Cold DMs That Aren't Spam: Getting Your First Users One at a Time

You shipped the thing. It works. And now the honest, uncomfortable question: who is going to use it? The advice you keep hearing is "just tell people," which is useless, because the moment you try, it feels like exactly the kind of spam you delete without reading. So you do nothing. The product sits there with three users, two of whom are you on different browsers.

Here is the part nobody tells builders clearly enough. The way most solo founders actually get their first handful of users is not a launch, not an ad, and not a growth hack. It is sending real messages to real people, one at a time, and it feels slow and awkward because it is slow and awkward. But there is a version of this that works and a version that gets you muted. The difference is not confidence or charisma. It is who you message and what the first line says.

Where does the money actually come from?#

Cold DMs do not make money. They start conversations, and conversations are what eventually turn into users, and some of those users pay. It helps to see the whole chain, because every step loses people and that is normal.

Someone publicly complains about a problem you solve
  ↓
You send a specific, one-to-one message (help first, not a pitch)
  ↓
A fraction reply, because the message was actually about them
  ↓
A real back-and-forth conversation happens
  ↓
They try the product (you make that step tiny)
  ↓
A few keep using it, and some convert to paying
  ↓
Those first users tell you what to build and refer the next ones

The money is at the far end, but the leverage is at the top. If you message the wrong people, nothing downstream can save you. If you message the right people, even a clumsy message often gets a reply, because relevance beats polish. This is the same mechanism behind almost every early-stage business: revenue comes from solving a real problem for a real person, which is worth internalizing in where does online money come from. For the wider picture of how these early conversations fit into a business, how making money online works lays out the shape.

How it actually works#

Break it into two skills, because people conflate them and then blame the wrong one when it fails. Skill one is finding the right people. Skill two is writing the message. Most beginners have a message problem, but plenty have a targeting problem and never realize it, because they are sending decent messages to people who simply do not care.

Finding people who already have the problem#

You are looking for public frustration. Not a demographic, not a job title, an actual person saying an actual sentence that means "this thing is annoying me right now."

• X (Twitter) search. Search the literal phrases your users would type when annoyed. If you built a tool that cleans up messy CSV exports, search things like "hate cleaning up CSV" or "why is this spreadsheet export broken." The complaints are right there, timestamped and public.
• Reddit. Niche subreddits are full of people describing problems in detail because they are asking for help. This overlaps heavily with market research, and the same digging that finds prospects also tells you what to build, which is the whole point of mine subreddits for topics people want and reddit for market research.
• Communities you already belong to. Discords, Slacks, forums, indie maker groups. You have context here, which makes a message land far softer than a true cold one.
• Comment sections and question sites. Anywhere people ask "how do I..." about the thing you made easier.

The signal you want is recency and specificity. Someone who complained about the problem an hour ago is a wildly better prospect than someone with the right job title who has never mentioned it. Frustration is the buying signal. Chase the frustration.

Writing a message that gets a reply#

The person reading your DM is silently asking one question: is this about me, or about you? If it reads like a template, they answer "about you" and move on. Your job is to make it unmistakably about them.

A message that works usually has three parts and is short enough to read in one glance.

The specific hook. Reference the exact thing they said. "Saw your post about spending an hour every Monday fixing that export." This one line does almost all the work, because it proves you are a human who actually read their words, not a script firing at a list.

The genuine offer or observation. Give something before you ask for anything. That can be a quick tip that helps even if they never touch your product ("one trick that helped me was..."), or a light mention that you built a thing for exactly this. Help first. The order matters.

The tiny ask. Not "will you sign up." Something almost free to say yes to. "Happy to send you the link if useful, no pressure" or even just "curious how you handle it now." The goal of message one is a reply, not a conversion.

Notice what is missing. No paragraph about your background. No feature list. No "I'd love to get 15 minutes on your calendar." You are trying to start a conversation between two people, and nobody starts a good conversation by reciting their resume.

The follow-up#

Most replies you will ever get come after silence, not after the first message. People are busy and your DM got buried. One follow-up, a few days later, short and low pressure, often does better than the original. "No worries if not useful, just thought it might save you the Monday headache." That is it. One follow-up. Not five. Five is where you cross from persistent into the thing people mute you for.

A worked example (hypothetical numbers)#

These numbers are a made-up illustration to show the shape of the funnel. They are not typical, not a promise, and your real numbers will look different depending on your product, your targeting, and your message. Treat them as arithmetic, not as a forecast.

