"Talk to customers" is the most repeated advice when it comes to building early-stage companies. And the founders who take the maxim to heart will eventually ask the same question, Where do I find them?
Most ask it like the answer is a list of places. dscout, userinterviews.com, your customer support logs, LinkedIn. And the list isn't wrong, but the list also isn't the answer.
It sounds like an ops problem, a sourcing checklist, something to solve once and move on from. But it's one of the most, if not the most, consequential design decision in the whole process.
More than your question guide, more than facilitation, where you recruit from determines whose story you hear. And whose story you hear determines what you learn which then determines what you build.
The list is where that starts — but not where it ends.
Start with who you know, then go further immediately
Tomer London and his Gusto co-founders started with a list of 30 people from Stanford and other networks. People they knew, or knew of. But here's what made it work: every conversation ended the same way. "Who are your friends who have small businesses?" Tomer shared this in Lenny’s newsletter in 2023.
That question is load-bearing. It's what turned a warm list into a research practice. In parallel, Tomer was cold-calling businesses off Yelp.
By the time they walked into the meeting that made them commit, they'd already heard the pattern. When Tomer asked people what they felt about their current payroll provider, more than half started cursing. Literally. Not "that sounds interesting." Cursing.
That's what you're looking for, not polite agreement. The feeling, the emotion that people can't contain.
Get the ask right
The Zip founders — Rujul Zaparde and Lu Cheng — spoke with ~75 CFOs, heads of procurement, and finance leaders in 2-3 weeks. They generated 110 pages of notes. Their approach on LinkedIn was simple: they asked for advice.
Rob Fitzpatrick warns about a failure mode in “The Mom Test.” He says, if you’re unclear why you want the meeting, it defaults to a sales call. And no one wants a sales call.
But the ask can work if you’re precise: you’re working on something, you don’t have it figured out, you believe they have perspective you can’t get elsewhere, and you want 20 minutes to learn.
No pitch, no hidden agenda. People generally want to be helpful when you’re specific about what you don’t know — by and large.
The five-part cold outreach message
Let's start with the message. The structure is simple: vision, framing, weakness, pedestal, and ask. But underneath each part is a sharper question — who did you see, what did you diagnose, what proves you've been here before?
An example using the work I do at How This Works co:
Let’s break it down:
Vision signals seriousness and focus — and names the trigger that made you reach out
Framing removes the sales-call reflex and opens space for honest diagnosis
Weakness gives them something concrete to respond to — and doubles as proof when it names a real pattern
Pedestal makes the outreach specific, not generic — the sharper your observation, the less it reads like spam
Ask is clear, bounded, and low-friction
Use this mnemonic to remember: “Very Few Wizards Properly Ask.”
Drop one of these and the message weakens. No vision, you sound like a student. No framing, you sound like a seller. No weakness, there’s nothing to respond to. No pedestal, it feels like spam. No ask, it goes nowhere.
The order can flex. But the components all build off each other to best effect.
One conversation should lead to three
Cold outreach works, the Zip founders proved it at ~75 conversations in three weeks, but it's a toll, not a strategy.
Each conversation should produce the next one. Ask who else deals with this, ask who you should talk to next, and so forth.
The person who just explained their problem to you knows three or more others living the same one.
That’s how cold turns warm — and warm turns into better signal. The snowball either starts rolling or it doesn’t. If every interview is a transaction, you stay cold. If every interview earns the next one, you don’t.
Open tabs in my browser
Spellman College commencement speech by Symone Sanders Townsend
Benedict Evans’ presentation AI Will Eat the World at SuperAI, the whole presentation is well worth your time but especially the IBM ad from the 1950s promising 150 extra engineers slide
Greg Storey’s Creative Intelligence Builder
Getting someone to say yes to talk is the first problem. The second is more consequential: not everyone who says yes gives you the same signal.
Who you recruit determines whose story you hear.
Next issue: what happens when you get that decision wrong.
For reference
The goal of cold conversations is to stop having them.
Fitzpatrick's point isn't about avoiding outreach or even cold outreach — it's about compounding. One good interview leads to the next. The person who shared their workflow problem knows a few more people living the same one. Warm referrals don't happen by accident. They're the signal that you found the right pool.
Thanks,
Skipper Chong Warson
Book a call
Only humans buy anything, build for your Bullseye Customer
—
Know someone else who should read this?
The folks who need this most aren't trying to figure it out alone — they're stuck on a list of places. If that sounds like a team you know, forward this their way.