First customers
Getting your first 10 customers, one at a time
The first ten customers come from single conversations, not from reach. Reach only pays once you know which sentence works and who buys; before that it spends money on an unfinished message. Five routes get you there in weeks. First: your own network one step removed. Second: places where your audience already gathers. Third: researched one-to-one outreach. Fourth: intermediaries who already have the access. Fifth: delivering the service by hand. Plan for roughly 125 approaches to reach ten customers, and count only one thing as a yes: somebody paid.
Reach is the expensive way to get ten customers
The instinct after launch is to get in front of as many people as possible. It feels like the fastest route and it is the slowest one, for a reason that has nothing to do with the budget: reach multiplies a message. If the message is right, that is leverage. If it is not, and at the start it never is, you have paid to be misunderstood at scale.
The second cost is quieter. A rejection over an ad is silent: somebody scrolls past and you never learn why. A rejection in a conversation comes with a sentence attached, and that sentence is worth more at this stage than the sale would have been. The comparison below uses the same target of ten customers for both routes.
Two routes to the same ten customers
Buying reach
around $300 in ad spend, 2 weeksTen customers means five times that: around $1,500 spent before you even know whether the message is right. The 9,998 people who did not react will not tell you why.
Approaching people one at a time
$0, around 15 hours of workTen customers means around 125 approaches, roughly 38 hours. Expensive in time, cheap in money, and the 38 rejections arrive with a reason you can build the next approach from.
The figures illustrate the arithmetic, they are not benchmarks; your own rates will differ. What the comparison is about is the last line of each block: what you hold in your hands after ten customers.
Five routes that work in weeks
None of them scales, and none of them needs to. Pick two, run them in parallel, and give each three weeks before you judge it.
1Your own network, one step removed
Lead time: daysSuited to
For any idea where somebody you know knows somebody in the affected group. That is almost always true, and it is almost always skipped, because it feels like begging.
How to run it
Do not ask for a purchase, ask for a name. "Do you know anyone who cycles to work every day?" is a question that puts nobody on the spot. Your contact makes the introduction, and the conversation happens with a stranger who arrives warm.
The list runs dry after about 30 names. And a yes from a close friend is weak evidence: friends buy out of loyalty, not because of the problem.
2Places where the group already gathers
Lead time: 1 to 2 weeksSuited to
For groups with a shared subject: specialist forums, professional groups, associations, trade bodies, meetups, club meetings, local groups.
How to run it
Spend two or three weeks being useful and answering questions that have nothing to do with you. After that you can mention your own offer once without it grating. Reverse that order and you have spent the access and will not get it back.
Every venue has house rules and many ban promotion outright. Getting thrown out costs you more than the venue, it costs your standing in the whole group.
3Researched one-to-one outreach
Lead time: immediateSuited to
For groups whose individual members are publicly visible: businesses, practices, clubs, people who have written about the problem themselves.
How to run it
Three sentences, no more. First: an observation that is true of this one person only. Second: the link to your offer, in one line. Third: a small ask, usually 15 minutes of their time. Ten of those a day, by hand, no template.
Regulated, and very differently depending on the channel and the country (see below). And it does not scale. For the first ten customers that is not an objection, because the first ten do not need to scale.
4Intermediaries who already have the access
Lead time: 2 to 4 weeksSuited to
For groups somebody else already reaches regularly: associations, repair shops, accountants, club officers, employers, staff councils.
How to run it
Offer the intermediary something that serves their people, not you. A talk, a member rate, a piece of analysis. A single intermediary can deliver ten customers at once, which no other route does.
Slow, and dependent on one person. If they say no, the whole channel is gone. So never as your only route, always alongside routes 1 and 3.
5Delivering the service by hand
Lead time: immediateSuited to
For anything software or a process is meant to handle later. Instead of building it, you do the work for the first customers yourself, visibly and in person.
How to run it
The promise stays the same, the implementation is a spreadsheet, a phone and you. Customers rarely notice the difference, and every job teaches you the exact exception a finished product would never have shown you.
It carries you to roughly ten or twenty customers, then you are the bottleneck. That is the point: the bottleneck tells you what to automate first.
The sequence, from list to yes
Four to six weeks in total, alongside everything else. The order matters more than the pace: every step produces the input for the next one.
Write down fifty names, not five hundred
2 hoursA sheet with fifty rows: the name, where you found this person, and a reason why this one. Without the reason you will be missing the first sentence later. If you cannot fill the list, you do not have a customer problem, you have an audience problem: the group is described too vaguely.
Result: fifty rows, each with a reason to make contact.
Write three sentences and send ten a day
1 hour a day, 1 weekThe message is adapted per person, otherwise it is a circular and gets read as one. Ten a day is the volume you can sustain by hand. More means a template, and a template means no reply.
Result: fifty approaches sent, and a column with the date.
Ask about their day, then say the price out loud
2 weeksFifteen minutes is enough. First how the problem gets solved today and what that costs, in money or in time. Then your offer in two sentences, then the price, said out loud and without apology. Name the price only after they agree and you did not get agreement, you got politeness.
