To get customers for a new app you have to win two separate fights, and almost everybody only fights the first one. You need to be found inside somebody else’s search box, and then you need to survive the first session after the download. The second one gates the first, because the stores rank you partly on what people do once they open you.
That’s why the number on your dashboard right now is probably the one number that will mislead you. There’s a second number sitting a click away from it that tells you the truth, and once you’re watching that one instead, most of the decisions get easy.
Why It’s Harder to Get Customers for a New App Than for Almost Any Other Business
Every other business shape in this cluster owns its own front door. A shop has a street, a service business has a phone, an online store has a website you control. An app has none of that, and the difference isn’t cosmetic.
Somebody else stands between you and your buyer. Apple and Google decide whether you appear, in what position, and next to whom. You don’t set the rules and you can’t appeal them.
You don’t get to know who downloaded you. The store keeps the customer relationship. No email, no name, no way to follow up, unless the person hands you those things inside your own app.
The unit everyone counts isn’t a customer. An install is a free download by someone who has committed nothing. Across the panels published by Adjust, AppsFlyer, and Sensor Tower, cross industry Day 1 retention runs around 25 to 26 percent and Day 30 lands somewhere near 5 to 7 percent, with wide variation by category. So roughly one in twenty of your downloads is still there a month later.
The store watches that number too. Both stores factor engagement and uninstalls into how they rank and recommend you, which means a leaky first session quietly throttles your distribution.
Put those together and the sequencing writes itself. Fixing what happens after the download isn’t a phase two nicety, it’s the thing that decides whether phase one is worth doing at all.
If your app is the delivery mechanism for a subscription business rather than a product in its own right, the churn and payback economics behind all this live in how to get customers for a SaaS or startup. What follows is the part specific to living inside a store you don’t own.
Step 1: Replace the Install Count With a Number That Means Something
The install count is the number that will mislead you. It goes up when a campaign runs and it never goes down, so it always looks like progress.
The number sitting next to it is the one that matters: how many people did the main thing. Not opened the app. Did the thing your app exists to do, once, on purpose. Booked the appointment, logged the workout, sent the invoice, placed the order.
Pick that single action and give it a name today. Then track three numbers every week:
- Installs, so you know how much attention you’re getting.
- Activations, meaning people who completed that main action at least once.
- Week four actives, meaning people still doing it a month later.
The gap between the first and the second is your onboarding problem, and the gap between the second and the third is your product problem. Those are two completely different fixes, and until you separate them you’ll keep spending money on the wrong one.
The other reason to define this early is that it’s the only way to know what an install is worth. Working out what you can afford to spend on a customer needs a real customer at the end of the arithmetic, and an install isn’t one.
Step 2: Win the First Session Before You Chase Another Download
Somebody downloaded your app, opened it, and had a few seconds of patience. What happens in that window decides everything downstream, and it’s the cheapest thing on this entire list to fix.
The pattern that loses people is always the same. The app opens on a sign up wall, asks for permissions it hasn’t earned, explains itself across four tutorial screens, and delivers nothing the person can feel.
Turn it around:
- Show something real before you ask for anything. Let people see the app working with sample data if you have to.
- Delay the account. Ask for a login at the moment it becomes necessary, not at the door.
- Ask for permissions in context. A location prompt makes sense when someone taps “find near me” and nowhere else.
- Get the person to the main action inside the first minute, with as few taps as you can manage.
- Kill the tutorial. If your app needs four screens of explanation, the explanation isn’t the problem.
Then watch what actually happens rather than what you hope happens. Both stores and every analytics tool will show you where people stop, and the drop off is almost never where the team guesses.
The arithmetic here is worth sitting with. If 100 people install and 30 reach your main action, you’re losing 70 people you already paid for, before the product has had a chance to be good or bad.
Move that 30 to 45 and you’ve grown your real user base by half without a single extra download. No channel, paid or organic, gives you a 50 percent lift for the price of one release, which is why this step sits second and not last.
Step 3: Spend Your 160 Searchable Characters on Apple Carefully
Here’s the fact that changes how you write an App Store listing, and most owners have never been told it. Apple indexes exactly three fields for search: the app name at 30 characters, the subtitle at 30, and a hidden keyword field at 100. That’s 160 characters, total, per language.
