Zway.ai

Playbook

Getting Your First 1000 Users

There is no channel that produces the first thousand users. They come from you, individually, in the places those people already spend time. Expect the first hundred to arrive one conversation at a time, and the next few hundred to come from those first hundred rather than from anything you post.

Last reviewed 27 August 2026

Outcome
A first thousand users acquired by hand, and a clear read on which single channel is worth scaling next
Timeframe
3 to 9 months

The steps

Run it in this order.

  1. 01

    List the ten places your buyers already gather

    Write down actual named places, not categories. This subreddit, that Slack group, this Discord, that conference, this GitHub repo issue tracker, that newsletter comment section. If you cannot name ten, you do not yet know your market well enough to acquire anyone, and that is the real finding of this step.

  2. 02

    Get the first ten users from people you can name

    Former colleagues, people from your old company, anyone who has complained to you about this problem. Message them individually and ask them to try it, not to buy it. Ten users who will tell you the truth are worth more right now than a hundred who signed up and never opened it again.

  3. 03

    Do a real call with every one of the first fifty

    Twenty minutes each, on video, watching them use it rather than asking them about it. Ask what they were doing immediately before they signed up. That answer tells you the trigger, and the trigger tells you where to find the next fifty. Founders who skip this step spend the next year guessing at positioning.

  4. 04

    Fix the thing three people hesitated on before adding anyone new

    If three of your first twenty stalled at the same point, stop acquiring and fix it. Adding users to a product with a known blocker converts your hardest-won attention into churn, and churned early users do not come back. This is the step that feels like it is slowing you down and is in fact the fastest route.

  5. 05

    Answer questions in public where your buyers ask them

    Pick two of your ten places and become a genuinely useful participant, with your affiliation stated in your profile and mentioned in thread when it becomes relevant. Answer questions in your area of expertise with no link attached. Standing accrues over weeks and is per-community, so two places done properly beats ten places done badly.

  6. 06

    Ask each happy user for one specific introduction

    Not a generic referral request. Ask whether they know one person with the same problem, and offer to write the message they can forward. Specific asks get acted on and vague ones do not. This is the mechanism that produces your second few hundred users, and it only works if the first hundred genuinely like the product.

  7. 07

    Do the unscalable thing deliberately and note when it breaks

    Onboard people personally. Set up their account for them. Sit on a call while they import their data. Keep doing it until the hours become impossible, and write down exactly which part became impossible first. That specific bottleneck is your next product priority, and you cannot find it by reasoning about it in advance.

  8. 08

    Only add a scaled channel once you can predict a conversation

    You are ready for paid, or content at volume, or partnerships, when you can say who converts, what they were doing beforehand, and what objection comes up on every call. Before that, a scaled channel just buys you more strangers, at cost, with no ability to tell whether the ones who left were the wrong people or had the wrong experience.

Why there is no channel for this

Channels work by finding people who already know what they want. The first thousand users of a product nobody has heard of do not know what they want, because the category the product sits in may not exist in their head yet.

That is the structural reason founders bounce off paid acquisition, SEO, and cold outbound at this stage. None of those are broken. They are simply instruments for capturing existing demand, and at a thousand users you are still manufacturing it, one person at a time, by explaining a problem to someone who had learned to live with it.

The work is therefore conversational rather than promotional. It is slow, it does not scale, and it is the only thing that reliably works.

The shape of the first thousand

UsersWhere they come fromYour job
1 to 10People you already knowAsk individually, watch them use it
10 to 50Their networks, plus direct outreachCall every one, find the trigger
50 to 200Communities where you have standingBe useful in public, disclose who you are
200 to 600Referrals from the first two hundredAsk for one specific introduction each
600 to 1000The first channel that showed real signalDo one thing repeatedly, well

Notice that the middle two rows are the largest and involve no marketing activity at all in the usual sense. They are participation and asking.

Being a person in a place

The single highest-yield activity available to an unknown founder is answering questions publicly, in a community where the answers are read by people with the problem.

It works because you actually know something. Whatever you built, you spent months learning the domain around it, and that knowledge is worth more to a stranger than any pitch. Answer the question fully, with no link, with your affiliation visible in your profile. Over several weeks you become someone the regulars recognise, and recognition is what makes a later product mention acceptable rather than removable.

The failure mode is obvious and extremely common. Arrive, post about the product in fifteen places in two days, get removed everywhere, conclude the community is hostile. Communities are not hostile to founders. They are hostile to strangers who arrive selling, which is a narrower and much more reasonable position.

The conversations are the product research

Every one of the first fifty calls should end with you knowing something you did not know before.

The most useful question is not what do you think of this. It is what were you doing immediately before you signed up. The answer names the trigger, and triggers are what you can go and find more of. A founder who knows that their users show up the week after a failed audit, or the day a teammate quits, has a distribution strategy. A founder who knows only that people like the product does not.

Write these down verbatim. The exact phrasing customers use becomes your headline copy later, and it will be better than anything you would have written yourself.

What breaks, and when

Doing things by hand works until it does not, and the point where it breaks is information rather than failure.

You will find that onboarding calls stop fitting in the week, or that the same setup question arrives twelve times, or that referrals dry up because the people you helped personally never got the same experience second-hand. Each of those is a specific product or process priority, discovered empirically rather than guessed at in a planning meeting.

That is the real argument for the unscalable approach. It is not virtue. It is that manual work surfaces the exact bottleneck, in order, and building against a real bottleneck is dramatically cheaper than building against an imagined one.

When to stop doing it this way

Stop when you can predict the conversation. When you can name who converts, what happened to them beforehand, and which objection appears on every call, you have the inputs a scaled channel needs.

Until then, more traffic just means more strangers, and strangers at this stage are expensive noise.

Playbook questions

How long does the first thousand realistically take?
For most B2B products, somewhere between three and nine months of daily effort, with the first hundred taking a disproportionate share of that time. If you get there faster, it is usually because you already had an audience or a distribution advantage. Treat any published timeline shorter than that as survivorship rather than method.
Is Hacker News or a launch site worth the effort?
Sometimes, but read the rules first and set expectations low. Hacker News guidelines state plainly that Show HN is not for blog posts, sign-up pages, newsletters, or lists, and asking people to upvote is explicitly penalised. A launch site produces a spike, and a spike is only worth having if you have the week free to talk to everyone it delivers.
Should I run ads to get to a thousand?
Almost never at this stage. Ads are an amplifier, and before product-market fit you do not yet know what you are amplifying, so you end up paying to learn something a dozen conversations would have taught you free. The exception is a small test used purely to check message resonance, treated as research rather than acquisition.
What counts as a user?
Someone who came back without you prompting them. Signups are a vanity number and they will make you feel considerably better than the situation warrants. Define an activation event that means the product did its job at least once, and count only the people who reached it. That number is usually a fraction of your signup count.
What if my total market is only a few thousand people?
Then this approach is even more clearly right, and the target changes. In a small market, individual relationships are the entire strategy and scaled channels never make sense at all. Aim for depth and revenue per account rather than a thousand logos, and treat the named list of every buyer in your market as the asset.

Or have us run it.

This playbook is yours to use. If you would rather not run it yourself, that is the entire point of Zway.

Book a demo

30 minutes. You leave with the plan either way.