The First Ten Customers Don't Come From Marketing

The night before we launched, I was in a Slack group of about forty operations people, the kind of group where everyone half-knows everyone and nobody posts much. I had been lurking in it for two years. I knew who was frustrated with their current vendor because they had said so, out loud, in the channel, months earlier. So I sent eight direct messages. Not a blast. Eight separate notes, each one referencing something that specific person had complained about. Three replied that day. Two of them became paying customers inside the week. That was the launch. There was no funnel. There was no campaign. There was a guy who had paid attention for two years and then asked.

I think about this every time a founder tells me they are "working on their go-to-market" before they have a single customer, and what they mean is they are picking fonts for a landing page and arguing about whether the blog should post twice a week or three times.

You are not running out of awareness, you are running out of trust

Here is the thing nobody says plainly. Marketing, the real discipline, is mostly a machine for taking demand that already exists and pointing it at you efficiently. It assumes there is a market that knows it has a problem, a category it understands, and a set of competitors it is choosing between. None of that is true for you yet. You have a product five people have seen and a category that exists only in your own head. Building a content engine for that is like building a drive-through for a restaurant that has no kitchen.

The first ten customers are not an awareness problem. They are a trust problem, and trust at that stage does not scale, because it is made of your own face and your own name. People buy the earliest version of anything because they believe a specific human will not let them get burned. That belief comes from knowing you, or from watching you say smart things in a room they respect, or from you emailing them something so obviously useful and specific that it could not have been automated. It does not come from a nurture sequence.

So the work, the actual job in month one, looks nothing like marketing and everything like sales you are too embarrassed to call sales.

What the embarrassing work actually is

It is making a list of every person you have ever worked with who has the problem you solve, and texting them, not emailing, texting, because you have their number and the email would be too formal and they would file it under "later." It is the founder who goes to the meetup that has nothing to do with software, the regional grocery-buyers' lunch, because that is where the customers physically are and nobody else selling to them is willing to fly to Columbus. It is reading 200 forum posts to find the 11 people describing your exact problem in their own words, then writing each of them a paragraph that proves you read it.

I once watched a founder get her first 30 restaurant customers by walking into kitchens at 3pm, the dead hour between lunch and dinner, with a clipboard and a free trial. No deck. She closed roughly 1 in 4 because she was standing there and the owner was bored and a real person is hard to say no to. Her competitor was running ads. The competitor is gone.

This is slow. It does not compound the way a viral loop is supposed to. You cannot put it in a board update without feeling like you have nothing to show. And it is, almost always, the entire game for the first stretch. The companies that win early are usually the ones whose founders were willing to do something that felt beneath a "real" company, and the ones that die quietly are often the ones who insisted on looking like a real company before they were one. They hired a head of growth at eight customers. They paid an agency to run paid acquisition against a 0.4% conversion rate and called the data "early signal."

The unscalable phase has an expiration date

Now, the honest part, because the founder essays that romanticize doing things by hand usually forget to mention when to stop.

Hand-to-hand is the job until two things are both true. First, you can describe the customer who buys and stays in one sentence, without hedging, because you have seen the same person say yes enough times to know the shape of them. Second, you are turning away demand, or losing it, because you personally cannot get to all of it. Not "it would be nice to scale." You are physically the bottleneck and you have proof people want through the door.

Until both of those are true, a repeatable engine is premature, because you would be making a machine that reliably produces the wrong customer. After both are true, staying in the trenches becomes its own kind of avoidance. I have seen that too, the founder who loves the dopamine of closing every deal personally and quietly resists building anything that would make them less essential. That is not grit. That is a different way to cap the company at the size of one person's calendar.

The tell that you have crossed over is boring and specific. The same objection shows up in the same words on the fifth call. The same three sentences close the deal every time. The same kind of customer churns for the same reason. When the pattern is that legible, you have earned the right to build something repeatable, and only then does the marketing apparatus, the positioning, the content, the actual funnel, start working, because now it has real demand to organize instead of imaginary demand to manufacture.

The mistake is not doing marketing. The mistake is doing it first, on an audience that does not exist, to avoid the conversations that feel too small and too personal to count. They are the only thing that counts. Go send the eight messages. You already know who the eight people are.