How to get your first customers
Your first ten customers come from people you can reach one at a time, by hand, in a conversation: your own network, the places your customers already gather, and anyone who has publicly complained about the problem you solve. They do not come from Product Hunt, a launch tweet, or SEO. Those channels need an audience to work, and at this stage you do not have one. The method is unglamorous and it is the same every time: make a list of a hundred specific people, contact them individually, and ask for money earlier than feels comfortable.
Make the list of a hundred
Before you write a single message, write a list of a hundred named people or companies who plausibly have the problem. Named, not a description of a segment. If you cannot get to a hundred, that is not a marketing problem, it is a sign you do not yet know who this is for, and no amount of outreach fixes that.
Sources, in the order they usually pay off:
- People you already know who fit. Uncomfortable, and by a wide margin the highest converting.
- Places your customers already gather. Subreddits, Slack and Discord groups, industry forums, a LinkedIn or X niche. Be there as a person for a while before you sell anything.
- People who have complained in public about the exact problem. Search the phrasing they would use, not the phrasing you use.
- Competitors' unhappy customers. Reviews, support forums, the replies to their posts.
- One degree out. Ask ten people who fit which two people they know who also fit.
What to actually say
The first message should be short enough to read on a phone, specific enough that it could not have been sent to anyone else, and it should ask for a conversation rather than a sale.
A shape that works: one line on why you are contacting them specifically, one line on the problem in their words, one question. No product description, no link, no calendar. The goal of message one is a reply.
The most common mistake is describing your product. Nobody has capacity for a product they did not ask about. They do have capacity to talk about a problem that annoys them.
Ask for money earlier than is comfortable
A free user tells you almost nothing. Enthusiasm is cheap, and "this is great, I would definitely use it" is the single most misleading sentence in early business. The only reliable signal is a card.
So ask early, and ask for a real price. If the product is not finished, sell the outcome rather than the software: an early access price, a pilot, a paid setup. If ten conversations produce a lot of warmth and no payment, you have learned something valuable and cheap, which is that the problem is real but not expensive enough to solve yet.
Do it in volume, and count
The honest numbers, roughly: warm contacts reply at maybe one in three, cold ones at one in twenty, and a fraction of replies become customers. A hundred contacts is not aggressive, it is the minimum sample where the result means anything.
Track four numbers and nothing else: contacted, replied, spoke to, paid. If you are getting replies but no conversations, the message is wrong. Conversations but no payment, the offer or the price is wrong. No replies at all, the list is wrong.
Why launches do not work yet
Product Hunt, a launch post, a Show HN: these distribute to people who already follow the channel, and they convert at a rate proportional to how well the product is already known. With no audience they produce a spike of traffic, a handful of signups who never return, and a day of dopamine that costs a week of preparation.
They are worth doing later, once you have customers and can say something specific about who this is for and what it did for them. At ten customers you have that. At zero you do not, and the launch is the most expensive way to find out.
Frequently asked questions
How do I get my first customer with no audience?
One at a time, by hand. List a hundred named people who plausibly have the problem, message each individually asking about the problem rather than pitching the product, and ask for money in the conversation. No audience is required for that, only a list and a fortnight.
How many people do I need to contact?
A hundred, to get a result that means anything. Warm contacts reply at roughly one in three and cold at one in twenty, so smaller numbers tell you about luck rather than about the business.
Should I offer it free to get the first users?
Free users tell you almost nothing, because enthusiasm costs nothing. If you need people using it to make it better, be explicit that it is a trade and time-boxed. For validation, a card is the only signal that is not noise.
Is cold outreach still worth doing in 2026?
Individually, yes. Generic and at scale, no, and it is worse than no outreach because it burns the list. The version that works is specific enough that it could not have been sent to anyone else, which is slow, which is the point.
What if nobody replies?
Almost always the list, not the message. Either the people on it do not have the problem badly enough, or you have described them as a segment rather than found them as individuals. Rebuild the list before rewriting the email.