Building an app used to be the hard part. It is not any more.
Somebody with no coding background can describe an app in plain English and have a working build the same afternoon. The tools got good, and they got cheap. That is genuinely a good thing.
It also means the store is filling up faster than it ever has, and almost none of what arrives finds users. We analysed 3,037,501 public Google Play listings to see how that plays out.
How many apps are published every day?
About 3,100 a day, and rising.
93,320 apps were first released in June 2026 alone. That is the largest full month in our data. In January it was 51,612, so the rate nearly doubled inside six months.
Part of that is AI. Apps with AI in the name went from 0.29% of releases in 2022 to 3.26% in 2026, which we broke down in our report on AI apps. But the bigger effect is quieter. AI did not just create AI apps. It lowered the floor for building any app at all.
So the question worth asking has changed. It is no longer whether you can build it. It is whether anybody needs it.
What happens to an app after it is published?
Usually nothing. This is the part nobody plans for.
Read that from the bottom. Only 7.2% of apps ever reach 100,000 downloads. That is 218,308 apps out of more than three million.
It gets starker higher up the chart. More than a third never reach 100 downloads. Almost two thirds never reach 1,000. The middle app on Google Play has 305 downloads in total.
None of that is a punishment for building a bad app. It is the base rate. Obscurity is what happens by default.
Does marketing fix this?
It helps. It does not fix it.
Marketing decides how many people see your app. It does not decide whether they keep it. If somebody installs your app, opens it once, and finds nothing they needed, you have paid to acquire a person who leaves.
Spending on ads before the app is worth keeping just makes you lose money faster.
The thing that decides whether an app survives is much less glamorous. The app has to solve a problem somebody actually has.
What makes an app worth keeping?
One question, asked honestly.
What problem does this solve, and for whom?
If the answer takes a paragraph, or needs the word “and” three times, that is a warning. The apps that work solve one problem clearly enough to say in a sentence.
Three failures we see over and over.
- The app solves a problem nobody has. It was built because it was buildable, not because anybody asked.
- The app solves a real problem badly. Something free and already installed does it better.
- The app solves it once. There is no reason to open it a second time, so it gets uninstalled by the weekend.
What to check before you build
Five questions worth answering before you write any code. None of them need a budget.
- Can you name a real person with this problem? Not a market. A person you could message today.
- What are they doing about it now? Everybody solves their problem somehow, even if it is a spreadsheet. That is your real competition.
- Search the store for it. If ten apps already do this and none has a rating, that is a market telling you something.
- Why would they open it again on Thursday? If you cannot answer that, you have built a tool, not a product.
- What is the smallest version that helps? Ship that, then find out if you were right.
What to measure after you publish
Downloads are the number everybody watches, and the least useful one. Four that tell you more.
- Day 7 retention. Of the people who installed a week ago, how many are still opening it? This is the single honest measure of whether you solved anything.
- Uninstall rate in the first 48 hours. A high number here means the listing promised something the app did not deliver.
- Sessions per user per week. One is a tool somebody tried. Four is a habit.
- Whether you get a rating at all. Only 12.2% of apps on Google Play ever collect enough reviews to show a star rating. Reaching that point puts you ahead of seven apps in eight.
If day 7 retention is near zero, more marketing will not save the app. Go back to the problem.
Where testing comes into this
Google now requires 12 testers, opted in for 14 continuous days, before a new personal developer account can publish to production. Most developers treat it as paperwork.
It is the cheapest research you will ever get. It is twelve people using your app before the store judges it, and it costs you nothing but the asking.
Our own records show that 92.5% of the rejected apps that come to us failed because the testers were never active. Those developers cleared the count and learned nothing, which is a waste twice over.
Ask your testers one question at the end: would you keep this? The answer is worth more than the approval.
What this data cannot tell you
- This is our dataset, not the whole store. Google publishes no total app count, so every percentage is a share of the 3,037,501 listings we collected.
- We cannot see retention or revenue. Downloads and ratings are public. What happens inside an app is not, so the retention advice above comes from working with developers, not from this data.
- We cannot count AI-built apps. We can count apps with AI in the name. An app built with AI tools looks like any other listing.
- Recent weeks are under-counted. A listing has to exist before we can collect it, which is why July and August are left out of the monthly chart.
- These are associations, not causes. Apps that reach 100,000 downloads tend to be better made, but resources and timing explain part of that too.
The short version
Building is solved. Distribution never was.
Around 3,100 apps arrive every day, and roughly seven in a hundred will ever reach 100,000 downloads. You do not beat those odds with a better launch. You beat them by building something a specific person would miss if you took it away.
Frequently asked questions
How many apps are published on Google Play every day?
About 3,100. In June 2026, the largest complete month in our data, 93,320 apps were first released. That is up from 51,612 in January of the same year.
What percentage of apps get more than 100,000 downloads?
7.2%. That is 218,308 apps out of the 3,037,486 with a download count in our dataset. Only 2.2% pass a million.
Is it still worth publishing an app in 2026?
Yes, if the app solves a problem somebody has. The odds against a generic app are worse than they have ever been, but they were never good. The middle app on Google Play has 305 downloads.
Are AI-generated apps flooding Google Play?
Apps with AI in the name grew from 0.29% of releases in 2022 to 3.26% in 2026. The larger effect is indirect: AI tools made it easier to build any app, which is part of why monthly releases nearly doubled during 2026.
Does marketing help an app that nobody keeps?
Marketing controls how many people see the app. It does not control whether they stay. If day 7 retention is near zero, paid installs lose money faster rather than fixing the problem.
What should I measure after launching an app?
Day 7 retention, uninstall rate in the first 48 hours, sessions per user per week, and whether you ever collect enough reviews to show a star rating. Only 12.2% of apps reach that last one.
Test it before the store judges it
Option 1: verified testers assigned to you
- Starter, $15 gives you 15 verified Android testers for the full 14 days.
- Pro, $25 gives you 25 testers, plus a detailed ASO report and priority support.
- Testers assigned within 6 hours, so your 14 days start today.
- They open your app daily and stay active for the full period.
- Anyone who goes inactive is replaced.
- A feedback report at the end, plus pre-filled answers for the production access questions.
- 100% production access or your money back.
More than 10,000 apps have reached production this way. Submit your app to start.
Option 2: join a free Pack of 16
If you would rather trade time than money, the Testers Community app puts you in a pack of 16 developers. Everyone tests each other’s apps for 16 days, opening every app daily and leaving feedback.
Next steps: the full publishing checklist, our 2026 app trends report, and what happened when Google removed millions of low-quality listings.
