If you follow the indie dev scene at all, you’ll probably have read a number of thinkpieces recently on whether or not there is an Indie Bubble, and whether that hypothetical bubble is in the process of collapsing. Jeff Vogel, Jonathan Blow and other big names have all weighed in.
A lot of the doomsayers refer to two particular graphs released by the marvelous SteamSpy. If you haven’t seen them yet, these are they:
They show that the number of games being released on Steam is increasing exponentially over time (although a not-insignificant portion of that is due to companies dumping their back-catalogue games on Steam now that the gates are open, mind), and that, while total monthly sales has remained fairly consistent, median sales is trending toward zero.
This proves, a lot of them claim, that the Doom is here, it’s upon us. Gather your loved ones and flee into the sea, the Doom is here, DOOM!
Now, I don’t know whether there is actually an indie bubble. Maybe, maybe not. What I do know is you can’t use that graph to draw the types of conclusions people are drawing from it. A little bit of statistics is a dangerous thing.
The problem lies in people claiming that the graph proves it is getting harder for the average developer to earn a decent income by selling their games. Because the median was higher in the past, and now it’s near zero. Makes sense, right?
No. It doesn’t.
That graph shows sales figures for Steam only, not games being sold as a whole. And Steam changed its policies to lower the bar to entry. Which means that a lot of developers that weren’t on Steam (and thus weren’t reflected in the stats) are now on Steam, and affecting values like the median.
Let me illustrate why this means you can’t draw real conclusions, if it isn’t already clear.
Let’s say you have the Olympic 100m sprint. And, for the sake of example, let’s suppose that 20 runners are allowed into the event, the top 20 in the world. And let us further suppose that big sports brands watch the event and award lucrative sponsorship deals to the top 10 runners in the race.
So 10 out of 20 earn lucrative sponsorship deals. That’s 50% of the Olympic runners who earn big bucks. So if you’re in that race those are some good odds at making money. The average earning will be fairly good, and the median earning for a runner in the sprint will also be pretty high, (the 10th place sponsorship amount)/2.
But let’s say that one year the Olympics committee decides that the Olympics needs to be more of a “people’s event”. They’ll let 1000 runners compete, and you don’t have to meet a particularly high barrier of entry, it’s first-come, first served. So what ends up happening is a whole bunch of “average joe” runners get in. They’re not going to beat Usain Bolt, but they figure it’s the Olympics, why not, what a lark!
They all run, and the top 10 get lucrative sponsorship deals, as before.
But the average and median will change. 10 out of 20 is 50% who earn money. 10 out of 1000 is only 1%. The overall amount of money given out stays constant, but the average earning of a runner competing in the 100m sprint drops way down, weighted by all those hundreds who didn’t earn anything. And the median will now be zero, because most of the sprinters didn’t earn a sponsorship deal.
So the question is, is the situation better or worse than it was before?
The answer is that it’s pretty much the same. The people who never had a chance to win lucrative sponsorship before still don’t have a chance. They’re dragging the stats down now that they’re being counted, but their earnings were always 0 and that hasn’t changed. The top athletes in the world were always competing with those people, it’s just that in the past they out-competed them to get a slot in the top 20 in qualifiers, now they’re out-competing them directly, on the day, in the race.
For the top athletes, the situation isn’t that much different. There will be some added competition from athletes who just missed the top 20 cutoff during qualifiers, but they had to out-compete them to get a spot in the top 20 anyway, so it’s just another qualifier.
The point here is that lowering the bar to entry and letting in a bunch of low-performers can dramatically alter the stats, simply because you’re counting people you never counted before. The median earnings of an Olympic runner will tend toward zero the more people you let run in that race, yes, but it doesn’t actually mean it’s really any harder to be a top earner.
It’s still as hard as it was before, ie fucking hard.
Which is why I don’t think you can really point to that graph as proving anything, besides that more devs are getting their games on Steam now.
Which we already knew.
What I’d be interested in, personally, is a graph showing number of indie devs earning above a certain amount, maybe a couple of hundred K, for the last few years. If that number has dropped, maybe The Doom is upon us. But I have a sneaking suspicion it’s actually gone up. That’s just my gut instinct from having followed the indie scene for a while, it seems like more indies are making a decent living in recent years, regardless of the “median” dropping. But that’s just a suspicion, it could be wrong.