Why Most Indie Games Fail on Steam: What the Data Shows
Updated August 5, 2026
Of the roughly 41,000 games released on Steam in the last three years, 76.5% earned less than $5,000 in gross revenue. Before Valve’s cut. Before tax. Before the years of work.
That is the number the “90% of indie games fail” headline is reaching for. A typical Steam release grosses a few thousand dollars, sells in the hundreds of copies, and has close to an even chance of ending its release year with fewer than ten user reviews. The odds are brutal, and almost every part of the distribution has a documented explanation.
How much does a typical Steam game actually make?
Gamalytic’s analysis of the Steam catalog, which covers more than 71,000 games on sale, breaks the whole platform into four brackets:
| Lifetime gross revenue | Share of all Steam games | Share of games released in the last 3 years |
|---|---|---|
| Under $5,000 | 67% | 76.5% |
| $5,000 to $50,000 | 17.2% | 13.4% |
| $50,000 to $200,000 | 6.9% | 4.5% |
| Over $200,000 | 8.9% | 5.6% |
Gamalytic, analyzing the Steam catalog including approximately 41,000 games released in the previous three years.
Read the two columns side by side and the trend is the story. Recent releases do worse in every bracket. The catalog is getting more crowded faster than the audience is growing, so each new cohort lands in a thinner slice of attention than the one before it.
One caveat that gets lost in these discussions: gross revenue is not income. Valve takes 30% of the first $10 million. Regional pricing pulls the average sale price below the sticker. Refunds, taxes, and any publisher or platform fees come out of what remains. A game grossing $5,000 has likely paid its developer something closer to $3,000, for what our own reporting on how long it takes to make an indie game suggests is usually one to five years of work.
Is “90% of indie games fail” actually true?
It depends entirely on what “fail” means, and nobody defining it that way has told you.
If failure means never earning a living wage, the real number is worse than 90%. If it means never earning back a modest budget, roughly 76% of recent releases missed $5,000. If it means never finding a single appreciative player, the figure is far lower than 90%, because plenty of games with tiny revenue have devoted small audiences.
The data gives you a distribution and no line drawn through it. Most games land in a long tail where they earn a few thousand dollars and are played by a few hundred people. Whether that counts as failure depends on what the developer was trying to do, and quite a lot of them were not trying to quit their jobs.
Where does the money on Steam actually go?
Into a very small number of games. Video Game Insights’ analysis of 2023 releases found that the top 10 games launched that year took 61% of all revenue generated by that year’s releases, the highest concentration since 2016. The top 100 took 91%.
The pattern held into 2025. More than 19,000 games shipped, and every one of the top 20 by revenue grossed more than $50 million, with 13 of them clearing $100 million.
This is a power law. Every average you read about indie revenue is misleading for that reason: mean revenue per game on Steam is a respectable-sounding number produced almost entirely by a handful of outliers. The median is the statistic that describes the platform, and the median is a few thousand dollars.
Does genre decide the outcome before you start?
More than most developers would like. GameDiscoverCo tagged every Steam game earning over $1 million and found revenue splits sharply by genre:
- Action: 58.37% of revenue
- RPG: 17.11%
- Strategy: 13.97%
- Simulation: 9.76%
- Sports: about 1%
Action and RPG together take three quarters of the money among Steam’s million-dollar earners. That is the slice of the platform this dataset covers, and it says nothing about how revenue splits across the long tail below a million.
There is a trap in reading that as advice. Those genres also carry the heaviest competition and the highest production expectations, and the subgenres at the top are effectively closed: across all of Steam’s history, only 29 games in categories like arena shooters, battle royales and hero shooters have ever cleared a million dollars. A crowded genre with a big pot is not obviously a better bet than a small genre with a reachable audience. What the data does say is that genre is a structural decision made on day one, and it constrains the ceiling more than most first-time developers expect.
What actually improves the odds?
Three findings survive scrutiny, and the first one is the least discussed.
1. It is probably not your first game that works
Video Game Insights found that a developer’s second game sells about 40% better than their first on average, and the third about 24% better than the second. The gain comes from having learned to market, to scope, to pick a genre, and to build an audience that carries between projects.
