How to get first users for your app

“Users” and “customers” get used interchangeably, and conflating them is how founders end up with a thousand signups, a dead retention curve, and no idea what went wrong.

A customer has decided your product is worth money. A user has decided it’s worth a bit of attention. Those are different bars, and they need different approaches.

Getting users is a habit problem

Nobody adopts a new tool because it’s good. They adopt it because it slots into something they were already doing, at a moment they were already doing it.

You’re not competing with other apps. You’re competing with the thing they do instead, which is usually nothing.

So the question isn’t “how do I reach more people” but “where is someone already mid-task and mildly annoyed?” That moment is the only reliable entry point.

Give before you ask

The highest-converting thing most early apps can do is make a piece of their value available without a signup at all. A free tool, a calculator, a single-use version, a public example. Something that works in one click and demonstrates the thing.

This feels like giving away the product. It isn’t. The signup wall isn’t protecting value; it’s preventing you from finding out whether anyone wants it.

Choosing what to give away

Not a trial, and not a crippled tier. The thing you want is one complete useful outcome, delivered in a single interaction, with no account.

Three tests for whether you have picked the right one:

• It finishes. Somebody arrives, does one thing, and leaves with a result they would have paid for. A preview that stops short of the answer is worse than nothing, because it converts curiosity into irritation.
• It is shareable on its own. The result, not the product, is what somebody sends to a colleague. That is your distribution, and it only works if the artefact makes sense without you explaining it.
• Repeating it is the paid thing. One is free because one demonstrates the judgement. The loop is the product. If your product has no loop, giving one away costs you nothing and you should do it today.

Where the first hundred actually come from

• Communities where the problem is discussed, not where your users nominally hang out, but where the specific annoyance gets posted.
• People you already know who have the problem. Ten personal messages beat any broadcast at this stage, and you get real answers back.
• One genuinely useful public artefact. A teardown, a dataset, a tool. Something linkable that exists whether or not anyone signs up.
• Being visibly present where the problem lives, for longer than feels comfortable. Most founders quit at week three.

The number that decides whether any of it counts

Day-two return. Not signups, not visits, not upvotes.

If a hundred people sign up and four come back tomorrow, more acquisition is actively harmful. You’re spending effort filling a bucket that empties, and you’re burning the communities you posted in while you do it.

Measuring it with no analytics stack

You do not need a product analytics tool for this, and installing one is a common way to spend a week not doing the thing.

You need one timestamp: the last time each account did the core action. One column in your existing table, updated on the action. From that, every question worth asking at this stage is a query: how many accounts created four weeks ago acted in the last seven days.

Reading a curve at these numbers

With forty users, be honest about what can and cannot be concluded. Two people returning out of nine is not a nine percent retention rate, it is two people, and the confidence interval on that is wide enough to drive through.

What is still readable, even at this size:

• Zero is meaningful. Nobody at all returning, out of forty, is a finding, and no larger sample will improve it.
• Flat beats declining. If a small group keeps coming back week after week and the number is not shrinking, you have something, however small.
• Names beat percentages. At this scale, go and look at who the returners are. If they share a situation, that is your positioning, and it is the most valuable output of the entire exercise.

The first sixty seconds

Retention at this stage is decided almost entirely in the first minute of first use, and it is the least examined minute in most products.

• An empty state is not a neutral state. A dashboard with nothing in it tells a new user that the product does not work yet. It has to say what to do, in one line, with one control.
• Prefill something. Sample data, a demo project, one working example they can poke. The distance from signup to seeing the thing work is the number that matters, and blank screens make it infinite.
• Ask for nothing you do not need. Every field before the first useful moment costs you a share of the people who would have stayed.
• Watch three people do it, in silence, without helping. This is unpleasant and it is worth more than any amount of instrumentation.

When users should become customers

There’s a common trap here: founders delay charging because they’re “still growing the user base”. Free users answer a question you’ve mostly already answered by a few hundred, and stop telling you anything new.

If your plan is to charge eventually, find out early whether anyone would. Getting your first paying customers is a different exercise and worth starting sooner than feels comfortable.

What not to do

• Buy installs or run growth hacks. You get numbers, not information.
• Post the same link in twenty communities. One coordinated spam run costs you all twenty.
• Chase a viral moment. Unrepeatable, unattributable, and it teaches you nothing you can do again on Tuesday.
• Add features to fix retention before asking the people who left why they left.

The short version

Find where the itch already is. Give something useful without a wall. Watch day two. Fix whatever day two tells you. Then do it again, and if you’ve just shipped and haven’t started any of this, how to market an app is the place to begin.

If you have users and the line has gone flat, reading a flat line is the next diagnosis. If nobody is arriving at all, launched and nobody came splits that case in two. Or paste your URL into the Product Check and have the reading done against your actual product. It is free and needs no account.