How Many Copies Does the Median Indie Game Sell?
The median paid Steam release of January 2026 has sold about 100 copies. The full distribution, why averages lie, and what moves a game up the curve.
A paid game released on Steam in January 2026 has, at the median, sold about 100 copies and grossed about $1,000, roughly eight months later. That number comes from a full cohort rather than anecdotes: GameDiscoverCo, the analytics firm that estimates Steam sales professionally, published its numbers for all 1,230 paid releases of that one month in a public spreadsheet. A third of them had sold no copies at all. Three quarters had sold fewer than 1,000. This article walks the whole distribution, shows why the averages you have seen quoted are fiction, and translates the curve into the arithmetic a solo developer can actually plan with. A disclosure up front: this blog belongs to Egmatic, a no-code 2D editor in pre-alpha, and what that has to do with the median is near the end.
If you came from the wishlist conversion benchmarks, this is the sequel from the other direction: that piece was about the rate at which interest becomes sales, and this one is about how many sales the median game ends up with at all.
Quick answer
| Question | Number | Where it comes from |
|---|---|---|
| Median copies sold, lifetime to date | About 100 | GameDiscoverCo data, all 1,230 paid Steam releases of January 2026 |
| Median gross revenue | About $1,000 | Same dataset |
| Games that sold zero copies | 412 of 1,211 with estimates, about a third | Same dataset |
| Games below 1,000 copies | 74.4 percent | Same dataset |
| Games above 10,000 copies | About 8 percent | Same dataset |
| Games above $1 million gross | 16 of 1,230, or 1.3 percent | Same dataset, and GameDiscoverCo's own writeup |
| Top 1 percent's share of the month's revenue | 72.8 percent | Same dataset |
Read the table as a map, not a verdict. The rest of the article explains how the data was measured, what the median does and does not mean for your game, and which levers actually move a title up this curve.
Where the numbers come from
Sales numbers for indie games are usually quoted from three unreliable sources: winner interviews, rage bait, and marketing. This dataset is different in kind. In September 2026, Simon Carless of GameDiscoverCo asked whether the game business is sustainable for the average indie, and to answer it his firm dumped its lifetime revenue and units-sold estimates for an entire release month, every paid game that launched on Steam in January 2026, into a public spreadsheet. That is 1,230 games, tracked from launch to the middle of September, about eight months of sales history for a complete cohort, the successful and the invisible alike.
Three honest caveats. These are estimates, modelled from the signals GameDiscoverCo tracks across the store, not Valve's internal books, though the firm's top-line claims have held up when developers published their real numbers. January is a slow month; Carless notes that busier months see 20 to 30 new releases grossing over $1 million rather than January's 16. And the long tail is still accruing: GameDiscoverCo's own long-tail research says the median game roughly quadruples its launch-week revenue over two years, so the medians here will edge up as the cohort ages. None of that changes the shape of the curve, which is the actual subject.
The average indie game does not exist
Take the mean instead of the median for the January cohort and you get a game that sold 7,253 copies and grossed about $110,000. That game does not exist. It is an artifact produced by adding one Devolver-published hit, Quarantine Zone: The Last Check with over a million copies and $21 million, into a sum with a thousand games that sold mostly to their developers' friends. The mean sits 72 times above the median because a handful of titles at the very top carry almost everything.
The concentration is the finding. In that single month, the top 1 percent of releases took 72.8 percent of all revenue, the top 5 percent took 92 percent, and the bottom half of the cohort, 600-plus real games, took 0.01 percent. Zoom out from one month to the whole store and it gets sharper, not softer: an independent census of 53,597 Steam games puts the top 1 percent at 94 percent of all revenue. When someone quotes average indie game sales, they are describing the winners and calling it a typical outcome. When you plan your own game, plan against the median.
The distribution, percentile by percentile
| Percentile | Copies sold | Gross revenue |
|---|---|---|
| 25th | 0 | $0 |
| Median | About 100 | About $1,000 |
| 75th | 1,000 | $5,000 |
| 90th | 5,700 | $51,000 |
| 95th | 20,100 | $197,000 |
| Top of the chart | 1,024,800 | $21.25 million |
Two features of this table deserve a slow read. The first is the floor: a quarter of a full release month sold literally nothing, and the 25th and 50th percentiles are nearly identical, zero and a hundred copies. The gap between releasing a game and having anyone buy it is where most of the cohort lives. The second is the compounding: each percentile step multiplies. From median to 75th is ten times. From 75th to 90th is about six times. From 90th to 95th is three and a half times, and above that the numbers stop being about game quality at all and start being about market events: a publisher push, a creator video, a feature placement.