One focused week of DMs (hypothetical)
  People messaged (all had complained publicly):  40
  Reply rate:                          25%  ->  10 replies
  Turned into a real conversation:     50%  ->  5 conversations
  Tried the product:                   60%  ->  3 tried it
  Still using it two weeks later:      ~1 to 2 keepers

  Result: 1 to 2 genuine early users from 40 careful messages

Two things fall out of this. First, the reply rate is high compared to real spam (which lives around one percent) precisely because every one of those 40 people had already raised their hand by complaining. Targeting is why the top of the funnel does not collapse. Second, you only need a small number of these to matter. Ten real early users who talk to you is worth more than a thousand anonymous signups, because they will tell you what to fix and hand you your next users. That compounding is the real return, and get your first 10 customers is built entirely around it.

What you need and what it costs#

Required:

• An account on wherever your people complain (X, Reddit, a Discord). Free.
• The ability to search public posts for the phrases your users would use. Free, built into every platform.
• A product or at least a working demo you can point someone to. If there is nothing to try, you are collecting feedback, which is fine, just be honest about it.
• A simple list. A spreadsheet or a notes file: who you messaged, what they said, whether they replied, what happened next. Without this you will message the same person twice and forget every lesson.
• Your time and a tolerance for being ignored. This is the real cost.

Optional:

• A saved-search or alert tool so new complaints surface without manual checking. Useful once the manual version already works, not before.
• A lightweight CRM or Trello board once your list outgrows a spreadsheet.

What to skip: any tool that promises to automate the sending. The instant you automate DMs, you are back to mass blasting, which is the exact thing that does not work and gets accounts banned. The value here is that a human read a specific post and wrote a specific reply. Automate that and you have destroyed the only thing that made it work.

How long it takes#

Faster than SEO, slower than you want. A first reply can land the same day, because you are messaging people who are annoyed right now. A first real user often comes within the first week or two of consistent effort. But "consistent" is the load-bearing word. This is a daily habit of a handful of good messages, not a one-time sprint you do once and abandon when the first ten people ignore you. The founders who get traction this way are the ones still sending thoughtful DMs in week four, after the early silence, when most people have already quit and gone back to tweaking the landing page.

What beginners get wrong#

• Messaging by demographic instead of by frustration. "Developers who might like this" is not a target. "This specific person who complained about this specific thing yesterday" is.
• Pitching in line one. Leading with your product name and a feature list guarantees the mute. Lead with their words.
• Copy-pasting. The moment two of your messages could be swapped without anyone noticing, you are sending spam, no matter how polite the wording.
• Making the first ask huge. "Book a call" or "sign up for a trial" is too much to ask a stranger in message one. Ask for a reply, not a commitment.
• Taking silence as rejection. Most people who eventually become users ignored the first message. Silence is the default, not a verdict.
• Going quiet the second it feels awkward. It always feels awkward. That feeling is not a signal to stop; it is just what doing this feels like. If reaching out at all makes you want to crawl under the desk, selling when you hate selling is worth reading before you send a single message.

How I would start#

1. Write down, in the user's own likely words, the sentence someone types when they hit the problem I solve. Not marketing language. The actual frustrated sentence.
2. Search those exact phrases on X and in two relevant subreddits, and collect 20 to 30 real people who said something close to it in the last couple of weeks.
3. For each one, write a three-line message: their words, a genuine tip or offer, a tiny ask. Personalize the first line every single time.
4. Send a small batch, maybe five to ten, and log every one. Do not send a hundred; you cannot personalize a hundred and you will learn nothing from a blur.
5. When someone replies, drop the pitch entirely and just talk. Ask how they handle it now. Let the product come up naturally, or not at all in message one.
6. Follow up once, a few days later, on anyone who went quiet. Then leave them alone.
7. Repeat daily. Track reply rates, and when they are low, fix the targeting or the first line before blaming the whole approach.

If you want the structured, higher-volume version of this once it is working, founder-led sales in your first 100 conversations turns the same instinct into a repeatable process, and the broader playbook for landing paying users lives in our first customers hub.

What I would not do#

I would not buy a list, load it into a tool, and blast the same message to a thousand strangers. That is the spam you are trying not to be, and it does not even work: the reply rate is near zero and you burn the account doing it. I would not fake familiarity ("hey, loved your recent work!") when I had not actually looked, because people can smell it and it poisons the message. And I would not treat this as beneath me or as a temporary hack to escape as soon as possible. Talking directly to the people who have your problem is not the grubby part of building a product. It is the part where you find out whether the product should exist at all. Do it by hand, do it with respect, and do it every day. The first ten users almost never arrive any other way.