How to market an app

The app works. You paid for the domain, fixed the mobile layout, and wrote the onboarding email nobody has received. Analytics shows four visitors, three of which were you.

Here’s the part nobody warns you about: building had instructions. There was a docs page, a Stack Overflow answer, a model that would write the migration for you. Marketing has none of that. It has a thousand posts saying “build in public” and none saying what to do on Tuesday.

First: fix the page, before you fill it

Before sending anyone anywhere, read your own homepage as a stranger with three seconds and no context. Does it say what breaks and what you fix? Or does it say something like “the modern platform for teams”?

Every visit you generate gets multiplied by whatever that page already does. If it converts at zero, a brilliant forum reply produces nothing, and worse, you conclude the forum doesn’t work and go somewhere else.

If nobody understands the homepage, buying more traffic just lets more people misunderstand it.

Writing the first line, mechanically

You do not need to be a copywriter. You need to answer two questions in one sentence: what breaks, and for whom. Mechanism last, or not at all.

• Category noise: “A modern platform for collaborative workflows.” True of nine thousand products.
• Mechanism first: “Event-sourced task graph with realtime conflict resolution.” Impressive, and it answers a question nobody asked.
• Problem first: “Your team stops overwriting each other’s work.” Recognisable in one read by exactly the person who has that problem.

The third one is worse writing by most measures and it converts, because a visitor is not evaluating your prose. They are scanning for whether this is about them.

The reliable source for that sentence is not your head. It is the words somebody used when they described the problem to you, before they knew you had built anything.

Second: stop looking for the channel

The instinct is to pick one. SEO, or X, or Reddit, or ads. Wrong first question, and it burns months.

The right one is: who has this problem badly enough to go looking for a fix? Not who might like your app. Who is, right now, annoyed enough to type something into a search box or complain about it somewhere public.

People don’t search for your category

This is the distinction that decides everything. Someone building a tool for golf societies will spend weeks chasing “golf society management software”. A phrase used almost exclusively by competitors. Meanwhile, society organisers are posting every week about losing an evening to a spreadsheet and a WhatsApp group full of scores.

Those are the same people. Only one group is findable, and it isn’t the one using your category name.

How to find where the complaint lives

This is a search problem with a specific trick: search the language of annoyance, not the language of the category. Half an hour, and it works for almost any product.

• Write down the complaint in their words. Not “inefficient scheduling workflow”. “Spent the whole evening chasing people for availability.”
• Search fragments of it, scoped to where people talk. Add a site restriction to your search engine, or use the search inside the forum itself. The fragments that work are the awkward ones: “anyone found a better way to”, “still doing this in a spreadsheet”, “how do you all handle”, “is there anything like X but”.
• Sort by recent, not by relevance. A thread from 2019 is research. A thread from last Tuesday is a person you can help today.
• Read the replies, not just the question. The tools people recommend to each other are your real alternatives, and they are usually not the competitors you had in mind.
• Keep the list. Ten threads with links and dates. This becomes the raw material for every week after this one.

If half an hour of this turns up nothing at all, that is the single most important finding available to you, and it is not a marketing problem. Either the problem is not painful enough to discuss, or you are searching in your vocabulary rather than theirs. It is almost always the second.

Third: be useful before you’re relevant

The fastest way to burn a community is to reply with a pitch. People smell it instantly, moderators ban for it, and you get roughly one shot per place.

The reply that works answers the question properly, genuinely, in detail, as though you had no product, and then mentions what you built in one sentence at the end, where it’s relevant.

What to skip in month one

• App store optimisation, if you have no installs to optimise. ASO ranks what already has signal.
• A launch. One non-repeatable spike, aimed at a page with no evidence anyone has used this. You get one.
• Paid ads. Paying to test an unvalidated page.
• An SEO content plan. Six-month payback on a this-month problem.
• Another feature. You already know.

Your first fortnight, concretely

Marketing advice fails on Tuesday because it never says what Tuesday contains. Here is a version that does, assuming a few hours a week.

• Days one and two. Rewrite the first two lines of your homepage. Send it to three people who do not know what you built and ask what it does. Rewrite again using their wrong answers.
• Day three. The half hour of complaint-searching above. Produce the list of ten threads.
• Days four to nine. Two genuinely useful replies a day, from the list. No link in the first reply unless somebody has literally asked for a recommendation.
• Day ten. Read what happened. Visits, replies, signups, or nothing. Write it down.

What “nothing happened” means

It usually means one of four things, and they have different fixes, so it is worth telling them apart before concluding the channel is dead.

• Nobody saw it. Wrong room, or posted into a dead thread. Check whether anything in that room gets replies at all.
• They saw it and did not click. Your one sentence about the product is the problem, not the room.
• They clicked and left. The page is the problem. Back to the top of this guide.
• They signed up and vanished. Not a marketing problem at all, and no amount of additional traffic will improve it.

Four different diagnoses that all feel identical from where you are sitting, which is exactly why writing down what happened is worth more than doing more of it.

The bit that actually compounds

Write down what you did and what it produced. Visits, signups, customers, or nothing. Nothing is a result too, and it’s the one most founders throw away.

Do that for a month and you’ll know something no article can tell you: which channel works for your app. That’s the point at which marketing stops being guesswork.

Users or customers?

Worth being clear which you’re chasing. The playbooks differ more than people expect. Getting first users is a volume-and-habit problem; getting first paying customers is a value-and-trust problem. Confusing the two is how founders end up with a thousand signups and no revenue.

If you have already done the launch and the line went flat, launched and nobody came picks up from there. If your users are developers, most of the advice above inverts. And if you want this applied to your own page rather than in general, the Product Check reads your site and is free.