Though this will vary greatly by app type, playability (if it is a game) - here are the Active Users vs. Total downloads figures I am seeing for 2 apps:
Active Users/Total Downloads:
21% in one app
31% in another app
Daily Active Users (from AppBrain stats) / Active Users (from Google stats):
0.06
0.10
This is a comparison of figures which WILL scale similarly (regardless of the age of the app, the established user base etc.).
In comparison, New Installs vs. DAU is LESS instructive - since if your app is new you will have fewer recurring users - so will be mostly new installs (ratio close to 1.0). And if it is an old app with a huge Active User base - out of which you will be getting daily recurring users - then there will be significant recurring users and the ratio will be less (since DAU = new installs plus recurring users that day).
Although in the long-term state even the New Install vs. DAU ratio can be instructive - since eventually if you will be reaching a STAGNANT state of discoverability - then your daily installs will wind up being nearly the same day to day. Your recurring installs (coming out of your Active User base) will also be reasonably similar day to day.
You will be LOSING nearly the same users from your Active User base (due to attrition/churn) as you are gaining - in the STAGNANT case - and so you will reach “equilibrium” when your churn rate is matching your new installs.
At this point your Total Downloads will be going up and up - BUT your Active User base will become stagnant.
So you can have 1M downloads - but eventually you will get so big with your Active User base - that you LOSSES from that Active User base become dominant and swamp your New Installs - at that point you will have reached equilibrium. Active User numbers will then be STAGNANT day to day - while Total Downloads going up.
Is that pretty much what the large download users are seeing ?
Like Mind with Word Hero and David and others (though I would place David in the “buzz” category as many people would be looking for iphone launchers) - but even at that scale it should eventually be something similar (unless at the high rankings in google - there are day to day GREATER fluctuations - as one or two position move can have greater consequences - so see greater jumps from one or two position change day to day).
Or something - anyway - just some thoughts - maybe folks can put out their analysis in a similar vein …