For example, I'd suspect that people who stay on old iPhone's are less likely to spend money on apps. But it's a small, very unfounded suspicion.
I'd be curious to see if anything works out though in how stuff like that lines up.
I'm thinking specifically of a 3rd party camera app, or maybe one of these modern 3D mapping apps that uses LIDAR to model a room. These apps can (and should!) continue to work on the hardware they were developed for, even if the author bumps the minimum version past what is supported. If the app author builds in reliances on a web api that expects an up-to-date companion app, that was/is shortsighted and should be discouraged.
This makes sense if that 10% are costing more money than they generate, which seems especially likely since people that are frugal enough to (for example) hang onto an iPhone 7 are also unlikely to spend much on other frivolous purchases.
Supporting those older devices can range from supporting multiple code paths due to different APIs, or fallbacks for missing HW performance and/or features. All of that adds up in terms of developer time, build/infra cost, etc.
I remember when I was in the app-writing business, and our devs would be constantly trying to bump up the minimum OS version supported. “See here, it says that only 10% of customer devices are on OS version 10.1 and below. We must bump our minimum version to 10.2 because us developers are tired of supporting this old shit and legacy code paths!”
This seems to me to be a lot like a salesman simply hanging up on 1 out of 10 of their potential customers. Makes no sense to me, but the version bumpers seem to be winning. The number of apps that no longer support my iPhone 7 is growing and growing…