Before you have an audience, only a few channels move: people you already know, communities where your buyer already gathers, direct outreach, and referrals from early users. SEO and paid ads are real but slow or expensive at this stage and rarely produce the first hundred. Pick one channel, work it for a month properly, and judge it on conversations started rather than traffic.
The advice you find is written by companies that already have an audience. Do SEO. Build a content engine. Run ads and optimise the funnel. All true eventually, none of it useful when the number you are trying to move is from zero to one.
The short answer is that the first hundred customers come from four places: people you already have some connection to, places your buyer already gathers, direct outreach to strangers, and referrals from the people you have already convinced. Everything else is a channel for later.
Why the usual advice does not apply yet
A distribution channel is repeatable if the same effort produces roughly the same result next month. Most of what gets written about is repeatable only above a certain scale, and below that scale it does nothing while still consuming all your time.
Search is the clearest case. A new domain has no authority, so a post published today competes with pages that have been accumulating links for years. It will rank eventually, and eventually is not a plan when you need to know within a quarter whether anyone wants this.
Paid ads have the opposite problem. They work immediately and tell you almost nothing, because you are buying traffic from people who were not looking for you, and the cost per customer at low volume is dominated by learning rather than by the market.
The four that actually work
| Channel | Effort | What it gives you |
|---|---|---|
| People you already know | Hours | Fastest yes, least reliable signal |
| Communities your buyer is in | Weeks of participation | Trust, and the language they use |
| Direct outreach | Ongoing, systematic | The only one you can turn up on demand |
| Referrals from early users | Asking, explicitly | Highest conversion, needs customers first |
| Launch platforms | Days of prep | A spike, some links, rarely retention |
| SEO and content | Months before anything | Compounds, but not this quarter |
| Paid ads | Money plus attention | Speed, at a cost that hides the signal |
Note the ordering. The channels at the top are unscalable and work now; the ones at the bottom scale and do not work yet. Founders reliably invert this, because the bottom half feels like building a company and the top half feels like begging.
People you already know
Start here and be honest about what it gives you. Former colleagues, people from your last company, anyone who has watched you build things. They are easy to reach and they will take the call.
The trap is the signal. People who like you will say encouraging things about anything you show them, and enthusiasm from a friend is not demand. The only evidence that counts is the same evidence a stranger would provide: they pay, or they introduce you to someone who pays.
So the useful ask is not "what do you think". It is "who do you know with this problem", which converts a polite conversation into a referral and costs them nothing but a name.
Communities, if you are actually a member
Every buyer gathers somewhere. Developers in specific subreddits and Discords, operators in Slack groups, indie founders on Indie Hackers, each industry in its own forum nobody outside it has heard of. Those places contain your customers, already assembled, already talking about the problem.
They also have excellent immune systems. A founder who joins and posts a link is recognised instantly and, depending on the room, either ignored or banned. Worse, on the platforms that matter most the domain can get filtered, which is a permanent cost for a temporary post.
The sequence that works is slow and not complicated:
- Join and say nothing for a fortnight. Read enough to learn what the room considers obvious.
- Answer questions you actually know the answer to, with no link and no mention of your product.
- Keep answering. Two weeks of useful comments buys more standing than any post you could write.
- When you do mention what you are building, disclose it in the first line. Rooms forgive a founder who is upfront and are merciless with one who is not.
- Use the designated threads for anything promotional. Most communities have them, and ignoring them is the fastest way to lose the standing you just built.
The second-order benefit is worth as much as the customers: you learn the exact words people use for the problem, which is the raw material for every piece of copy you will write.
What a month of this looks like
Concretely: week one you read and answer nothing. Week two you answer two or three questions properly, at length, with no link. Week three you keep answering and, if someone describes exactly the problem you solve, you say what you are building and offer to show them, in a reply rather than a post.
By week four you are a recognised name in a room of several thousand people who share a problem. That is worth more than any single post would have been, and it is not reversible the way a banned account is.
Direct outreach, the only channel you control
Outreach is the one channel that works on demand. You do not wait for an algorithm, a community, or an audience. You decide to talk to forty people this week and then you do.
