How to get your first 100 users after launch

The first 100 users don't arrive from a launch. You go and get them, one at a time, from a group narrow enough that you can name where they spend their time. That's slow and unglamorous, and it works because each early user is also a conversation about what to fix next.

Paul Graham's 2013 essay Do Things That Don't Scale is still the clearest description of this. His point is simple: nearly all startups have to recruit their first users manually, and founders underestimate how much small numbers compound. At 10% growth a week, 100 users becomes roughly 14,000 after a year.

Start with one narrow group

The mistake I see most is launching to everyone. "Small businesses" is not a group you can reach. "Independent coffee shops in Nottingham that run pop-up events" is.

Facebook is the famous example of starting narrow. It began only for Harvard students, and because they felt it was really for them, enough of them signed up quickly to make it useful. The same logic applies to a B2B tool or a marketplace: a tight group that talks to each other gives you word of mouth for free.

Write down three things before you contact anyone:

• Who exactly. A job title, a type of business, a place, or a community.
• Where they gather. A Slack group, a subreddit, a trade event, a market, a LinkedIn hashtag, an industry newsletter.
• The one problem you solve for them. In their words, not yours.

Recruit by hand

Then go and talk to people. Direct messages, emails, phone calls, turning up in person. Keep each message short, specific to the person, and about their problem.

When someone says yes, set them up yourself. The best-known version of this is what Graham calls the Collison installation: when someone agreed to try Stripe, the founders didn't send a link, they asked for the laptop and set it up on the spot. For your app, that means a 20-minute call where you create the account with them, import their data, and watch where they hesitate.

Then make those first users feel looked after. Wufoo sent every new user a handwritten thank-you note for as long as it could. You don't need stationery, but a personal message from the founder in the first week is cheap and memorable.

Real examples

These are the public, well-documented ones, all from Graham's essay:

• Airbnb. Early on, the founders went door to door in New York, recruiting hosts and helping existing ones improve their listings. Graham describes the company as so fragile at the time that about 30 days of in-person work made the difference.
• Pinterest. Ben Silbermann noticed many early users were interested in design and went to recruit at a conference of design bloggers.
• Stripe. The Collison installation above: the founders onboarded people in person rather than waiting for them to try it alone.
• Wufoo. Handwritten thank-you notes to every new user.

None of these needed a big audience. They needed founders willing to do work that wouldn't scale, aimed at a specific group.

Channels compared

Here's how I'd weigh the usual options for a first 100. The ratings are my judgement, not data.

Show HN and Product Hunt

Both can bring a spike of visitors. Both have rules worth reading before you post.

Hacker News describes Show HN as being for something you've made that other people can play with. Landing pages, sign-up pages and fundraisers don't qualify, and if your work isn't ready for users to try, they ask you not to post it. They also say plainly: don't ask friends to upvote or comment. Your title starts with "Show HN", and the guidelines suggest making it easy to try, ideally without sign-ups.

Product Hunt's launch guide says a launch is visible from midnight Pacific Time and runs for the entire day, and you can schedule a date up to 30 days ahead. Its rule on promotion is short: ask for feedback, never for upvotes. Midnight Pacific is 8am in the UK for most of the year, so a UK founder can be awake for the start.

Treat either as a bonus once you have users who like the product, not as the plan.

Make the product ready for strangers

Hand-recruited users forgive a lot because they know you. Strangers from a launch post don't. Before you send traffic anywhere, check the basics:

• Sign-up works on a phone. Test it on a real device on mobile data, not the desktop preview.
• The empty state explains what to do. A new account with nothing in it is where most people leave.
• Errors are readable. No raw stack traces or blank screens.
• You'll know when it breaks. Error alerts go to you, not to a log nobody reads.

My app launch checklist has the full list, and the guide to launching an AI-built app covers the fixes AI tools usually leave behind.

Measure the right few things

With under 100 users, dashboards mislead. A handful of numbers per user tells you more:

• Where they came from. Record the source at sign-up.
• Did they reach the first useful moment? For a booking app, a first booking. For a dashboard, first data connected.
• Did they come back in week two?

The simplest way to record the source is to carry the campaign tag from the link into the sign-up form:

"use client";
import { useSearchParams } from "next/navigation";
 
// Wrap the parent in <Suspense> when the page is statically rendered.
export function SignupForm({ action }: { action: (data: FormData) => Promise<void> }) {
  const params = useSearchParams();
  return (
    <form action={action}>
      <input type="email" name="email" required autoComplete="email" />
      <input type="hidden" name="source" value={params.get("utm_source") ?? "direct"} />
      <button type="submit">Create account</button>
    </form>
  );
}

Store source on the user row and you can answer "which channel brought the people who stayed" with one query. For the wider set-up, read my guide to analytics for an MVP, which also covers consent.

Talk to every one of them

The real value of the first 100 is the conversations. I'd aim to speak to as many as you can in the first month and ask the same few questions each time: what were you using before, what nearly stopped you signing up, what would make you recommend it.

Patterns show up quickly. When three people stumble on the same screen, that's your next sprint.

When the product is the bottleneck

Sometimes outreach works and the product doesn't hold up: sign-ups fail, pages are slow, payments break. That's the point where the prototype needs to become a product. My prototype to product guide covers that step, and the free health check is a quick way to see what's likely to break in an AI-built app. If you'd rather hand it over, my Launch package takes a prototype to production, priced from the MVP Build. Book a call and tell me who your first 100 are.