How to revive a dying indie game: what 195 Steam comebacks did

Of 1,493 paid indie games on Steam that faded to under a third of their first-year pace, 195 (13%) later had a month back at that pace or better.

What came just before those months was rarely a discount. It was a release.

A faded Early Access game that shipped 1.0 came back 57% of the time. For games already at 1.0, paid DLC, a Major Update or a console version came back about 1 time in 20. In a faded month where the developer posted and shipped nothing, the chance of a comeback that month or the next was 0.5%.

This is for the developer whose game sold, then slowed, and who is deciding whether there is anything left to do with it. Every number here comes from Steam’s public store data, and the method is at the end.

Nine in ten games faded, and that is normal

We make games too. Watching the reviews on a game you love go from dozens a week to a few a month feels like failure. It is mostly just what happens.

We took the 1,651 paid indie games that launched on Steam from 2020 to 2022 and drew at least 100 Steam reviews in their first year. Reviews are the best public trace of buyers over time.

We called a game faded in any month, from month 18 on, when the six months before it averaged under a third of its first-year monthly pace. A game with 240 reviews in year one averages 20 a month, so it had faded once it ran under about 7 a month for half a year.

By August 2026, 1,493 of the 1,651 (90%) had faded at some point. Steady decline after launch is the ordinary shape of a game’s life, not a sign something went wrong. Our long-tail study found the median game still drew about half its reviews after year one. It just drew them slowly.

195 faded games came back to their first-year pace

A comeback month is a faded month in which the game drew at least as many reviews as its average first-year month. 195 of the 1,493 faded games (13.1%) had at least one.

The median comeback month ran 6.2 times the faded months before it. More first comebacks came in year three than in any other year.

1. Year 2 (months 18 to 23)40
2. Year 377
3. Year 442
4. Year 523
5. Year 611
6. Year 72

1,651 paid games under Steam's Indie tag launched January 2020 to December 2022 with 100+ Steam-purchase reviews in their first 12 months. Monthly review histograms read September 16 and 21, 2026, through August 2026.

Later years have fewer games old enough to reach them, so the drop after year three is partly the calendar. But it also matches what we found for quiet launches: the odds fall every year a game is left alone.

Score mattered, but less than you might expect. Among faded games rated 80% positive or better, 14.3% came back. At 70 to 79%, 12.7%. Under 70%, 8.4%.

A 1.0 brought back more than half the faded Early Access games that shipped one

For each thing a developer can do, we asked: when a faded game did it, how often was that month or the next a comeback month?

The 1.0 stands alone. Of 89 Early Access games that had faded and then released 1.0, 51 (57%) had a comeback month right then.

Steam treats 1.0 as a launch: Valve gives it full release visibility and emails everyone with the game wishlisted. A faded Early Access game is still holding that launch in reserve, and our Early Access study found 35% of indie Early Access games from those years still hadn’t spent it.

A 1.0 often came with other things in the same weeks: a console version, a soundtrack, a Major Update. So for every other move below, we left out any that fell within a month of a 1.0. Otherwise a console launch timed with a 1.0 gets credit for the 1.0.

Without a 1.0, the best moves worked about 1 time in 20

Away from a 1.0, three moves stood out, and they did about equally well: paid DLC, a Major Update and a first console version.

1. Released 1.0 (89)57.3%
2. Paid DLC released (126)5.6%
3. Major Update post (528)4.9%
4. First console version (86)4.7%
5. Sale post (683)2.2%
6. Developer's next game launched (792)2.0%
7. Smaller update post (523)1.7%
8. Nothing posted or released (45,724 months)0.5%

1,651 paid Steam indie games launched 2020 to 2022 with 100+ Steam-purchase reviews in year one. Faded: from month 18, the six months before averaged under a third of the year-one monthly pace. Comeback: a faded month at or above that pace. Brackets: moves made while faded; a run of consecutive months counts once; moves other than 1.0 within a month of a 1.0 left out. Reviews through August 2026.

One in 20 is not good odds. Of 528 Major Updates by faded games, 26 came right before a comeback. But it is about ten times the odds of a month in which nothing happened, and those are the moves a developer controls.

The counts are small enough that the order of the three means little. What they share is that each gives Steam, and the people who already know the game, a reason to look again.

A sale alone brought a game back about 1 time in 45

A faded game that posted a sale was followed by a comeback 2.2% of the time (15 of 683). That beats doing nothing. It is under half the rate of a Major Update or paid DLC.

Comebacks that came with a sale also held up worse. Over the next three months, 6 of 16 we could follow (38%) kept at least twice their faded pace. For comebacks that came with a 1.0, 35 of 46 (76%) did.

This fits how Steam works. Valve’s update visibility rounds put a game with a Major Update on the Steam homepage for people who own it or have it wishlisted. A discount puts a lower price in front of people who mostly passed on it already. If the game is the same, most of them pass again.

We can only see sales a developer posted about, so this undercounts discounts. A posted sale is the most visible kind, though, and it still did less than any release.

The studio’s next game did no better. A later release credited to the same developer came before a comeback 2.0% of the time (16 of 792). Sequels were too few to count fairly, because some aren’t tagged Indie on Steam, so treat that case as unmeasured.

