There is a specific trap in early business, and almost everyone falls into it. You build something, launch it, and wait for customers. Nothing happens. So you improve the product, relaunch, and nothing happens again.

The problem was never the product. The first hundred customers do not arrive — they are fetched, one conversation at a time, by methods that feel far too manual to be a real strategy.

Customers 1 to 10: go and get them

These come from direct, individual effort. Not marketing — asking.

Start with people who already know you, and with anyone you spoke to while validating the idea. Those conversations are warm leads you have already had, and most people never go back to them.

The message that works is specific and low-pressure. Name the problem you know they have, say what you built, and ask whether it is worth ten minutes. Not a broadcast announcement that you have launched — a message to one person about their situation.

Expect this to feel uncomfortable and inefficient. Ten personal messages producing two customers feels like failure compared to imagining a launch that produces a hundred. It is not — it is the actual mechanism, and the two customers are real.

Deliver these first ones manually, even if you plan to automate everything later. Watching someone use the product teaches you what to fix in a way no analytics dashboard will.

Customers 10 to 50: be useful in public

Direct outreach runs out. The next stage is reaching people who do not know you, and the reliable route is being visibly helpful where they already are.

Pick two or three communities full of your buyers — not communities full of people doing what you do. Then answer questions properly, for weeks, without mentioning what you sell. When someone describes a problem you can solve, explain how to solve it for free.

Some people implement your advice. Others message you asking whether you would do it for them. That second group arrives pre-sold, because they have already watched you demonstrate competence.

Content works the same way and compounds better. One piece a week answering a question your buyers actually search for keeps working long after you publish it. It is slow for the first few months and then it is not.

Customers 50 to 100: find the one channel that works

By fifty customers you have data. Ask each of them how they found you, and a pattern will be visible.

Usually one channel is producing most of your customers while you have been spreading effort across five. The move is to concentrate on that one rather than continuing to do a little of everything.

This is also the point where paid acquisition starts making sense, because you now know what a customer is worth and can judge whether the spend clears. Before fifty customers you are guessing at both numbers — our ROAS simulator and CAC guide cover how to check.

Referrals are the channel everyone forgets to build

Referred customers typically cost the least to acquire and stay the longest. Most businesses receive some by accident and never ask for any deliberately.

• Ask at the moment of success, when the customer has just got the result they wanted. Not months later.
• Ask a specific question. “Do you know anyone else dealing with this?” gets answers. “Please refer me” does not.
• Make it easy. Offer to write the introduction message yourself so they only have to forward it.
• Say thank you visibly. People who are thanked for a referral make another one.

Formal referral programmes with rewards can work, but for the first hundred customers simply asking outperforms building a system.

Keeping the ones you get

Acquisition gets all the attention and retention decides whether any of it was worth it. Losing customers as fast as you find them means running hard to stay in the same place.

Early customers churn for reasons that are usually fixable and usually invisible unless you ask.

• They never got the result. The most common cause. They signed up, got confused, and quietly stopped. First-week onboarding matters more than any feature you could add.
• It solved a one-off problem. Not a failure — it means your model should be one-off purchases rather than recurring.
• Something broke and nobody responded. Early customers forgive bugs and do not forgive silence.
• They forgot you existed. A monthly email costs nothing and prevents more churn than most product work.

With a small customer base you can do something that stops working later: talk to every single one. Thirty personal conversations will tell you exactly what to fix, and they also make people considerably less likely to leave.

Track where each one came from

This is dull, takes seconds per customer, and is the single most useful habit in early business.

Keep a simple sheet: name, date, how they found you, what they paid, what nearly stopped them. After fifty entries the pattern is unmistakable, and you will almost certainly find you have been spending most of your effort on a channel producing almost nothing.

The most reliable source of the “how they found you” answer is asking them directly at signup or in the first email. Analytics will tell you the last click; the customer will tell you what actually convinced them, and those are frequently different things. If you are running campaigns, tag the links properly so the two can be reconciled — our UTM builder handles the formatting.

What not to do

• Waiting for a launch. Launches produce a spike and then silence. Steady manual acquisition beats a launch you spent two months preparing.
• Spending on ads too early. Before you know your conversion rate and customer value, paid traffic is buying data at a premium price.
• Posting into the void. Announcing your product to an audience of zero on five platforms is activity, not marketing.
• Building more features. The most comfortable way to avoid selling. If nobody is using what exists, more of it will not help.
• Discounting to get started. A low price attracts people who were never going to be good customers and sets an anchor you cannot move.

Talk to everyone who says no

The rejections in this phase are worth more than the sales, and they are free.

Ask what stopped them. The answers cluster quickly: too expensive, missing a specific thing, solved it another way, or did not understand what it does. Each of those points at a different fix, and the fourth one — they did not understand — is the most common and the easiest to correct.

Ten nos with reasons attached will tell you more about what to change than a hundred page views ever will.

Frequently asked questions

How long should the first hundred take?

Anywhere from a few months to a year, depending on price and market. High-priced B2B products reach a hundred customers slowly; low-priced consumer products need volume faster. Neither pace is wrong.

Should I offer a free tier?

Only if free users convert or bring in paying ones. Early on, free users consume support time while teaching you little about willingness to pay, which is the thing you most need to learn.

What if I have no network at all?

Then start at stage two. Communities and content build a network rather than requiring one. It takes longer, and it works.

Is cold outreach worth it?

Researched, specific outreach works. Templated mass messaging does not. Ten messages that each name something real about the recipient beat two hundred generic ones.

Why a hundred is the milestone

At a hundred customers you can see which channel works, what objection recurs, who your best customer actually is, and what they are worth. Those four answers are what let you spend money on growth with any confidence.

Below that number you are guessing, and scaling a guess is how marketing budgets disappear. The manual, unscalable work of the first hundred is what earns you the right to stop doing manual, unscalable work.

If you have not tested demand yet, start with validating the idea before chasing customers for something nobody has agreed to pay for.