Why nobody uses my app — three causes that look identical

You look at "5 users" and think: I need ads.
Maybe. Or maybe your signup is simply broken — and you can't see it, because you've been logged into your own app since day one and have never once gone through the entry flow as a stranger.
That's the whole difficulty with the symptom "nobody uses it". There's one symptom and three diseases behind it, and they're treated in opposite ways. Pour traffic into a broken signup and you'll spend money to lose people faster.
First, split the symptom into three
Don't fix anything until you've answered three questions in order:
- How many people actually saw the page?
- How many of them got through to the first useful action?
- How many of those came back at least once?
Three numbers. Then look for where the zero is — that's your disease. If you don't have those numbers, tagging events is today's job: without them you're fixing blind.
One correction before counting: exclude yourself. Your own visits, your phone, deploy previews. In a fresh project half the stats are the author, and they give a false sense of life.
Cause 1. They simply never got to you
The most common and the most boring. The product is fine; nobody knows it exists.
How to check. Open your analytics and look at sources. If nearly all visits are direct or from a messenger, you have no channel at all: the only visitors were the people you sent the link to yourself. Check search separately — open a private window and search your name. Not there? That's normal for a new site, but it means search won't bring you people for months, so don't plan around it. If search was your plan, see why your site isn't in Google.
How to fix. Not "launch ads" — do the one thing people postpone for months: go where people already complain about the problem you solve. Topic chats, forums, comments under other people's posts. Not with an ad — with an answer to a concrete question and a link at the end. Ten such answers bring more first users than a week of configuring an ad account, and they show you the words people use for their pain.
An honest calibration: distribution isn't a phase after development, it's half the work. Shipping is the easy half.
Cause 2. They got there and broke at the door
The most galling one, because it takes an hour to fix and lives for months.
An author almost never walks through their own onboarding again. They're logged in, their database is populated, their email is whitelisted. So the classic breakages live on unnoticed:
- the code email lands in spam, or never sends at all;
- social login redirects to
localhost— nobody changed the callback URL for production; - the form doesn't submit on mobile, and 70% of people came from a phone;
- after signup the person lands on an empty screen with no hint and no idea what to do.
How to check. Take someone else's phone, someone else's email, a private window — and walk the whole path from link to first result as a stranger. Time it. Anything that took more than a couple of minutes or required a guess is where you're losing people.
The second method is more precise: hand the phone to a real person, say "do this", and stay silent for two minutes. Staying silent is hard. Those two minutes show you everything.
How to fix. Build a four-step funnel and find the transition that loses the most. Fix that one first — those people are already motivated, they want this, and if they still leave you're blocking them with something specific. And remember the arithmetic: removing an unnecessary step almost always beats improving an existing one.
Cause 3. They tried it and didn't come back
The hardest, because it's no longer a question of interface but of need.
How to check. Compute retention: of the people who signed up last week, how many showed up at least once on day seven. Then look at the shape of the curve by day. If it falls and settles onto a shelf, you have a core and you just need to grow. If it crawls to zero and doesn't slow, there is no core, and traffic won't fix that.
How to fix. No amount of analytics will give you the reason — numbers only show the place. The reason comes from talking to a real person, and you ask about the past, not the future: not "would you use this" but "walk me through the last time you dealt with this". Five such conversations with people who left are worth more than any dashboard.
What not to do
- Don't add features to win people back. They didn't leave because something was missing; they left because they didn't see the value in the first minute. A new feature only makes that minute longer.
- Don't buy traffic before you've found the zero. Ads multiply what you have. Multiply zero and you get zero, but paid for.
- Don't watch an overall percentage instead of groups. It rises on its own when you pour in new people and falls when you stop. That's not about the product — more on that in cohorts.
How many people before I can conclude anything?
For numbers — at least a couple of hundred at the first step, or percentages will swing on one person. For conversations, five is enough: qualitative conclusions arrive far earlier than quantitative ones. While numbers are small, watch individual sessions, not averages.
Maybe the product is just not needed?
Sometimes, yes. But test that third, not first: make sure signup works and that people reach the value at all. Far too many products were buried over a broken verification email. If the retention curve still crawls to zero after you fix the door — then yes, you're solving the wrong problem, and it's better to learn that now.
Should I relaunch for a different audience?
Before switching audiences, look hard at the people who did stay, even if there are three of them. They usually share something — a job, a task, the route by which they found you. That's cheaper than guessing a new audience from scratch: you already have three pieces of evidence about who this fits.
The AI Architect 2027 course: step by step from computer basics to your own app and first customers. 543 short lessons on your phone.