Carless's own reading of the same data, in the article that published it: a game with a $500,000 budget needed to be one of those 16 million-dollar releases to be profitable self-published, and only about seven in that month would have fully recouped a publisher's $500,000. Even at a $100,000 budget, only about 6.5 percent of the month's games would have broken even net of Steam's cut. His conclusion is not despair; it is that the market rewards two positions, a strong catalog or a compact team shipping something notable from a low cost base. The median single release, priced to disappear, is not one of them.
Why you have never met the median game
Everything you have ever read about an indie game's sales was written about the right side of the table. Post-mortems come from developers with an audience for the post-mortem. Sales milestones are announced when they are worth announcing. Conference talks are given by the people the conference wants. The median game, the one with 100 copies and no thread about it, is structurally invisible: it has no marketing reason to speak and no audience to speak to, so it does not appear in your feed at all.
This is survivorship bias doing what it always does, but on Steam the filter is unusually brutal because the spoils are so concentrated. When the top 1 percent takes 94 percent of a market's revenue, the median participant is not slightly unsuccessful, they are outside the economy that the discourse is about. The practical consequence is calibration: every sales number you have absorbed by osmosis, the million-copy launches, the six-figure months, is drawn from the top sliver of the top percentile. If your internal picture of "normal" comes from that sliver, then a 2,000-copy launch, which beats four out of five games in a real release month, will feel like failure. It is not. It is the 80th percentile.
What separates the median from the 90th percentile
The curve is steep, but it is not random, and the same dataset plus the public case studies show where the separation happens.
Demand before launch. The single best predictor available to a developer is the wishlist count, because the median game sells about 0.2 times its wishlist balance in launch week and converts 5 to 20 percent of the list over its life. A game that reaches launch with 7,000 wishlists is playing a different game from one that reaches launch with 300, and the mechanics of getting from 300 to 7,000 are covered in our guide to the first 100. The cohort data is consistent with this from the other side: a third of January's releases sold nothing, which is what zero launch-day interest looks like on a store where the release email to wishlisters is the most reliable conversion event a small game gets.
Price. The median January release was priced at $5.49, and half the cohort sat at $5 or below. The 94 games that sold more than 10,000 copies had a median price of $12.99, and GameDiscoverCo's monthly top-50 charts put the median price of a top-selling new release at $15.64. The pattern is not that cheap games cannot win, it is that cheap is the default setting of games built without a distribution plan, so cheap and invisible travel together. Picking a price deliberately, with the tradeoffs in mind, is its own discipline; our pricing guide walks it.
Iteration after launch. The 90th percentile is not a launch-week position. The median game multiplies its launch-week revenue by 1.39 in month one, by 3.16 in year one and by 4.2 by year two, and the games that climb are the ones that keep shipping: the public case of Tower Factory shows conversion rising from 6.8 percent to 13.7 percent across a year of updates, sales and creator coverage, finishing at 52,000 copies. What you do with the tail has its own playbook in the post-launch strategy guide.
From copies to actual money
Copies are the scoreboard; money is the constraint, so here is the translation at a $15 price with Steam's standard 30 percent cut.
| Copies sold | Gross | Net after 30 percent |
|---|---|---|
| 100 (the median) | $1,500 | About $1,050 |
| 1,000 (75th percentile) | $15,000 | About $10,500 |
| 5,700 (90th percentile) | $85,500 | About $59,900 |
| 20,100 (95th percentile) | $301,500 | About $211,000 |
Reality takes a few more cuts than the table shows: refunds, chargebacks, and regional pricing, which sells the same game for half or a third in some markets. The $100 Steam Direct fee, refunded at $1,000 adjusted gross, is the median game's entire margin story: the median cohort member grosses almost exactly the threshold at which the fee comes back. The fee, the cut tiers, and the whole mechanics of getting paid are in the Steam publishing walkthrough, and the fuller map of store commissions and revenue models is in how indie games are monetized.
Set this against the cost side and the real arithmetic of indie viability appears. The cost of a solo 2D game can honestly start near $0 in tools plus $100 for the store, but the largest line is months of evenings, and the median outcome does not come close to paying for them. This is why Carless's conclusion lands on the cost base: the market did not get kinder, so the sustainable positions are a catalog that compounds, or a small team whose per-game costs are low enough that the 75th percentile pays. The less each attempt costs you, the fewer copies each one has to sell.
How to estimate any game's sales yourself
You do not need GameDiscoverCo's models to get the right order of magnitude for a game you are studying, competing with, or calibrating against.
- The review multiplier. Roughly one buyer in 30 to 50 leaves a Steam review, a rule of thumb Chris Zukowski popularized, so multiply the review count by that band. It is loose for a single game and reliable across a genre or a competitor set. The anchor worth memorizing: 1,000 reviews correlates with $150,000 or more in revenue.
