I've been thinking a lot about the Steam ecosystem. When we released our first game, Curious Expedition 1, back in 2014, it was a completely different world.
Back then, Steam guaranteed every new game 1 million page impressions. It was their way of giving you a good shot, by giving you a minimum viable share of views. The platform was still curated through the Greenlight system, where the community had to vote for your game to even be allowed on the store.
When we started, there were around 6,000 games on Steam. today, there are over 120,000! In 2024 alone, around 19,000 new games were released on the platform.
The Greenlight system is long gone, replaced in 2017 by the "Steam Direct" program and a $100 fee. Game engines are easy to get, documentation is everywhere, and digital distribution gives anyone a global platform. This big growth in content forced a change. Steam couldn't give 1 million views to every game anymore, especially with the flood of "asset-flipper" and "shovelware" titles.
But the biggest change wasn't just the removal of Greenlight. It was a big change in Steam's main idea.
The Year the Algorithm Changed
I remember a specific moment, maybe seven years ago, when many of my game dev peers saw a big, sudden drop in revenue. It was up to 30% from one day to the next. Developers at the time called it the "October Discovery Bug" of 2018; traffic from key discovery widgets simply vanished, and for many... it never returned. One developer, Grey Alien Games, publicly reported their sales had "halved" almost overnight.
This happened at the same time as a change to the Steam algorithm. Steam had decided to stop "AVERAGING" revenue and instead focus on making the most money. On November 30, 2018, Valve made another announment: new revenue-sharing tiers: their 30% cut would drop to 25% only after a game earned $10 million, and to 20% after $50 million.
This was seen by some indies as a "slap in the face". For Valve, the reason to do this was now clear. They started feeding the winners. By 2019, Valve's own data showed the effect: the bottom third of games (small indies) earned less money than in 2018, while the top two-thirds earned more.
This trend has only grown. We are now in an age of big, out-of-nowhere hits: Peak by Aggro Crab, which sold 11 million copies ; Supermarket Simulator, from a tiny team, which has made over $34 million ; Content Warning, which got 6.2 million free downloads and then 700,000 paid sales in its first week ; or the brand-new latest super hit RV There Yet? that sold 1.28 million copies in just five days.
Some call this a new "golden age" because of these many new megahits. To me, it points to a very big change in the ecosystem, one I would compare to the difference between YouTube and TikTok.
Algorithm is King
For the longest time, the main logic for creators was the YouTube model. You know the trope: "Please subscribe and like!" The focus was on gathering subscribers. The idea was that if you built a large subscriber base, they would see your new videos, giving you an increasingly large minimum viewership over time.
Then came TikTok.
TikTok famously cares less about the subscriber count and introduced the "content graph." On TikTok, the content is king, not the creator's history. A brand-new account with zero followers can post one good video and get millions of views. If a video performs, the algorithm rewards it with more views, which gets it more views, in a feedback loop.
My hot take is this: Steam is now a TikTok platform.
The old Steam logic was the YouTube model. The common consensus was:
- Announce early.
- Put your Steam page up and collect wishlists (subscribers).
- Work marketing beats, go to trade shows, talk to journalists.
- Seed your game long-term.
Those things aren't useless, but their role has changed. They are no longer the goal, they are only the entry fee now. This is the key difference. The old "YouTube" logic of collecting wishlists is now the thing you must have before the "TikTok" algorithm will pay attention. You must "prove to the Steam algorithm that my game is worth it". What most people say is that you must gather 7,000 to 10,000 wishlists just to unlock the "Popular Upcoming" widget, which is the main gate to launch-day visibility. Data from 2025 shows wishlists still have a 70% correlation with first-month sales, but they are dwarfed by the viral feedback loop that kicks in after the algorithm takes notice. Nothing trumps the power Steam has to direct the eyeballs of 185 million monthly users.
The Middle Ground Is shrinking
In this new ecosystem, if your game pops up in "New & Trending" and has a good thumbnail, a good title, and an trailer that clicks. If it uses a genre trend or a theme people instantly "get", Steam will send more and more people to your page in a "viral loop".
This has a hard and simple result: The middle ground is breaking away.
In the past, you could have an "okay" success or a "break-even" game. Now, you are either a winner or a loser.
It's no longer a linear calculation. A game with a 15% click-through rate (CTR) won't get three times the sales of a game with a 5% CTR. The 15% CTR game will get all the views and all the sales. The 5% CTR game, whose low CTR is seen as "poison" by the algorithm, will get no sales.
This is the "winner-takes-all" platform Steam has become. The numbers are amazing: a 2024 report found that the top 10 games of 2023 made 61% of all revenue for the year. If you are not a top-performing, top-converting game, you are invisible. For context, a 2023 analysis found the median revenue was a "whopping" $799.
This, I believe, another reason why we see so many big indies coming out of nowhere, and at the same time, why so many established, larger companies are struggling and falling apart. The middle-market is crashing hard AA studios are squeezed and the clear losers, lacking both the blockbuster budgets of AAA and the viral novelty of hit indies. They were built for a world where "middle success" was possible.
It's exactly like the old joke about two people running from a bear. You don't have to outrun the bear; you just have to outrun the other person. In the new Steam world, if your game is that slowest runner with a low click-through rate, the algorithm is the bear and won't just give you fewer sales, it will kill you.
What This Means for Developers
A common consensus used to be: "Ideas are nothing, execution is everything." I think we have to reverse that. We are back to the point where ideas are everything. Specifically, an idea that is easy to understand, clickable, and marketable on its face.
Community building is still important, but its role has changed. Your community is no longer just your launch audience; they are your validators. They are the ones who help you guide your game, your trailer, and your store page until you have a package that makes people click.
For new developers especially, you won't have a big community. You must focus on getting the packaging right. Everything is now about "first glance" appeal.
The talk of the town is "validation," and for good reason. I have been speaking more about this recently. This is the whole point of tools like Steam Next Fest. As marketing expert Chris Zukowski notes, Next Fest is "not a testing ground... it's the grand debut". Your demo must be bulletproof and vetted by streamers before the festival to prove it's a "winner". In a "winner-takes-all" market, you can't afford to release a "loser." You have to validate, reduce risk, and focus hard on one thing: creating a game that the algorithm wants to feed.