Your 4,000 character description does not feed App Store search at all. Neither does your promotional text. They sell to a person who already found you; they don’t help anyone find you.
So treat those 160 characters like the scarce inventory they are:
- Put your strongest search term in the name, next to your brand.
- Use the subtitle for the second term plus the reason someone would pick you.
- Fill the keyword field with single words, comma separated, no spaces after the commas, and never repeat a word you already used in the name or subtitle.
- Skip plurals, skip your own brand name, and skip category words Apple already knows.
- Leave out anything you can’t defend, because a term you rank fortieth for sends you nothing.
Write those three fields as one budget rather than three separate boxes, because that’s how the store reads them. Every wasted word in the keyword field is a term you could have owned and didn’t.
Here’s what spending it well looks like. Say you built a scheduling app for barbers, and you call it “Trimly: Barber Booking”. That uses 23 of your 30 name characters and buys you barber and booking.
The subtitle “Appointments and reminders for shops” spends 36, so it needs trimming to fit 30, and it adds appointments, reminders, and shops without repeating a single word from the name. That leaves the whole 100 character keyword field for terms neither field touched: haircut, salon, chair, walkin, client, schedule, waitlist, deposit, noshow, calendar. Ten more terms, no duplication, every one of them plausible for a small app to rank on.
Compare that with the listing most owners write, which puts the company name in the title, repeats “barber booking app” in the subtitle, and leaves the keyword field half empty. Same 160 characters, maybe a third of the coverage.
Step 4: Write the Google Play Listing Like a Web Page
Google Play works the opposite way, and copying your Apple listing across is the mistake almost everyone makes.
Play indexes the app name at 30 characters, the short description at 80, and the full description at 4,000. That’s roughly 4,110 indexed characters against Apple’s 160, about twenty five times more room, and Google reads it the way it reads a web page.
Which means the Play description is a writing job, not a metadata job:
- Lead the short description with the search term and the promise, because it’s the highest weighted field after the title.
- Write the long description in real sentences, using your main terms naturally two or three times each.
- Cover the questions people ask before installing, in their words.
- Use short paragraphs and subheadings, because it’s read on a phone.
- Update it when you ship something meaningful, the same way you’d update a landing page.
The practical upshot is that you maintain two different listings for the same app and you stop feeling guilty about it. On Apple you’re rationing characters, on Play you’re writing copy, and a listing tuned for one store is quietly underperforming in the other.
Step 5: Fix the Product Page Before You Buy a Single Install
Both stores show you a funnel, and it’s the most useful free data you have. Impressions, product page views, then installs. App Store Connect and Play Console both report it.
Run your own numbers and the priority becomes obvious. Say 10,000 impressions produce 500 page views and 100 installs. That’s a 20 percent page conversion. Lifting it to 30 percent gives you 150 installs from exactly the same traffic, half again as many, for the cost of new screenshots.
Almost all of that lift lives in three places:
- The icon. It’s the only thing visible in search results, at about the size of a thumbnail.
- The first two screenshots. Most people never swipe past them, so those two carry the whole pitch.
- The rating. It sits under your name in every list you appear in.
Design the first two screenshots as a headline and a proof, not as a gallery of your interface. Put a short caption above each image that says what the person gets, use high contrast text large enough to read on a phone, and show the outcome rather than the settings screen. Apple and Google both let you test alternatives against your live listing, so you can measure this instead of arguing about it.
Step 6: Earn Ratings Without Nagging Anyone
Your star rating is doing two jobs at once. It’s the single biggest trust signal on your product page, and it feeds how the stores rank and recommend you. A new app with eleven ratings and a 3.4 average is fighting with one hand tied.
The mechanics matter more here than in most channels, because the platform limits you. Apple caps the in app review prompt at three requests per customer per year and enforces it in the system, so you get three shots and they’re gone whether you used them well or not.
Use them at the right moment:
- Ask right after somebody completes the main action successfully, never on launch and never mid task.
- Ask people who’ve come back at least twice, not first timers.
- Never ask right after an error, a payment, or a permission prompt.
- Route unhappy people to a support form first, so a fixable complaint doesn’t become a one star.
The broader craft of asking, and what makes people actually respond, is covered in how to get more online reviews from happy customers. The app specific part is simply that you have three prompts a year per person, so spend them on people who just had a good moment.
