The first ten customers: a playbook for founders with no audience
Every startup story compresses the same awkward chapter into one sentence: "and then we got our first customers." The chapter is worth expanding, because customers one through ten behave nothing like customers one hundred through one thousand. They are not found by marketing. They are hunted, one at a time, by the founders.
Paul Graham's 2013 essay "Do Things That Don't Scale" named the pattern, but the mechanics deserve a field guide. Here is what actually works when the logo is unknown, the budget is a credit card and the product still has sharp edges.
Direct outreach is the channel, not a phase
Cold email and direct messages convert poorly at scale — one to five percent reply rates are normal — and wonderfully at ten. At this size you can research each recipient, reference their actual situation and offer something specific. A message that names the prospect's current tool, the workflow it breaks and a two-sentence fix outperforms any funnel at this stage, because there is no funnel yet, only conversations.
The constraint that makes it work: send fewer messages than you want to. Ten deeply researched outreach notes a day beat a hundred templated ones, and the replies teach you the language the landing page should later use.
The Collison installation
Y Combinator lore calls it the Collison installation: when an early Stripe user agreed to try the product, the founders did not send a link. They said "give me your laptop" and set it up on the spot. The principle generalizes: for the first ten customers, the founder personally absorbs all friction — onboarding, data import, the first workflow, the training call. Whatever the customer would have to figure out alone, do it for them.
Airbnb's founders flew to New York in 2009 and photographed hosts' apartments themselves. Superhuman's Rahul Vohra ran thirty-minute onboarding calls with every early user for years. None of this scales, and that is the point: the learning density of a hand-delivered customer is worth twenty self-serve signups who churn silently in week two.
- Define the ideal first customer narrowly enough to list fifty real names.
- Founders write every outreach message; no agencies, no sequences yet.
- Offer done-for-you setup in every pitch until customer twenty.
- Log why each prospect said no; the no-log is your real roadmap.
Price them from day one
Free pilot customers feel like progress and teach almost nothing. A customer who pays even a symbolic $50 a month has made a decision and will complain honestly; a free user has made no decision and will ghost politely. If ten people will not pay a discounted founder price, the answer is not "we need more features." It is a demand signal, and it arrived cheap.
| Channel | Founder effort | Typical reply rate | Works best when |
|---|---|---|---|
| Warm network | Low | 30–50% | You know the industry personally |
| Researched cold outreach | High | 5–15% | Clear niche with findable names |
| Communities (Slack, Reddit, forums) | Medium | Varies | You contribute before you pitch |
| Content and SEO | Medium | Slow | Months later, not for customer one |
| Paid ads | Low | Under 1% to trial | After retention is proven |
What ten customers actually give you
Ten paying, hand-onboarded customers are not a business, but they are proof of the three things everything else depends on: a reachable audience, a pain worth money and a delivery you can repeat. Only after those exist does it make sense to ask which channels scale. The next stage — turning conversations into a repeatable acquisition motion — is exactly what the unit economics guide on this site is for.