paired with a speed rule (respond to any unsolicited customer signal immediately) and a bounded-programme trick for controlling inbound volume, it replaces long UXR and survey cycles with a roadmap you can generate in weeks.
Origin
Jeff Weinstein's own practice as a product lead at Stripe, hardened by a failure: at his earlier SQL-tooling startup, the service accidentally bricked for 20 minutes and customers barely reacted. He high-fived his team for the quick recovery and only realised a year later that the absence of fury was the signal that he had no product-market fit. Stripe's identity products emerged directly from this listening practice — customers kept raising fraud/bot-signup problems that Stripe had no mandate to solve, so Stripe became an identity company.
Core principles
- 01People don't get out of bed for their second problem — they get out of bed for their first problem.
- 02Pitching anchors the conversation to your guess and guarantees you miss the burning problem; it is not the customer's job to interrupt your demo to tell you what actually hurts.
- 03Not pitching is itself a trust-building act. Silence earns you the right to hear the hardest thing.
- 04The time between a customer bothering to complain and your reply should be near-zero — that person gave you a finite, unbeaten gift.
- 05Speed of response converts detractors into promoters; the bar is not 'solve the problem' but 'turn them into a promoter'.
- 06You don't need 10,000 customers. Being text-message-friendly with five or ten of the right ones gives you overwhelming signal.
- 07Give customers write access, not just read access — some of them will design the product for you.
How to run it
- 1
Kill the pitch
Do not open with who you are, what your company does, and a demo. The moment you do, you've limited the surface to your guess at their top problem. Say almost nothing about yourself.
Watch out This is the single most common failure Weinstein sees from well-meaning founders on calls: 'hi I'm the CEO of XYZ, we do one two and three, I want to show you a demo.'
- 2
Ask a wide-open question about their actual life
Use prompts that don't mention your product at all: do you mind just opening your email — what's in there? If you weren't talking to me, what would you be working on? What grinded your gears last week? Magic wand — what do you wish was just off your plate immediately, forget my company, forget work? What did you do yesterday?
Pro tip It feels awkward because they don't know why you're asking. Sit with the awkwardness; ambitious people building their own business want to offload their hardest thing to someone.
- 3
Sit in the silence
Then say nothing. Do not rescue the pause. People will start to spill, and they get to their biggest problem fairly quickly. Weinstein recommends taking an improv class or two to get comfortable being out of your skin like this.
Pro tip Silence is also how you get the pricing signal — sit long enough and customers will sometimes volunteer 'I'll pay you money to solve X'.
- 4
Treat any unsolicited signal as a P0
When a customer goes out of their way to post, tweet, or email about your product, that is an unbelievable gift — most unhappy people just silently move on. Leave the meeting; change what you're doing; get one message back to them, even if it's 'I've got this, I'm heading to dinner, can I hit you up tomorrow?'
Pro tip That instant reply creates the feeling of a secret portal between a big brand and an actual curious human. It puts you on the right trajectory even before you fix anything.
Watch out When someone is hot on an issue — an SLA breach, a real bug — this is when speed matters most, and the bar is to turn them into a promoter, not merely to close the ticket.
- 5
Use a light structured ask to self-select the deep ones
To triage volume without a heavy process, reply asking for a screenshot, a video, or an email with three to five bullet points on how they hit the issue. It's not a 30-minute meeting and not a homework assignment, but it's enough structure that only those who want to go deep will do it.
Pro tip If someone does that work for you, you owe them a follow-up — and at that point you have a product friend for life you'll keep coming back to.
- 6
Bound the effort with an arbitrary programme
Invent a named, time-boxed programme to give people permission to engage and to cap your own exposure. Weinstein made up the 'Stripe Bugfinder Program' while driving to work, tweeted it, and said he'd take videos for three days.
Pro tip The programme can be deeply arbitrary and people will still ask to apply and treat selection as an honour. Report the results back publicly afterwards ('we ran our first programme, got 65 videos, found dozens of issues').
- 7
Let customers draw the product
Rather than running UXR about a future surface, invite customers into a shared diagramming tool (Stripe uses Whimsical) and ask them what they'd hope this dashboard would have. They will grab a shape and start drawing what they want.
Pro tip You don't need a hundred of them, and not all customers are designers at heart — but more of them are than you think. Sit in a bit more silence and someone will raise their hand.
In the wild
At Weinstein's SQL-analysis startup, a mistake bricked the service for 10-20 minutes with a couple of hundred customers. The team hustled it back up, high-fived, and moved on. Customers didn't email in fury, didn't text, didn't turn up at the door — just a small murmur.
→ Weinstein realised a year later that the calm was the signal: no burning problem, no product-market fit. He spent several more years on the product before selling the company, and now runs his whole discovery practice to avoid repeating it.
By listening without pitching, Stripe kept hearing the same adjacent complaint from customers building their businesses: how do I know the person signing up isn't fraudulent, or a bot, or not who they say they are? The internal reaction was 'we're not an identity company'.
→ 'Well I guess now we are' — because every customer had the problem. Weinstein describes this as building the roadmap via silence, dropping the long UXR/survey/build cycle.
An Atlas bug generated legal docs granting founders, say, 25 shares instead of 25% of shares. A founder spotted it in her documents and tweeted about it. Stripe fixed and regenerated the documents for everyone affected quickly, and the team engaged the founder directly rather than defensively.
→ The founder — who has a law degree — offered to read Stripe's docs any time and brainstorm on document generation. A near-disaster produced a long-term ally the Atlas team still works closely with.
Common mistakes
Opening with your demo
You've just anchored the customer to your guess at their top problem, and social norms mean they will not interrupt to redirect you to the real one. Companies that launch, get press, raise, then vanish usually traced back to pitching instead of listening.
Treating the absence of complaints as success
If your service goes down and people barely react, that isn't resilience or goodwill — it's the loudest possible signal that the problem you solve isn't burning. Most people apathetically move on rather than complain.
Delegating customer contact to a research function
Weinstein's whole practice depends on the PM being directly text-message-friendly with five to ten of the most ambitious, technical, fast-growing customers. Routing signal through a UXR pipeline strips the immediacy that makes it infectious internally.
Rescuing the silence
The awkward pause is the mechanism. Filling it with a follow-up question, a clarification, or a product mention hands the frame back to you and closes the window where the customer would have named their real burning problem.
Is it for you?
Best for
B2B product leaders and founders who need to find or validate a burning problem and want a roadmap in weeks rather than through a long research cycle
Not ideal for
Consumer/social products where Weinstein explicitly says he has less intuition, and situations where you genuinely have no access to real target customers yet
From the transcript
“people don't really get out of bed for their second problem right they get out of bed for their first problem”
“they were just pitching rather than sitting in silence and just waiting for their customer to just open their heart out about what's the most…”
“I'll say hey do you mind just opening up your email like what what's in there or if you weren't talking to me right now…”
“speed is an important one which is red just reducing the time between the moment the customer felt compelled enough to go out of their…”
“so I will leave a meeting I will CH change what I'm doing to just get one message back to them”
“respond to those folks is not to just sort of solve the problem but is to turn them into a promoter”
“if you're text message friendly with five or 10 of those you are going to have so much direct signal”
“do you mind sending me an email with three to five bullet points about the details of how you got to this issue”
“you're giving people permission through a pro program even if it is deeply arbitrary”
From the episode
Building product at Stripe: craft, metrics, and customer obsession
Jeff Weinstein (Product lead)