Step 7: Only Then Pay for Installs
Paid installs look like the obvious accelerator and they’re the fastest way to waste a small budget, for a reason the arithmetic makes plain.
Published 2026 benchmark sets disagree wildly on what an install costs. They put global iOS somewhere between roughly $2 and roughly $6, and Android between roughly $1 and $2, with North American iOS commonly quoted at the top of that range.
Take a middle figure and run it through the retention numbers from earlier.
Buy 1,000 installs at $3 each and you’ve spent $3,000. At a benchmark 5.4 percent Day 30 retention, 54 of those people are still using the app a month later. That’s about $56 per retained user, roughly eighteen times the price on the invoice.
That multiplier isn’t an argument against paid acquisition. It’s an argument for doing it in this order, because every point of retention you add before you start divides that number:
- Fix onboarding first, since it moves the denominator.
- Fix the product page second, since it lowers the effective cost per install on the same spend.
- Start with search ads on terms where intent is highest, rather than broad discovery placements.
- Set a test budget you’d be comfortable losing entirely, and judge it on activations rather than installs.
The general question of whether you’re ready to pay for customers at all is worked through in when to start paying for customers, and the answer for an app is usually later than the founder wants.
What to Do in Your First 30 Days
Work in this order. Each step makes the next one cheaper.
- Week one, define the main action. Name it, instrument it, and start reporting installs, activations, and week four actives side by side.
- Week two, rebuild the first session. Remove the sign up wall, delay the permissions, get people to the main action inside a minute.
- Week three, rewrite both listings separately. Ration your 160 characters on Apple, write real copy on Play, and redo the icon and first two screenshots.
- Week four, turn on the rating prompt properly. Trigger it after a successful main action, for returning users only.
If you’re not sure which problem you have, the numbers will tell you. Plenty of installs and almost no activations means your first session is broken, and no amount of traffic will fix it.
Plenty of activations and no installs means you’re invisible in search, so the listing work in steps three and four is where your month goes. Almost no impressions at all means your keyword targets are too competitive, and you need narrower terms you can genuinely win before anything else matters.
Frequently Asked Questions
How do I get people to download my app?
Being found in store search does most of the work, and that comes down to targeting terms you can realistically rank for and then converting the people who land on your page. Fix the icon and your first two screenshots before anything else, because they decide whether a page view becomes an install. Paid installs work but cost far more than they look like they do. Sequence the free work first.
Why is nobody using my app after they download it?
Almost always the first session. If your app opens on a sign up wall, asks for permissions it hasn’t earned, or takes more than a minute to deliver something useful, most people leave and never come back. Cross industry benchmarks put Day 30 retention at roughly 5 to 7 percent, so some loss is normal. Watch where people stop and fix that one screen.
How much does it cost to get an app install?
Published 2026 benchmarks disagree a lot, putting global iOS anywhere from about $2 to about $6 and Android from about $1 to $2, with North American iOS at the top end. The number that matters is not the install price though, it’s the cost per person still using the app a month later. At benchmark retention that figure runs close to eighteen times the install cost.
Is app store optimization worth it for a small app?
Yes, and it’s the highest return work available to you because it costs time rather than money. You get 160 indexed characters on Apple and roughly 4,110 on Google Play, and most small apps waste both. Rewriting your name, subtitle, keyword field, and screenshots is a weekend of work that keeps paying. Do it before you spend anything on ads.
Do I still need a website if I have an app?
Yes, for two reasons. It’s the only place you control where somebody can learn about the app, and it’s the only way to collect an email address, since the stores will never tell you who downloaded you. A single clear page with screenshots, an explanation, and store links is enough. Without it you have no audience you actually own.
How do I get more app ratings and reviews?
Trigger the in app prompt right after somebody completes the main action successfully, and only for people who’ve returned at least twice. Apple limits you to three requests per customer per year and enforces it, so timing is the whole game. Never prompt after an error or a payment. Send unhappy users to a support form first so a fixable problem doesn’t become a one star.
An app is the only business shape where somebody else owns your front door, your search results, and your customer list. You can’t change that, but you can control what a person meets in the first minute after they open you, and that single thing feeds the rankings, the ratings, and every dollar you’ll ever spend on traffic.
So stop watching the install count and start watching the number of people who did the main thing this week. If you want a second pair of eyes on which of your numbers is actually broken, send us the question and a real person will work through it with you.
Previous 