- The wishlist multiple. For unreleased games, take the wishlist count and apply 0.1 to 0.2 for launch week and 5 to 20 percent for lifetime, the bands from the conversion benchmarks. A page sitting at 4,000 wishlists is signalling several hundred copies in week one at the median.
- Read the shape, not the total. A game with 329 reviews in month one and slow growth afterwards is a different story from one that accrued the same reviews over two years of sales. Sales curves are public in outline form through review timestamps.
- Check your own dashboard against the bands. For your released game, the Steamworks wishlist report and the sales page give you the real curve; the game analytics guide covers which numbers to read weekly and which only cause anxiety.
Common mistakes
- Planning against the mean. Any plan built on average indie sales figures is built on the winners. Ask for the median, then ask what the 75th and 90th percentiles did, because those are the outcomes you can actually influence toward.
- Treating the median as destiny. The median game is invisible, not doomed; the levers that move a title up the curve, launch demand, deliberate price, post-launch iteration, are all craft decisions rather than lottery tickets. The curve describes the cohort that mostly did not pull them.
- Quoting one month as the whole market. January 2026 is a full cohort, which is its virtue, and one month, which is its limit. Busier months put more games over $1 million. The shape holds; the exact medians drift.
- Confusing gross with take-home. $1,000 gross is about $700 net of the 30 percent cut, before refunds and regional pricing. Run every milestone you celebrate through the net number.
- Skipping to conclusions from estimates. These figures are modelled, good enough for planning and calibration, wrong in the details for any single game. Use them to set expectations, then replace them with your own dashboard the week you launch.
How Egmatic fits
The median game in that dataset is not a game that was outplayed on code. It is a game that shipped into silence, at a default price, and then stopped moving: no demo to convert curiosity, no update cadence to give the tail a reason to compound, no second release to start the catalog math. Those are shipping problems, which is the layer Egmatic is building. It is a no-code 2D editor with a ship layer that carries a scene to a released game, no code to maintain, no backend to run, so cutting the demo, shipping the update and standing the game back up in front of players are movements of one tool. The AI in the editor executes under your direction, and every file stays yours. We are pre-alpha, we do not announce dates we cannot keep, and the waitlist at egmatic.com is where the build's priorities get set.
Ship the game the median forgot.
The median sold 100 copies and went quiet, and the difference was shipping cadence: demo, update, next release. Egmatic is a no-code 2D editor with a ship layer built around exactly that loop. Reply on the waitlist with the number you are privately hoping for; requests sent during pre-alpha are the ones that steer the build.
Conclusion
The median paid Steam release of January 2026 sold about 100 copies for about $1,000 gross, a quarter of the month sold nothing at all, and the top 1 percent took 73 percent of the money. That is the honest baseline, and the honest response to it is neither despair nor the survivor stories: it is arithmetic. One thousand copies beats three quarters of a real release month; ten thousand puts you in the top 8; both are reachable through launch demand, a deliberate price and a tail you keep shipping into, and all three are craft. Decide before launch which bill the game has to pay, plan against the median, work the levers, and let your own dashboard, not the discourse, tell you where you stand.
Sources
- Simon Carless, GameDiscoverCo — Is the game biz sustainable for the average indie dev?, September 15, 2026 (the January 2026 cohort of 1,230 paid releases; 16 games above $1 million; breakeven arithmetic at $500,000 and $100,000 budgets; the linked public spreadsheet of lifetime revenue and units per game, retrieved and analyzed October 15, 2026)
- GameDiscoverCo — January 2026 top Steam paid releases, full cohort spreadsheet (public data dump; all percentile, median and revenue-share figures in this article computed from it)
- GameDiscoverCo, Simon Carless — Revealed: the state of Steam wishlist conversions and long tail in 2024, April 15, 2024 (long-tail revenue multipliers of 1.39x month one, 3.16x year one, 4.2x year two)
- Chris Zukowski, How To Market A Game — Do wishlists matter any more?, January 29, 2024 (observed conversion range of 1 to 40 percent; the review-count multiplier of roughly 30 to 50)
- Chris Zukowski, How To Market A Game — Leftover research from 2025, December 15, 2025 (Tower Factory conversion rising from 6.8 to 13.7 percent over a year, finishing at 52,000 copies; the correlation between 1,000 reviews and $150,000 or more in revenue)
- Independent census of 53,597 Steam games, cited and verified in our AI games on Steam analysis (top 1 percent of titles take 94 percent of all revenue)
- Valve — Steam Direct documentation ($100 fee per product, recoupable at $1,000 adjusted gross revenue; the 30, 25 and 20 percent revenue tiers)