Result: one note per conversation with what the problem costs them today and how they reacted to the price.
Correct the approach after every fifth conversation
ongoingFive conversations are enough to see which sentence gets a nod and which one grates. Change one thing at a time, or you will not know afterwards what worked. The pitch from week three usually has little in common with the one from week one, and that is the output of the work, not its failure.
Result: an approach built out of answers rather than assumptions.
Think about reach only once ten customers have paid
afterwardsTen paying customers give you the sentence that works, the price that holds and the route they came through. Only with those three does advertising become an investment. Before them it is a bet on a message you do not know yet.
Result: a decision on which of the five routes to amplify, based on numbers rather than instinct.
Which price to say out loud in step 3: startup pricing, five steps to your first price →
Which yes counts, and which does not
Most encouragement a founder hears is free to give. The list below sorts the signals by what the other person actually put at risk.
| Signal | What it means | Weight |
|---|---|---|
| "Sounds great, good luck with it" | Politeness. Costs nothing and says nothing. | none |
| A sign-up to a waiting list | Mild interest, often just curiosity about the price. | low |
| A scheduled meeting the person actually keeps | They spent time, and time is scarcer than an email address. | medium |
| "Yes, if you can do X" | A named condition. Useful, because you can test it. | medium |
| A deposit, prepayment or signed order | Money changed sides. The only signal that does not lie. | high |
| A referral by name to a third person | The customer is spending their own reputation. Almost as strong as money. | high |
Where outreach becomes advertising
Talking to somebody is free of rules. Sending them an unsolicited commercial message is not, and the rules differ sharply depending on where the recipient sits. In the EU and the UK, commercial email generally needs prior consent, and business recipients are covered too. In the United States, federal law permits it as long as the sender is identifiable and honours an opt-out, with individual states adding requirements. Phone calls are treated more strictly than email almost everywhere.
The practical consequence is not that outreach is off the table, it is that the channel matters. Speaking to people where the group already gathers, replying to a request somebody posted publicly, being introduced by an intermediary and a physical letter all sit on much safer ground than an unsolicited email to a list you bought. Routes 1, 2, 4 and 5 in this article are unaffected by any of it.
General orientation, not legal advice. The rules differ by country and change; check what applies where your recipients are.
The same subject in full: cold outreach rules for founders, what you may send and to whom →
Five mistakes that cost the first ten
Every one of them feels productive from the inside. That is precisely why they last for months.
Building reach before the sentence exists. A channel, an account, a posting plan, all before the first customer conversation. That distributes a message nobody has tested, and the numbers afterwards only tell you that nobody reacted, not why.
Asking for approval instead of asking for the order. "What do you think?" always gets a friendly yes, because disagreeing is uncomfortable. The question that counts is: "Would you take this at $40 a month, starting next week?"
Automating too early. A sending tool, a sequence, a CRM for fifty contacts. The effort feels like progress and replaces exactly the part where you would have learned something: the individual, adapted message.
Counting friends as customers. A purchase out of loyalty proves nothing about the problem. They may be among the first ten, but list them separately, or you will talk yourself into demand that is not there.
Hiding the price. Wait until the end and you did not have a conversation about value, you had one about rapport. The price belongs in the first conversation, because the reaction to it is the real answer.
Frequently asked questions
How do I get my first customers with no budget?
Through five routes that all cost nothing in money. First: your own network one step removed, asking for names rather than for a purchase. Second: places where your audience already gathers. Third: researched one-to-one outreach of three sentences. Fourth: intermediaries such as associations or trade businesses that already have the access. Fifth: delivering the service by hand instead of building it. All five cost time rather than money, and each rejection arrives with a reason, which paid reach never gives you.
How many people do I have to approach to get ten customers?
As a planning figure: around 125 researched approaches for ten customers, so roughly 12 approaches per customer. That assumes a message adapted per person. Send the same circular to the same list and the number multiplies. Treat the order of magnitude as arithmetic, not a benchmark: it depends on how well the list matches the audience and how expensive the offer is.
Can I just cold email companies?
It depends heavily on where the recipient is. In the EU and the UK, unsolicited commercial email generally requires prior consent, and that applies to business recipients too. In the United States, commercial email is permitted under federal law provided the sender identifies themselves and honours an opt-out, and some states add their own rules. Either way, phone calls are treated more strictly than email. Check what applies where your recipients are. This is general orientation, not legal advice.
Should the first customers pay less?
Less yes, nothing no. An introductory price with an end date and something in return (feedback, a quote, a referral) is defensible and can be ended later without losing face. At zero nobody pays, so you learn nothing about willingness to pay, and moving from free to paid usually costs more customers than the discount ever brought in.
How do I know a route is working and I can do more of it?
Three things have to be true at once: the route has produced at least three paying customers, not just conversations. The effort per customer won is falling rather than rising. And you can name the sentence the yeses followed. If one of the three is missing, it was luck. Only when all three hold is it worth putting money into reach.
Read next
Will the idea survive the first ten?
The STRATEGIST checks the route to your first customer, the PESSIMIST names where it breaks. 4 to 12 AI experts, in minutes, from € 5.99.
Assess my idea →