I think it helps significantly to qualify what kind of app we're talking about (because then the demographics wanting each will vary) and revenue models. Game apps in general seem to do significantly better on the Apple store. At the very least, there's significantly more games in the Apple App Store than the Play Store. That likely has to do with the demographics that hold Android vs iOS phones. It also doesn't account for the alternative revenue model that most Android games take, that is the ad-driven model. If you're in it for the revenue, Android's ad-driven allowances (as much as I dislike them) may be a valid approach. For most other types of apps (productivity, education etc it's a toss up in my opinion, and will likely come down again to pricing model.<p>Might also help to consider where you're developing for. Different countries each have VASTLY different usage numbers for the different platforms. Big statistics that cover the whole globe aren't altogether that useful.<p>As a last point, I highly doubt that Android fragmentation is going to be the major reason why development is going to cost 20% more. That just seems wrong and arbitrary, especially with all the measures Google has taken to patch it (a la ICS). I could just as well say that the Apple app submission process is a hellish procedure that will cause massive delays and increased cost. It's really hard to say which is more costly to develop for due to all the variables.<p>Last throw in, just on an intuition, I'd develop for Android first, simply due to the significant reach in my country, greater experience with it and the type of app I'm considering. All in all, with so many variables, it's really hard to explicitly say.