Do Things That Don't Scale to Get Your First Customers

There is a specific voice in your head when you are a solo builder trying to get customers, and it says things like "this won't scale." You are about to reply to a signup with a personal email and the voice says a real company would automate that. You are about to hop on a call to walk one person through setup and the voice says you should build better onboarding instead. So you close the tab and go write code, because code feels like leverage and answering one human feels like a rounding error.

That voice is right about the future and completely wrong about right now. The unscalable work is not a distraction from getting customers. Early on, it is the entire way you get them.

Where does the money actually come from?#

The money comes from a specific person crossing the line from interested to paying, and early on that line only gets crossed when a human helps them across it. Look at where the effort actually goes:

A stranger with the problem you solve
        |
        v
You find them yourself and reach out personally  <-- no reach to do this for you
        |
        v
You help them try it, hand-holding through setup  <-- no onboarding flow yet
        |
        v
They hit the moment where it obviously works      <-- you dragged them to it
        |
        v
They trust you because a real person showed up     <-- no reputation to lend trust
        |
        v
They pay, and tell you exactly why they almost didn't
        |
        v
Revenue + a lesson about what to automate next

Every arrow in that chain is manual because you have nothing to make it automatic yet. A scaled company gets people to the "it works" moment with slick onboarding built from thousands of past users. You do not have thousands of past users. You have you, and you can personally carry a handful of people to that moment right now. That is where the first dollars come from. Not from a funnel. From you.

How it actually works#

The core move is simple to say and uncomfortable to do: for your first customers, be the automation.

Do not build a self-serve onboarding flow. Get on a call and set the product up with them while they watch, or set it up for them and hand it back working. Do not write a help center. Answer the one question the one customer has, in a reply that arrives in minutes. Do not build a drip email sequence. Write the person a real email that references the actual thing they are trying to do. Do not wait for people to find you. Go find them and start the conversation yourself, the way getting your first ten customers lays out in detail.

This feels wrong to a builder because the whole appeal of software is that it works without you. But the product is not the point yet. Learning is the point. When you personally onboard someone, you see the exact place they get confused, the feature they expected that is not there, the word on the button that means nothing to them. You cannot get that from analytics on a page nobody visits. You get it by sitting in the discomfort of watching a real person struggle with the thing you built.

There is a hidden bonus. When a founder shows up personally, people buy who otherwise would have drifted off. A stranger who gets a thoughtful, fast, human reply feels something a support ticket never delivers: that someone is actually behind this thing and cares whether it works for them. That feeling closes sales. It is a real advantage you have over big companies, and it evaporates the moment you decide you are too important to answer your own email. This is founder-led selling, and it stays the job well past the first handful, which founder-led sales through your first hundred conversations covers.

The manual work also doubles as research. Every hand onboarding is a live user interview whether you planned it or not. If you want to be deliberate about pulling signal out of these moments, how to talk to users is worth reading before your next few calls, because most builders talk too much and learn nothing.

A clearly hypothetical example#

Here is an invented scenario to show the shape of it. The numbers are made up to illustrate the trade, not a promise. Yours will differ.

Say you built a tool for a narrow group and you have two ways to spend the next two weeks.

Option one, the scalable instinct. You spend the two weeks building a polished self-serve onboarding flow, tooltips, a welcome sequence, a help center. At the end you have a product that could, in theory, onboard a thousand people without you. You launch it to your zero-person audience. A handful of signups trickle in, poke around alone, get confused somewhere you cannot see, and leave. You end the two weeks with a beautiful funnel and, hypothetically, one paying customer, and no idea why the others left.

Option two, the unscalable path. You spend the two weeks finding fifteen people who have the problem and personally walking each one through the tool. It is slow and a little draining. But say, hypothetically, six of them pay because you were right there when it clicked, and every single one told you something: three got stuck on the same setup step, two wanted an integration you did not have, one only paid because you set it up for them. You end the two weeks with six customers and a to-do list written by the market itself.

Same two weeks. Option two produced more revenue and, more importantly, told you exactly what the scalable onboarding should do when you eventually build it. That is the whole argument. The manual work is not slower to revenue. It is faster, and it comes with a map.

What you need (required vs optional)#

Required:

• A product that does one useful thing well enough that a real person, helped by you, would pay for it.
• Willingness to do work that feels beneath a "real" software company: manual setup, personal emails, jumping on calls.
• A way to reach individual people directly and a way to take payment. Both can be free and basic.

Optional but helpful:

• A simple way to record what you learn from each onboarding, even a plain notes file. The patterns across ten sessions are your product roadmap.
• A short screen recording of you setting the product up, so you can send it when a live call is not possible.
• A running list of the manual tasks you keep repeating. That list is the spec for the first thing you should automate.

What it costs#

The cost is entirely your time and your ego. There is no tool to buy. What it actually asks of you is the willingness to do repetitive, personal, unglamorous work while a quieter part of you insists you should be coding something clever instead.

The real risk is not the time. It is quitting the manual work too early because it feels inefficient, automating before you have learned what to automate, and shipping a self-serve flow built on guesses. That flow will convert worse than you personally would have, and you will not know why, because you skipped the part where the customers told you.

How long it takes#

Plan on the manual phase covering your first ten to fifty customers. There is no fixed number of weeks, because it depends on your segment and how many conversations you start. The signal that you can begin replacing manual work with automation is not a calendar date. It is repetition: when you have onboarded enough people that you can predict where they will get stuck before they do, you finally know enough to build the flow that handles it. Automating before that point is guessing. After it, it is engineering.

What beginners usually get wrong#

The biggest mistake is treating "this won't scale" as a reason not to do something, when at this stage it is often a reason to do it. Scale is a problem you earn. You do not have it yet, so optimizing for it is optimizing for a future you may never reach if you cannot get the first customers.

The second mistake is building automation as a way to avoid talking to people. It is disguised procrastination. Writing an onboarding flow feels productive and lets you never face a live human who might not like your product. The discomfort is the signal that you are near the thing that actually matters.

The third mistake is hiding behind the product. "I'll just make it self-serve so I don't have to sell." Self-serve is what you build after you understand the sale, not instead of understanding it. If you build it first, you are automating a conversion you have never once made by hand.

The fourth mistake is quitting manual onboarding the moment it gets tedious, right before the patterns become clear. The tedium is the data collecting. Push through enough sessions that the repeated problems start to obvious themselves, then build. What you learn here feeds directly into real onboarding that turns signups into customers later.

How I would start#

1. Pick the next person who could plausibly use the product and reach out to them personally, by name, referencing their actual situation. One human, not a broadcast.
2. Offer to set it up with them or for them, on a call or over messages, rather than sending a signup link and hoping.
3. Watch or listen for every moment of friction and write down the exact thing that confused them, in their words.
4. Do whatever manual work it takes to get them to the moment the product clearly works for them, even if that means doing part of their job once.
5. Ask them plainly what almost stopped them from paying, and who else they know with the same problem.
6. Repeat with the next person, and keep a running tally of the problems that show up more than once.
7. Only when a manual task has repeated enough times that you can predict it, build the thing that automates that one task, and nothing more.

The bottom line#

The advice to do things that do not scale confuses builders because it sounds like permission to be inefficient. It is the opposite. When you have no reach and no reputation, manual effort is the only lever you have, and applied one customer at a time it converts strangers, closes sales that automation would lose, and hands you a precise list of what to build next. It ends after the first ten to fifty customers, because by then the work will have taught you what to replace and why. Until then, be the automation. If you want the hands-on version of finding and closing those first people, our first customers walkthrough turns this into a concrete plan you can start today.