Setting the expectation
Launching does not produce users. A post announcing your product reaches people who follow launches, most of whom will look once and leave. That's not failure — it's just what launching is, and expecting more of it is the source of a specific and avoidable disappointment.
The first ten come from going to specific people who have the problem and talking to them individually. It's slow, it doesn't scale, and it's the only thing that reliably works. The value isn't only the ten users; it's that ten conversations tell you what to say to the next hundred.
Where to actually look
| Source | Why it works | The catch |
|---|---|---|
| People you already know with the problem | Trust exists; they'll be honest | A small pool, and they may be too kind |
| Communities where the problem is discussed | Concentrated, and already talking about it | Participate first; drive-by promotion gets removed |
| Reviews of competing products | Named people with a stated, specific complaint | Reach out about their complaint, not your product |
| Direct outreach to a defined list | Complete control over who you approach | Requires real personalisation to land |
| Answering the question in public | Compounds over time | Slow — this is a channel, not a start |
The message that gets a reply
Cold outreach works when it's evidently written for one person and asks for something small. It fails when it's a template with a name merged in, which everyone recognises instantly.
Subject: the itinerary question you posted Hi [name], You mentioned in [specific place] that you rebuild clientitineraries by hand every time a booking changes. I'm buildingsomething for exactly that and I'd rather show you than describe it. Could I have fifteen minutes this week? If it's not useful I'llsay so myself and stop taking your time. - [you]- Reference the specific thing they said or did — proof you're not sending five hundred of these
- State the problem in their words, not your product in yours
- Ask for something small: fifteen minutes, not a trial and a card
- Give them a graceful exit; it raises reply rates rather than lowering them
- Keep it under a hundred words
What to do once you have one
A first user is a research relationship, not a support ticket. Treat them accordingly and they'll teach you more than any analytics tool at this stage.
- Set them up yourselfOn a call, watching. Every place they hesitate is a design finding you'd otherwise never see.
- Give them a direct line to youYour email, your number. Ten users is not a support-queue problem.
- Follow up after a week with one question"Did you use it again?" — the answer separates polite interest from a real need.
- Fix their blocker immediately and tell themNothing generates loyalty like a complaint fixed the same week, with a message saying so.
- Ask who else has this problemReferrals from an early user are the highest-quality source available, and only exist if you ask.
When you're ready to scale a channel
Don't try to build a repeatable channel until you can answer three questions from evidence rather than hope.
- Do users come back without being prompted? If not, more users means a bigger leak
- Can you describe who it's for in one sentence a stranger would recognise themselves in?
- Did any two users arrive the same way? That repetition is the seed of a channel
Until all three are true, more effort on acquisition mostly buys you a larger sample of people who leave.
Common mistakes
- Waiting to launch instead of talking to ten people this week
- Posting in communities without ever participating in them
- Template outreach with a merged first name
- Optimising a funnel before knowing whether anyone returns
- Treating the first users as customers rather than as the best research you'll ever get
Key takeaways
- The first ten users are found individually; no channel skips this
- Specific communities beat large audiences
- Reference something real, ask for something small, keep it short
- Retention and a repeated arrival path are the signals you're ready for a channel
Try it yourself
Make a list of twenty specific people who have the problem — named individuals, not a category. Send five personalised messages this week. The replies, including the refusals, will teach you more than another month of building.