Paid DLC did as well as a Major Update, and few games made it

373 of the 1,651 games (22.6%) have content DLC on Steam, not counting soundtracks, art books and supporter packs. Only 168 (10.2%) released paid content DLC after their first year.

When a faded game did, 7 of 126 releases away from a 1.0 (5.6%) came right before a comeback. That is the same as a Major Update, within the noise of a count this size. DLC has one thing an update doesn’t: it earns money on its own, even in the months when the base game doesn’t move.

Soundtracks and other extras did nothing on their own. Every soundtrack release followed by a comeback came within a month of a 1.0.

Nearly half the comebacks had no visible cause

87 of the 195 comebacks (45%) came with none of the moves above, in that month or the one before. 74 of them came with no developer post at all.

We read a dozen of these by hand. One came the month a film based on the game was released. One came the month the studio’s sequel launched on its own store page, which our Indie listing didn’t include.

The rest had no public trace we could find. A streamer, a store feature or a discount without a post would all look like this.

Those comebacks were the least likely to last. 36 of 85 (42%) kept twice their faded pace over the next three months, against 76% after a 1.0. Luck brings a game back for a month. Shipping something keeps it there more often.

When reviving a game is the wrong move

The 1.0 number is the one most people will take away, so first the honest limits.

• A comeback month is not a comeback. Only 53 of 188 comebacks we could follow (28%) held two-thirds of the first-year pace for the next three months. Most games went back down, just not all the way.
• The 1.0 only works once. Valve says it can’t turn Early Access back on for a game that has left. If your game is already at 1.0, your best moves came back about 1 time in 20.
• Ports and updates cost money and months. Our porting cost study found labour quotes from $5,000 to $100,000 a platform. A one-in-twenty chance at a month of first-year sales has to be worth that to you, or the work has to pay for itself some other way.
• Some games are finished. A game that faded after a good first year did its job. Leaving it on sale costs nothing. Spending your next year on it, instead of your next game, costs a lot.

What to do with a game that has faded

1. Read your own curve. The monthly counts we used are public for any game at store.steampowered.com/appreviewhistogram/ followed by its app ID. If six months in a row ran under a third of your first-year average, you’re where these games were.
2. If you’re in Early Access, the 1.0 is your best card. Treat it as a launch: a trailer, press, a price plan, a date you believe. Don’t let the game slide into Steam’s not-updated notice first.
3. If you’re at 1.0, ship something big enough to announce. Post it as a Major Update and spend one of your update visibility rounds on it. A patch with a sale attached doesn’t do the same job.
4. Judge a console version on its own store, not on Steam. It lifted Steam about as often as a Major Update, and its own sales don’t show up in Steam reviews at all. Half the late console ports in our data came through a partner not credited on Steam, so you don’t have to fund or build one alone.
5. Don’t count on a sale to do it alone. Put the discount behind a release, not in place of one. If you have content worth selling, paid DLC did as well as a free update and earns on its own.
6. If you can’t do any of this yourself, the game still has value. Its reviews, wishlists and owners stay with it if it changes hands, and some partners take on released games to do exactly this work. Strike Two is one of them, so weigh that knowing where it comes from.

How we measured it

The games are every title in Steam’s US store search under the Indie tag (listing pulled September 16, 2026) that launched between January 2020 and December 2022, had a US price above zero on September 21, 2026, and drew at least 100 Steam-purchase reviews in their first 12 months: 1,651 games.

A game that sold in Early Access first is dated from its first review month, the rule from our quiet launch study.

Monthly review counts come from Steam’s review histogram, which counts reviews from Steam purchases by the month each was written, read September 16 and 21, 2026, through August 2026. On 15 random comebacks, we reread the live histogram on October 19, 2026, and the comeback month and the six months before it matched exactly on all 15.

Faded means a month from month 18 on in which the six months before averaged under a third of the game’s first-year monthly pace. Comeback means a faded month with at least the first-year monthly pace. A move counts for a comeback when it happened in the comeback month or the month before.

The moves come from these sources:

• Posts: each game’s Steam events feed, read September 22, 2026. Valve doesn’t document the type codes, so we identified them from post titles: Major Update, Regular Update, Game Release, Sale, livestream and cross-promotion among them.
• 1.0 dates: Steam’s store service (IStoreBrowseService/GetItems), read October 16, 2026, for games that opened in Early Access.
• Console versions: the first Switch, Xbox or PlayStation release found in our matches against each store’s US catalogue, read September 29 to October 14, 2026.
• DLC: each game’s DLC list from Steam’s appdetails, and each DLC’s release date and name from GetItems, read October 19, 2026. Soundtracks, art books and supporter packs were separated from content DLC by name; a hand read of 25 random DLC found no misfiled ones.
• The developer’s next game: any later game in Steam’s Indie listing credited to the same developer, read the same day, matched on developer credit.

The limits: reviews stand in for sales and aren’t sales. Games removed from the store, games that went free and games not tagged Indie are missing. Discounts are visible only when the developer posted one.

A move and a comeback in the same months doesn’t prove the move caused it: console versions, soundtracks and sales often landed in the same weeks as a 1.0, which is why we separated them. And the games that made a move were the ones whose developers were still working on them, which is part of what we measured.