It is also the one most technical founders refuse, and the refusal is usually framed as taste: cold email is spam, it does not work, it is beneath the product. Some of that is true of bad outreach. Well-targeted outreach to people who genuinely have the problem is just a letter, and the reason it feels wrong is usually that it is uncomfortable rather than that it is ineffective.
If you are going to do it, do it properly: a narrow list of people who plausibly have the problem, a message that names their situation rather than your features, and a sending setup that does not burn your main domain. The mechanics are covered in cold email for developers and the deliverability side in why cold email goes to spam.
This is the part of the job Operater's Sales agent does, and the reason it exists is that it is both the most reliable early channel and the one builders are least willing to work by hand.
One thing worth setting expectations on: the numbers at this stage are small and that is normal. Forty well-targeted messages is not a campaign, it is a sample. You are looking for whether the message provokes any reaction at all, not for a conversion rate, and a single genuine conversation out of forty is a signal worth continuing on.
What kills most founder outreach is not the response rate. It is stopping after one week because the response rate of a sample that size is indistinguishable from noise, and reading that noise as a verdict on the product.
Referrals, which you have to ask for
Referrals convert better than anything else and almost nobody asks for them, because asking feels like imposing. The asymmetry is that a happy customer usually knows two or three people with the same problem and will name them if prompted, and will never think to do so unprompted.
Timing matters more than wording. Ask within a couple of days of the moment something worked for them, while it is still concrete, and ask for a name rather than a favour: "who else do you know dealing with this" is easier to answer than "would you refer us".
The launch, and what it is actually for
Launch platforms deserve a paragraph because founders plan entire quarters around them. A launch produces a spike of traffic for a day, some backlinks that persist, and occasionally a handful of customers. What it does not produce is a repeatable motion, and most platforms limit how often you can go back.
That makes a launch a one-off asset rather than a channel. Spend it when there is something worth clicking through to and a plan for the week afterwards, because the traffic arrives once and the question of what happens next is entirely yours to answer.
Where this advice is wrong
You already have an audience
If you have spent three years building a following, ignore most of this. Your first hundred come from telling the people who already listen to you, and the constraint is having something worth telling them rather than finding anyone to tell.
You are selling to enterprises
Long sales cycles and procurement change the arithmetic. Communities and launches are close to useless; the channels that work are direct relationships, introductions and events. The first hundred may also be the wrong target entirely, because the first ten might be the whole business for two years.
Your product is genuinely viral
If using it necessarily exposes it to other people, distribution is a product problem rather than a marketing one, and effort is better spent on the moment of exposure than on any channel here. This is much rarer than founders hope, and the test is whether it already happens on its own at small numbers.
Pick one, and measure the right thing
The common failure is doing a little of all four. A few community comments, a bit of outreach, a launch that got postponed, some SEO. Nothing gets enough volume to produce a signal, and after two months there is no data and no customers.
Pick one channel. Work it properly for a month. Measure conversations started, not visits: at this stage visits are noise, and ten real conversations tell you more than a thousand sessions. If a month of genuine effort produces no conversations at all, that is a result too, and it is worth more than another month of spreading yourself across four channels that each get a tenth of the attention they need.
A reasonable first month, if you have no idea where to start: pick outreach, because it is the only one that does not depend on anybody else showing up. Build a list of a hundred people who plausibly have the problem. Write to forty of them properly, follow up, and count the conversations rather than the replies.
At the end of that month you will have either some conversations, which tells you the message works and the constraint is volume, or none, which tells you something is wrong upstream of volume and more sending will not fix it. Both outcomes are worth a month. Four channels at a quarter of the effort each buys you neither.
Key takeaways
- Channels that scale later and channels that work now are almost never the same channels.
- Your first customers are found one at a time, by hand, and that is not a failure of strategy.
- A launch is a spike, not a channel: it produces a day of traffic and no repeatable motion.
- Communities work when you are a member first and a founder second, and punish the reverse.
- Judge an early channel on conversations started, not on visits, because visits at this volume are noise.