The First 100

Stripe asked for your laptop and set the product up on the spot. Airbnb knocked on doors in New York. PostHog spent roughly a thousand dollars and one Hacker News post. Dropbox paid up to $388 to win a customer who was worth $99, then changed course. None of them got their first 100 customers the way the growth blogs say you should.

The First 100: How Founders Actually Win Their First Customers is a research edition field guide by Angel Cee, founder of ROIpad.com. It starts from the startups instead of the tactics. Sixty four of them, from Stripe and Slack to Linear, Cursor, Harvey, Paystack and Chowdeck, each traced back to the moment the earliest users arrived, and to what the founders tried that did not work.

The question every founder faces and almost nobody answers well

Most advice about growth begins after the first hundred customers, when there is something to optimize. Before that point nobody is waiting for you. You are manufacturing demand by hand, and the useful question is not which channel scales but which channel works at all, for this product, with no brand behind it.

This book answers that question with evidence. Every startup is coded by product type and by the channels that account for its earliest users, so the patterns can be counted instead of guessed. Every claim is tied to a numbered source, and every case carries an evidence grade so you can see how much weight it can bear.

What the research found

• Trust beats marketing. The warm network was the most common source of first users, in 27 of the 64 cases. It led in B2B software, consumer apps and developer tools alike.
• The channel follows the product. Every marketplace in the dataset relied on field work. Developer tools leaned on communities and open source. B2B software leaned on networks and targeted outreach.
• Paid acquisition never won a first channel. Where it appears at all, it is in the failure ledger.
• A launch converts demand, it rarely creates it. Slack launched into a queue it had built with friends. Canva launched after a two year waitlist and still saw only about 2,000 signups in its first week.
• Attention is not demand. One founder with 50,000 email subscribers still needed about 15 months to reach 50 paying customers.

Inside the book

Sixteen chapters in five parts, plus two appendices.

• Part I, The Evidence. What the first 100 really means (users, accounts or suppliers), the compound growth arithmetic behind it, and the dataset, with a heat map of which channels each product type used.
• Part II, The Channels. One chapter for each of eight ways first customers arrive: the warm network, founder led outreach, communities and open source, launches and press, waitlists and content, field work, referral and invite loops, and investor and accelerator networks. Each covers the mechanism, the cases and exactly where it breaks.
• Part III, Context. How AI era products such as Cursor, Lovable and Harvey found their first customers, and three Nigerian case studies: Paystack, Chowdeck and PiggyVest.
• Part IV, The Graveyard. An eleven row failure ledger and seven patterns behind it, built from what founders later admitted did not work.
• Part V, The Playbook. A channel by product matrix, five diagnostic questions, a 12 week operating plan and a frank chapter on where the book could be wrong.
• Appendix A. A case library of all 64 startups: what each does, how its first users arrived, what failed or limited it, and how strong the evidence is.
• Appendix B. The numbered source list behind every claim.

A few of the stories you will meet

• Gusto found its first ten customers among friends starting businesses, including a children's swimming camp and a flower shop.
• Retool filtered Crunchbase for operations heavy startups, deliberately excluded one whole category, and landed DoorDash, Rappi and Brex.
• Superhuman put up a bare landing page that replied to every signup with two questions, and turned its waitlist into an interview pool.
• Segment posted an open source library to Hacker News and stayed near the top all day, after an earlier product had failed.
• Harvey began with thousands of LinkedIn messages, then trialled with a major law firm: 3,500 lawyers and about 40,000 queries.
• Chowdeck started in two Lagos neighbourhoods with 67 partner restaurants, and daily orders grew from about 40 to about 650 in a matter of months.
• PiggyVest grew on Twitter with almost no marketing spend, and its own users supplied the proof of trust.
• Facebook turned a single dorm mailing list into 1,200 to 1,500 registrants in 24 hours, then expanded school by school.

The failures nobody puts in a case study

Founders love to tell you what worked. This book also collects what did not: paid search that cost more than the product earned, two years of content that produced almost nothing, big launches that fizzled, selling to friendly founders who could not pay, and waitlists that never turned into customers. A failed channel is rarely a bad channel. It is usually a good channel used too early, at the wrong scale or for the wrong product, and the book shows you how to tell the difference before you spend the money.

Questions you will be able to answer

• Which one or two channels should I build around, given what I sell and who buys it?
• How many outreach messages do I need to send before I can judge whether the channel works?
• When is a launch worth doing, and what has to exist before it?
• Should my first customers be friends, or strangers who owe me nothing?
• How do I tell attention from demand?
• What did the founders who failed wish they had done in the first month?

A plan you can start this week

The final part turns the evidence into a 12 week sequence: narrow your segment and build a named list in weeks one and two, win your first ten by hand in weeks three to six, add one channel with enough volume to read the whole funnel in weeks seven to ten, then measure the weekly growth rate and double down on the single source that produces paying users. Each stage comes with the evidence behind it and a clear decision rule for when to keep a channel and when to drop it.

Honest about its limits

Few startups publish an exact first 100 breakdown, so the book uses each founder's earliest described traction as a proxy and says so. Cases are graded A, B or C by source quality, the channel coding is openly the author's judgement, and a whole chapter covers survivorship bias, founder retellings, period effects and geographic skew. You are told how far each finding can be trusted, which is rarer than it should be.

Who it is for

Solo founders and small teams who have a product, or are about to, and need their first paying customers. Developers who would rather ship than sell and want to know where to start. Founders in Africa and other emerging markets who rarely see their own context in Silicon Valley case studies. Anyone about to spend money on ads or a big launch who would like to see how that went for the companies that tried it first.

The most consistent finding in the book is also the least glamorous: first customers come from specific people who trusted a specific founder, and the founder did something that did not scale to earn it. The channels differ. The pattern does not. Find out which circle is closest to you, and how to work it.