How to Get Your First 100 Users With No Audience

A founder sent me his launch plan last spring. It was a Notion doc with eleven rows in it: Product Hunt, a Reddit post, a newsletter swap, six X creators he wanted quotes from, and a line at the bottom that said "waitlist conversion push." He had 1,900 emails collected over five months and he wanted to know which three creators to book.

I told him not to book anyone yet, which is a strange thing to say when booking creators is what my company does. Then I asked him how many of those 1,900 people he had ever spoken to. The answer was none.

That's the whole problem, and it's why the first hundred users don't come from marketing.

How to get your first 100 users

You get them one at a time, by hand. Find people already complaining about the problem in public, message them about the specific thing they said, give them the product free with your attention attached, and watch what they do. No paid distribution until you can predict what happens after a click.

That's the compressed version. The rest of this is the part people skip.

Why the first hundred are carried and not marketed to

Marketing is a multiplier. You point it at a message and it makes more of whatever that message already does. If your landing page converts 1% of warm traffic, buying more traffic gets you 1% of a bigger number and a lighter bank account.

At user zero you don't have a message. You have a guess about a message. You think people want the scheduling thing, and it turns out they want the scheduling thing only if it imports their existing calendar, and nobody will ever tell you that in an analytics dashboard. They tell you in a DM, at 11pm, after you asked a specific question about a specific screen.

Paul Graham wrote the canonical version of this in Do Things That Don't Scale, and it's fifteen-odd years old and still the most ignored essay in startups. People read it, nod, and then go build a waitlist page.

There's a second reason, and it's less philosophical. Hand-carried users retain. Somebody who got a personal message from the person who built the thing, and got a reply within an hour when they hit a bug, doesn't churn at the same rate as somebody who clicked a sponsored post at 2am. Your first hundred are also your first testimonials, your first referrals, and the people who defend you in a reply thread when a stranger dunks on your pricing.

Where the people who already have your problem are sitting

Stop thinking about audiences. Think about complaints.

Somebody, right now, is posting about the exact pain your product removes. They are annoyed, they are specific, and they are findable. That person converts at a rate that makes cold-audience advertising look like arson.

Here's where I actually look, in order.

Twenty of those a day for a week is 140 candidates, and if your product is any good you'll convert somewhere between five and twenty percent of them. That's your hundred, in about three weeks, without spending a dollar.

The part nobody wants to hear: this is boring. It's four hours a day of reading other people's mild irritation. Founders would rather spend two weeks building a waitlist page with a referral leaderboard on it than do fifteen minutes of this, and I understand the impulse, and it's still wrong.

The cold DM, and why every line is in it

Here's the one I give people. Copy it, then read the annotation, because the structure is doing more work than the words.

Hey Priya. Saw your post Tuesday about re-typing receipts into your accounting tool every month and losing the paper ones. I built a small thing that reads a photo of a receipt and pushes it straight into Xero, mostly because I was doing the same thing for my own books and got sick of it. It's free right now and honestly a bit rough. Want me to set it up on your account, or would you rather I just process your last month and send you the file so you can see if it's any good?

Four sentences. Here's what each one is for.

Line one is the filter. If you can only write "Saw your post about productivity," you didn't read the post, and they can tell in under a second because that's what every spammer in their inbox opens with. The detail is the credential.

Line two is narrow on purpose. "A tool for modern finance teams" gets ignored. "Reads a photo of a receipt and pushes it into Xero" gets a reply, because it's either the thing they wanted or it isn't, and either answer is useful to you.

Line three is the one founders delete, and deleting it is a mistake. Admitting the product is rough does two things: it takes away the fear that saying yes means being sold to, and it sets up the ask you actually want, which is for them to tell you what's broken. Free-and-rough invites feedback. Polished-and-priced invites silence.

Line four is a fork, not a question. "Would you like to try it?" is a yes-or-no, and no is easier. Two options where both end with you doing work turns the decision into a preference. And the second option, where you just do the thing manually and send back the result, is the one that converts. That's a concierge onboarding wearing a DM costume.

What I don't send: links in the first message on X (they suppress reach in DMs and read as bait), the word "quick," anything about a call, or a second message if the first gets no reply. One follow-up after five days, only if you have something new to say. Beyond that you're a nuisance and you deserve the block.

Volume expectation, plainly labeled as an example rather than measured Wiral data: if you send 20 well-researched DMs a day, expect single-digit replies. Twenty percent reply rate on genuinely specific outreach is good. Three percent means your line one is generic.

Waitlists rot, and here's the shape of the rot

A waitlist is a promise you made and then didn't keep for four months. The person who signed up in March had a problem in March. By July they've either solved it with a spreadsheet, paid a competitor, or forgotten your name entirely.

What to do instead, and this is the whole fix: let people in the same day, in whatever half-broken state the product is in, and gate on your capacity rather than on a launch date. Ten seats this week. Ten more next week. If the product genuinely can't be used yet, don't collect emails at all. Collect conversations. Reply to the person, ask them two questions about how they do it now, and put them in a text file with the date and what they said.

And if you already have a list of 1,900 that's been sitting for five months, don't blast it. Email the most recent 200 first, individually if you can stomach it, and watch the open and reply rates before you touch the rest. A dead list you burn in one send is a dead list forever. A dead list you probe carefully sometimes gives you thirty real users.

When paid creator distribution starts making sense

Here's where I'm supposed to sell you the thing I sell. Instead: at user zero, hiring creators is close to setting money on fire, and I turn this work down regularly.

The reason is arithmetic. Paid distribution multiplies your conversion rate, and at user zero you don't know your conversion rate, so you can't predict anything. Say you pay a creator with 40,000 followers for one post, at an illustrative rate of $600. Say that post does 25,000 impressions and 300 people click through, which would be a decent result. What happens next depends entirely on a number you haven't measured.

So the honest sequence is: fix the page with your hand-carried hundred, then buy distribution.

Concretely, I think you're ready to pay creators when you can say yes to all of these:

• You've talked to at least 40 users and you can quote the sentence they use for the problem, in their words, not yours.
• Your landing page converts warm traffic at a rate you've actually measured, not guessed.
• Something retains. Some group of your hundred came back in week two without you poking them.
• You can describe the product in one sentence and have a stranger repeat it back correctly.
• You have enough money to book three creators, not one. One post is a coin flip. Three is a read.

That last one is the part founders argue with me about most. A single sponsored post has enormous variance. It can land on a slow Tuesday, or the creator's audience can be the wrong slice of their followers, or the post can just be worse than their usual. Booking one and concluding "creator marketing doesn't work" is like judging a restaurant on one chip.

When those boxes are ticked, paid creator distribution stops being a gamble and starts being a multiplier on something you've already proven. That's the point where the creator network is worth your money, and where bringing us a brief makes sense. Before that, I'd rather you kept the cash and sent forty more DMs.

Your first hundred users aren't a growth problem. They're a research budget you pay in hours instead of dollars, and the founders who skip it end up buying attention for a product nobody has told them the truth about yet.