The posts about Android's overwhelming marketshare miss the greater story that Tim Cook tried to highlight during the iPhone 5S launch around about usage versus unit volume. We used to live this at HP, and especially in the printer buisness where most of the economics were not in the "placement of the socket" but instead in the number of pages printed after the "install" (and therefore the amount of ink consumed). It is why HP smoked Lexmark and has held the lion's share of the economics despite pressure from Epson and Canon in inkjet printing. Whereas analysts worry about socket marketshare as a leading indicator, the real battle is in utilization.
Today we are hardly at the same place with mobile platform ecosystems because most of the gross margin is being made up front with the sale of the devices (well at least for Apple and Samsung) and with a complex value chain that relies on carrier subsidies in a lot of their core markets. But increasingly as it becomes harder to differentiate on hardware alone and software/user experience takes front stage, the bulk of the economics will shift to the services. Most people think of this shift as represented by the paid downloads in the AppStore and Market respectively but I don't believe that this is durable over time (as this blog post) argues and will shift to in-app payments and services that make recurring gross margin completely outside of the app stores but totally enabled by the mobile platform.
Just like the momentum story for app developers is going to shift from app install to app engagement over the next couple of years, the momentum story for the platform owners is going to shift from installed base to usage in a way that may expose some fairly naked swimmers in the receeding tide of large numbers.