Here are the data underlying all the calculations in the previous post. Please feel free to use them for your own calculations if you would like.
Of course, no data are perfect. ComScore data come from repeated surveys of smartphone owners. Chitika data come from tabulating hits to websites on Chitika’s ad network. There are two ways that this figure could overstate Android’s decline in Verizon net additions.
- ComScore might be understating Android’s overall growth. My ComScore estimates imply that the Android base grew 58% from Feb 2011 to Aug 2011 (from 22.9 million units to 36.1 million units). It’s possible that this is an underestimate, but that seems unlikely. In fact, ComScore’s estimates of US Android market share are higher than any others I’ve seen.
- Chitika might be overstating the reduction in Verizon’s Android share. Chitika’s numbers are not exactly right, because Verizon Android users may have slightly different browsing habits than Sprint or T-Mobile Android users. Our concern, however, pertains to changes in shares. The question is whether large numbers of Verizon Android users have suddenly stopped visiting Chitika-affiliated websites since March, while Sprint and T-Mobile Android users have not. While technically possible, this seems unlikely.
In my view, the most likely explanation for the observed data is that Android has lost momentum on Verizon. If so, I’m certain that Google knew this many months before any of us did.