Welcome to the all-you-can-eat-App-buffet!
Let's start off by looking at previous mobile app usage patterns. Early adopters and new mobile platform customers (Tablet and handset) are initially 'wowed' by whichever App store they visit. The wealth and availability of applications in an App store is impressive, and they are all available within a few clicks. People have little to no experience with any of these new applications now available to them, so they skip around a lot, downloading applications without a great deal of thought. We snack on different applications until we are gorged with a wide selection of them. There are no repercussions to installing a lot of these things, after all the only currencies are disk space (plentiful on modern devices) and a few pounds if you are purchasing applications. We haven't really got a plan, or a usage pattern here, it's an exciting new world of easily downloaded applications.
This slightly chaotic, unfocussed view on downloading applications gives people an opportunity to find the programs that work for them. It would be really interesting to see the discard rates for installed applications. Users try them, find that it wasn't what they thought and get rid of them. Often to then try something else very similar to see if the issues they had with the first App are fixed in a competitor or if that extra bit of missing functionality they thought was there is present.
Building Application loyalty and discovering a usage pattern
There is no App loyalty at this point. People are still adding and removing Apps until they settle on a relatively stable core set of applications. Consider the last time you upgraded a device. Chances are you simply reinstalled half of the same applications that were on the old device onto the new device. Apple and Google have actually built this data migration functionality into their ecosystems now. Its become the expected behaviour. Users develop a pattern of usage, and then they tend to stick with it. Barring significant life changes that introduce new categories of interests such as Marriages, children, house moves etc. people stay within their comfort zone. It takes a lot of marketing, or a trusted social recommendation to instigate someone to try out a new application.
A statistic that has an important bearing on this is the early adopter (or repeat upgrader) figure. New Smartphone sales are slowing. This is due to all of the early adopters having already adopted them, and in all likelihood are now on their third or fourth device due to the length of contracts. So the user base is more familiar with the way smartphone work, and integrate with everyday life. There are less and less completely new users to the ecosystem. The pattern described above is attributed to these new users, so the 'snacking' phenomenon is decreasing. This is evident in modern examples of Apple and Samsung's advertising campaigns which are both aimed at the casual user, the fringes of their traditional market segments, such as the 40+/50+ or budget conscious.
Application market Stagnation? Or User Stability?
I think when you take these factors into account the figures make a lot more sense and can be explained more as a repercussion of user stability and established practices. Experienced users have experience using online App stores and don't download a dozen applications in the first ten minutes just to test them out and remove them. Downloads are more considered and planned, time is taken to find the right application as the novelty value of having ALL THOSE APP's(!!) right on your fingertips has gone. This really helps to separate the chaff, and poor applications find themselves falling by the wayside quickly.
Rather than having a statistic around downloaded applications it may be more useful to view how many times an application was opened on a mobile device. Typically applications that are kept by a user are more heavily used. Users end out in a model of fewer applications, but more heavily used.
If you want to view more or this, or see a statistical view then have a look at either of the following two articles: