The first 100 users of a small product almost always arrive by hand: people you already know who have the problem, conversations in the places where the problem gets discussed, and direct replies to strangers who are asking for exactly what you built. You can count them one at a time, and you should. Launch sites and ads come later, if they come at all.
That is not the answer people want, because it does not scale. It is not supposed to. Paul Graham's essay Do Things That Don't Scale made this point years ago and it has not stopped being true: at a hundred users the job is learning what the product is, and you learn that from conversations, not from traffic.
Write down who has the problem
Before you post anything anywhere, write a list. Not "small businesses" or "developers". Twenty specific people or places.
- People you know who have the problem, or had it last year.
- Communities where the problem comes up: a subreddit, a Slack or Discord, a forum, a group for a profession.
- Searches people type when they have the problem. Write the actual phrases.
- Tools they already use whose forums, issue trackers or communities are full of people asking for what you built.
If you cannot write twenty, that is a finding. Either the problem is rarer than you thought, or you do not yet know who has it, and both are worth knowing before you spend a launch on it.
Start with the people you know
Your first ten users are usually people who already trust you a little. That feels like cheating. It is not, as long as they genuinely have the problem.
Do not ask them to "check out" your product. Ask whether they still have the problem, and offer to set it up for them.
You mentioned last month that chasing invoices eats your Fridays. I built a small thing that sends the reminders for you. Want me to set it up on your account this week? Takes ten minutes, and I would like to know where it falls over.
That message gets a yes far more often than a link does, because it costs them nothing and it is about their problem rather than your launch.
Reply, do not broadcast
The best source of early users is someone who has just written, in public, that they need the thing you built. They exist, and they are searchable.
- Search the communities on your list for the problem, in the words people actually use. Check every few days.
- When someone describes it, answer the question they asked first, properly, as if you had nothing to sell.
- Mention your product only if it directly solves what they described, and say plainly that you built it.
A helpful reply with a disclosed link reads very differently from a link with a sentence around it. The first one gets thanked. The second one gets removed, and in most communities it gets you banned.
Onboard every user yourself
At this stage, every sign-up is a person you can talk to. Email each one, from your own address, within a day.
Thanks for signing up. I am the person who built this. What were you hoping it would do? If anything was confusing in the first five minutes, I would really like to hear it.
Some will reply. Those replies are worth more than any analytics dashboard you can set up, because they tell you why, and the dashboard only tells you what.
Watch for three things in what people say:
- The words they use for the problem. Put those words on your landing page.
- The thing they expected that is not there. That is your next feature, or a sign you are describing the product wrong.
- Who they are. If the users who stay are not the people you built it for, follow the users.
Use launch sites as one burst, not a channel
A launch on Product Hunt, Show HN or a directory brings a day of attention, mostly from other makers. That is useful for feedback and a small amount of traffic. It is not a steady source of the users you want.
Do one or two, prepared properly, and treat them as a feedback event. Where to launch your product compares the main ones, what they cost, and what they are each good for.
Write the page people search for
The slow channel that compounds is search. Somebody with your problem types a question into a search engine today, and will again next month.
Write one honest, useful page that answers that exact question, with your product mentioned where it is genuinely relevant. Not ten thin pages; one good one. It will do nothing for weeks and then, if it is good, it will keep working while you sleep. Your landing page should use the same words: writing a landing page for a developer tool covers how.
Count where each user came from
Keep a simple table. A spreadsheet is fine. For every user in the first hundred, note where they came from and whether they were still using the product a week later.
| Source | Signed up | Still active after a week |
|---|---|---|
| People I know | ||
| Replies in communities | ||
| Launch day | ||
| Search |
After fifty users this table will tell you something no blog post can: which channel brings people who stay. Very often it is not the one that brought the most sign-ups. Double down on that one and drop the rest.
What not to do
- Do not buy lists or send cold mass email. It burns your domain's reputation and it reaches people who did not ask.
- Do not fake activity. Upvote rings, fake reviews and sock-puppet accounts get noticed, and getting caught costs more than the launch was worth.
- Do not spread thin. Ten communities with one low-effort post each reach nobody. Two communities where people recognise your name do more than all ten.
- Do not wait for "ready". If you are embarrassed by the first version, show it to five people anyway. They will tell you which embarrassment actually matters.
When you hit a hundred
At a hundred users you will know three things you did not know before: who the product is really for, which words make them sign up, and which channel brings the ones who stay. That is the point of the first hundred. Not the number, but the answers.
Then the work changes. You stop finding users one at a time and start building the one channel that worked into something that runs without you.