Run a sequence, not a campaign. Forty named companies. One artifact only those companies would want. One message carrying one real observation about their business. A diagnosis call instead of a pitch. A small paid first step. Then take the sentence the first customer used about you and write it into everything. The first ten come from named people and evidence of judgment. Volume comes later, and only after the sentence exists.
- Forty named companies beats four hundred contacts. Small lists force real messages.
- Early buyers purchase evidence of judgment, not a finished product.
- A free pilot buys politeness. A small paid step buys information.
- The diagnosis call replaces the pitch. Do the work live, in front of them.
- The deliverable of the first ten customers is a sentence, and the sentence is what a channel gets built on.
Two founders in a kitchen with a camera
Airbnb had listings nobody booked. The hosts had photographed their own apartments badly, so the founders rented a camera and shot the New York listings themselves, one apartment at a time. Bookings went up. It was not a feature. It was not a growth channel. It did not scale, and it was never supposed to.
Stripe did a version of the same thing. When a founder said yes on a call, the Collisons asked for the laptop and set up the integration right there instead of emailing documentation and hoping.
Both companies get taught as growth stories. They were not. They were two teams doing embarrassingly manual work in front of ten people who could not believe someone finally understood the problem.
That is the whole answer, and most founders skip it because it looks too small to matter.
Why the usual approach stalls
The default is to build a list of four hundred contacts, write a sequence, send it, and watch a 0.8% reply rate come back. Then the diagnosis is that the copy was wrong, so the copy gets rewritten. Then the channel was wrong, so a new channel opens.
Neither one was the problem. Four hundred contacts forces templates, templates read as spam, and spam gets ignored by exactly the people who had the problem you solve.
The other failure is quieter. A founder posts consistently, gets likes, gets DMs from other founders, and calls it traction. Attention from peers is not demand. It feels identical from the inside, which is what makes it expensive.
The sequence
Seven steps. Roughly thirty days if you are honest about the first one. Nothing here needs a budget, a sales hire, or a finished product.
Write forty names, not a persona
Not "Series A fintech CMOs." Forty actual companies, with the human who owns the problem and the trigger that made it urgent. A funding round, a new competitor, a churn number that showed up in a board deck.
If you cannot get to forty, you do not have a distribution problem yet. You have a market definition problem, and no amount of outbound fixes that.
Cut forty to the ten who are on fire
Keep only the companies where the problem is costing money this quarter. Not next year. This quarter.
Fit tells you who could buy. Urgency tells you who will. Early on, urgency is the only qualifier that pays rent.
Publish one artifact only those ten would want
Build the thing that is useless to everybody else. A teardown of a company they compete with. A benchmark from data they cannot assemble alone. A diagnostic that scores the exact failure they are living in.
The test is uncomfortable on purpose. If a generalist audience would enjoy it, it is content. If only ten companies would care and one of them would forward it internally, it is an artifact.
Send it with one specific observation
One message per company. It contains one thing you noticed about their business that proves you looked, and the artifact. No cadence, no follow-up automation, no "just bumping this."
Ten messages that each took twenty minutes will out-convert four hundred that took four minutes, and it is not close.
Ask for a diagnosis, not a meeting
"Can I show you what I think is actually broken" gets a yes far more often than "do you have 30 minutes to hear about what we built." One offers them something. The other asks them for something.
Then do the work live on the call. Diagnose in front of them, out loud, including the parts that are unflattering. Slides perform expertise. Diagnosis demonstrates it, and the buyer can tell the difference inside of two minutes.
Charge for the next step
Give away the artifact. Give away the diagnosis. Charge for the work that follows, even if the number is small.
Free pilots produce enthusiasm and no information, because nobody defends a decision that cost them nothing. A paid first step tells you three things fast: whether the pain is real, who actually signs, and how slowly the company moves. Scope it to one decision so the yes is easy and the outcome is visible.
Take the sentence and reuse it
After the first customer sees a result, ask what they told a colleague about you. Write down the exact words. Do not improve them.
That sentence is the real deliverable of the first ten customers. Put it on the site, at the top of the deck, in the next outbound message. When five customers say roughly the same sentence without prompting, you have a position instead of a description.
What to do instead of the obvious thing
Most early-stage advice points at a channel. The channel is the last decision, not the first one.
| The instinct | What it produces | Do this instead |
|---|---|---|
| Build a 400-contact list and sequence it | Templates, 0.8% replies, a dead domain reputation | Forty names, ten messages, one observation each |
| Post consistently and wait for inbound | Peer applause that never converts | One artifact aimed at ten companies who feel it |
| Offer a free pilot to reduce friction | Polite enthusiasm, no signature, no signal | A small paid step scoped to one decision |
| Book a demo call | You presenting, them evaluating | A diagnosis call where the work happens live |
| Hire a salesperson at customer three | Someone selling a position that does not exist yet | Founder-led until five customers repeat one sentence |
Notice what the right column has in common. Every step trades reach for specificity. That trade is correct until roughly customer ten, and wrong after roughly customer fifty.
How you know it is working
Three signals, in order.
- A prospect replies with detail you did not ask for. That means the observation landed and they think you already understand the situation.
- Someone forwards the artifact internally. That is the first evidence the problem has a budget owner other than the person you emailed.
- A customer describes you to someone else in words you did not give them. That is position-market fit showing up before you named it.
If none of the three shows up across ten conversations, the failure is upstream. Either the forty names are wrong or the problem you are solving is not urgent for them. Rewriting the email will not fix either one.
This is the earliest form of relevancy engineering. Authentic core is you knowing the problem better than the buyer. Category ownership is the sentence they repeat. Distribution mastery, at ten customers, is forty names and a well-aimed artifact. Same three legs, smaller scale.
Find out what customers believe when you are not in the room.
Customer Truth Extraction applies the Customer Belief Audit and produces a Customer Belief Map, Perception-Reality Gap Analysis, and Customer Voice Synthesis. 2.5 hours. Internal adjectives get replaced by evidence.
Frequently asked
How many prospects do I need for my first ten customers?
Forty named companies is enough to produce the first ten customers, and the number is deliberately small. A list of four hundred forces templates, and templates are the reason early outbound fails. Forty is small enough that every message can carry one real observation about that specific business, which is the only thing that separates a founder message from spam.
Should I give the first customers the product for free?
No. Give away the artifact and the diagnosis. Charge for the work. A free pilot produces polite enthusiasm and no information, because nobody defends a decision that cost them nothing. A small paid first step tells you whether the problem is real, who signs, and how fast the company moves. Price the first step against one decision, not against your hours.
Do I need a finished product to get the first customers?
You need evidence of judgment, not a finished product. Airbnb's founders photographed listings themselves. Stripe's founders set up the integration on the spot during the sales call. Neither one was a scalable feature. Both proved the team understood the problem better than the buyer did, which is what an early customer is actually buying.
When does this sequence stop working?
Around customer ten to fifteen, when founder attention runs out. The sequence is manual by design and it does not scale. Its purpose is to produce the sentence customers repeat about you. Once five customers describe the company the same way, that sentence becomes the position, and the position is what a repeatable channel gets built on. Skipping ahead to channels before the sentence exists is why paid acquisition gets expensive early.