The single strongest predictor in that dataset is having shipped before. Which means the most valuable outcome of a first game that earns $4,000 may be the second game it makes possible. Our piece on the rise of micro-studio games covers the survivors who got there.
2. Wishlists matter, but not at the ratio people quote
You will see a confident “wishlists convert at 20%” claim repeated everywhere. The measured figure is different and far less flattering. GameDiscoverCo tracked launches from September 2024 to August 2025 and found a median wishlist-to-first-week-sales ratio of 0.15x for games entering with more than 25,000 wishlists. A game launching with 100,000 wishlists sold roughly 15,000 copies in week one. Above a $10 price, the median fell to 0.10x.
The variance is the real headline: the same study put the spread between games at 10 to 20 times, not 10 to 20 percent, with individual titles landing as low as a tenth of the median and outliers converting many times above it. A wishlist count is a reading of interest with an enormous error bar attached to it.
3. Visibility binds harder than quality
SteamDB counted more than 19,000 releases in 2025, and nearly half finished the year with fewer than ten reviews. About 2,200 got none at all.
A game with no players generates no signal, and by Valve’s own description the store algorithm follows player interest. We broke that mechanism down in how Steam’s discovery algorithm actually works, and it is the reason most games do not lose a competition so much as fail to enter one.
What does this mean if you are a player, not a developer?
That the quality signal you rely on is much weaker than it feels. A game with 40 reviews is not a game the market judged and rejected; it is far more likely a game the market never saw. Review counts measure exposure at least as much as merit.
It also means the discovery work you do yourself has unusual leverage. With three quarters of recent releases earning less than $5,000, the gap between “good” and “found” is where most of the interesting games live. That is the entire premise behind our indie game iceberg list.
Where we sit in this data
We are Overbaked Studio, a small independent studio in Istanbul, and Grandma Left Me a Building is our first game. By the numbers above, that puts us in the worst-performing cohort on the platform: a first title, from an unknown studio, in simulation, which takes under 10% of the revenue among Steam’s million-dollar earners.
We are not going to pretend the strategy solves that. What the data changed for us is the definition of a good outcome. If the baseline for a first game is a few thousand dollars and a few hundred players, then the useful goal is to be the studio that gets to make a second one, since that is where the 40% lives. That shapes what we build. We put the work into depth in one narrow place: 30+ tenants across four social tiers, 300+ distinct interactions between them, and a trait-based system that decides whether the philosopher who answers every question with “but why?” settles a floor down or empties it. It is unreleased, so we have nothing to point at yet, and the distribution above does not care what we intended.
Frequently asked questions
How much does the average indie game make on Steam?
The average is misleading because a handful of hits distort it. Use the distribution instead: 67% of all Steam games have earned under $5,000 lifetime, and that rises to 76.5% for games released in the last three years.
How many copies does a typical indie game sell?
Most sell in the low hundreds of copies. Revenue data is the more reliable public measure, and at typical indie prices the under-$5,000 bracket implies a few hundred copies for the majority of releases.
Do wishlists guarantee sales?
No. The median game with more than 25,000 wishlists converted at 0.15x into first-week sales in GameDiscoverCo’s 2024 to 2025 study, and the spread between individual games ran to 10 or 20 times rather than 10 or 20 percent. A wishlist count tells you roughly how much interest exists and very little about how many copies will sell.
Is it still worth making an indie game in 2026?
That depends on what you want from it, and the data does say the second one goes better. If the plan requires the first game to replace a salary, the distribution is against you. If the plan is to finish something, find a small audience, and build from there, the odds look very different.
Key takeaways
- 76.5% of Steam games released in the last three years earned under $5,000 gross, up from 67% across the whole catalog. New releases perform worse than the historical average.
- Revenue follows a power law. The top 10 games launched in 2023 took 61% of that year’s release revenue and the top 100 took 91%.
- Genre sets the ceiling early. Among Steam games earning over $1 million, action and RPG take roughly three quarters of the revenue and simulation takes under 10%.
- Shipping before is the strongest predictor. Second games sell about 40% better than firsts, third games about 24% better than seconds.
- Wishlists convert at a median of 0.15x into first-week sales, with a spread between games measured in multiples of 10 to 20 times, not percentage points.
If you want to help a first game clear the visibility problem described above, you can wishlist Grandma Left Me a Building on Steam.
