Published on April 7th, 2020 📆 | 5839 Views ⚑0
Ars analysis: ~80% of Steam games earn under $5K in first two weeks
It has been roughly two years now since Valve shut off the source of Steam Spy’s huge, randomly sampled sales estimates and promised a “more accurate and more useful” replacement to come. We got our first glimpse of what that replacement might entail today, as Valve gave a rare glimpse into its treasure trove of aggregate sales data across thousands of PC games.
The blog post sharing that data correctly points out that the raw number of games finding some minimum level of sales success on Steam has increased vastly since 2012 (when Valve launched Steam Greenlight and loosened its tight control of what games could appear on the storefront). But Valve’s selective view of the data leaves out a huge mass of games that make less than $5,000 in their first two weeks on Steam’s virtual shelves. An Ars analysis finds those titles have made up the vast majority of Steam releases for the last five years.
Filling in the holes
To get at that data for the charts above, we started with the graphs Valve itself provided in its blog post today. These lay out the number of games making over $5,000, $10,000, $50,000, $100,000, and $250,000 in their first two weeks, respectively, by release year. I used photo editing software to measure and convert the bars in those graphs into raw numbers, but the actual numbers may be off by a fraction of a percentage point from Valve’s internal benchmarks (we didn’t decipher the graphs for 2005 and 2006, when the total number of Steam releases was too small to draw much meaningful data).
We then compared those numbers to the total number of Steam releases for each year, as collected by Steam Spy (I removed non-game software and free-to-play games from consideration, as Valve did in its data). Using that, we were able to figure out the one number that Valve pointedly failed to mention for all of these years: the games making less than $5,000 in that two-week launch period.
As you can see in Fig. 3, this lowest tier of sales made up well under half of all Steam releases through 2014, a year when Steam still saw less than 1,400 total games hit the storefront. By 2015, though, the raw number of releases had ballooned to over 2,300, and our data shows that increase came almost entirely from games that were struggling to make any significant sales impact at launch.
Since 2015, both the annual growth rate in Steam releases and the proportion of those releases that struggled at launch continued to increase exponentially, before leveling off in 2019. At this point, though, the ratio of Steam releases that can’t hit $5,000 in initial sales has hovered around 80 percent for the last three years.
As Valve acknowledges in its blog post, these launch window revenue cutoffs are somewhat arbitrary. A game that makes $4,000 in a couple of weeks could be considered a huge success for a hobbyist programmer just getting their solo start. Initial sales also don’t always capture the full picture for games that may find success long after launch, as well.
But Valve also points out that “most recent games earning around $10,000 in the first two weeks earned between $20,000 and $60,000 over the course of 12 months following release.” Convert that to an annual revenue number for a struggling indie developer, and you can see how launch window sales of under $5,000 can easily be considered a commercial failure (though the calculus changes if a developer can crank out multiple games per year or keep sales steady for a multi-year period).
What does it mean?
There are two ways to look at these numbers. The first, which Valve focuses on in its blog post, is the raw number of games in those “high initial revenue” groupings (i.e. above $5,000). As you can see in Fig. 2 (and below), the number of such titles generally rose slowly after Steam launched, absolutely exploded in 2014 amid looser Steam Greenlight standards, and continued to rise from that expanded number since then. Many more total games are quickly finding a significant audience on Steam than ever before, and Valve can rightly feel good about that.
But even as the number of successes on Steam has increased, the number of failures has increased even more, both on an absolute and relative basis. For developers, that means just getting a game on Steam isn’t anything close to a guarantee of being able to find a significant audience these days. For players, that means sifting through Steam storefront listings full of a lot of crap that apparently very few people want to buy.
Valve is constantly tinkering with its “discovery” algorithms to help with both sides of this equation, and those algorithms do the best they can to direct deserving games to their deserving audiences. But on a platform with increasingly hands-off submission guidelines and dozens of games launching every day, there are inevitably going to be a lot of losers in Steam’s sales lottery.
There are signs that things are improving, though. As Valve points out in its blog post, the proportion of games reaching that $10,000 initial sales milestone ticked up 11 percent from 2018 to 2019, slightly reducing the “failure rate” for a generic Steam release in the last year. And the average initial earnings for Steam games increased from 2018 to 2019 at all but the lowest percentile levels.
Those numbers suggest that quality games are starting to find even larger audiences on Steam these days, while weaker titles are proving even less lucrative. That doesn’t mean we’ll be getting back to the relative “all killer no filler” world of early Steam any time soon, but maybe we’re starting to get a little closer with each passing year.
Listing image by Getty / Aurich Lawson