Good question.

Let me ask you the reverse with a hypothetical: imagine that you spend a great deal of time building a library for generating realistic engine sounds, like this guy. Now you make an OEM deal with Sony and your work goes into the next version of Gran Turismo. Now let’s say everybody loves the new version, because of the great engine sound and a number of other awesome features. Would you want your work to be rewarded by how much value Sony extracted from it? You would right? (otherwise tell me why not and we’ll have that discussion, but I can hardly imagine you’ll say no to this)

Then put yourself in Unity’s position. It’s not one company you’ve got to track, but perhaps hundreds of thousands. New ones popping up, old ones dying without a trace. You want to be rewarded for your continuous effort based on how much value people are getting from your product. This is only reasonable, right? Now you’ve got to come up with a way to do that. So one way to do that would be to track the revenue of each developer and charge a percentage. This is mission impossible. Perhaps you can do that with the larger companies, who are less likely to forge data and easy to get hold of, but you’ve got thousands on thousands of developers that are making peanuts or making just enough that are one man shops. There is also no reliable way to get accurate revenue data from developers across the world. You can’t just ask the tax office of the Philippines or Norway for income statements of random developers. So instead they use a heuristic, which is very common by the way. The heuristic goes like this: revenue ∝ usage ∝ installs ∝ downloads (∝ means “is proportional to”, but in this context I think it would be better to say: “correlates highly with”) .

Now if you proof to me that downloads does not positively or significantly correlate with revenue made then I’ll agree with all the people who feel they need to hate on Unity right now, but the way I see it this isn’t an unreasonable business model.

One last thing. It is an oversimplification to say that Unity doesn’t have any cost to usage. Sure once the binaries have been built, there are no costs to Unity them being copied across the globe, but more usage means more demands on the developer, which translates to demands on Unity to make sure their engine works well on all platforms and devices and is able to keep up with the queries and demands of the developer. Imagine just having to QA the Unity engine; it’s gotta be an enormous undertaking. They’ve got to offer active support on a number of versions (n) of their platform for a number of platforms (m) and supported devices/hardware (o). That makes (n^m)^o combinations that could cause issues and then still that is an oversimplification. A game that is used a lot is going to hit a lot of these combinations and that’ll certainly translate into a lot more work for Unity to ship updates. So I would even argue that usage ∝ costs.