Originally Posted by Tallest Skil
Explain the usage reports, then.
I don't think there are any credible usage stats. They are all based on viewing pages where analysts have scripts, which are mostly just in western countries that include a higher percentage of iPhone users. Those script pages account for a small fraction of the total pages on the Internet. Couple that with the fact that by in large, Android users who do have fast web capable phones and do a lot of web surfing are perhaps spending a significant amount of time on Google services and especially YouTube, Twitter, Facebook, all of which do not count in the surveys because there are no analyst scripts on those pages as they are totally private.
I also think typical Android users have a much younger age median than iPhone users. This brings a couple metrics to the equation. One, younger people are using more social networking than older people and those pages are not part of the surveys. Also they are less likely to be able to use their device while at work since they probably aren't in management positions that allow surfing. Young people spend a lot more of their free time out and about where there is no WiFi compared to the older users being at home where there is high speed Internet. Despite LTE, surfing on cellular is slow and it uses up your data plan hence people don't do as much surfing when not connected to WiFi.
Another related reason is that since Android phones are usually a little cheaper, many users may not even have a computer, so home WiFi might not be as common.
Lastly, iPhones are handed down and have a longer lifetime so virtually all iPhones minus the original are still in active use making the total number of iPhones vastly greater than the active number of premium Androids. The premium distinction I think is important because those would be the users most likely to actually surf the Internet on their phone. The low end Androids are probably just used as phones and that is it.
Sorry for the long winded explanation but I've been trying to make sense of the statistics discrepancy for a while now and these are just some of my latest ideas on the subject.