
Could this mean that iOS is easier to use and get connected, there are after all many more Androids in circulation.
Or is it that iOwners have the money to play while most Android phones are owned by teenagers that have run out of credit.
Apple's iOS devices account for 65 per cent of mobile web traffic versus Android's 20 per cent, or the two operating systems are neck-and-neck at about a quarter of all mobile web traffic apiece. Yes, you read that correctly. There is, indeed, that much disparity in two recent mobile web-usage reports, one from Net …
My reaction is to assume the latter; iOS devices are much more likely to come on a contract with free 'unlimited' data than Android phones. It's a natural result of Apple not being competitive at low price points.
So what conclusions to draw? Nothing we didn't already know: failing to support either iOS or Android is going to cost you a large chunk of your potential market.
Sort of the latter. I'd imagine Apple owners are much more likely to make full use of their relatively expensive devices, and probably have a contract that allows them to do so. Android on the other hand includes a lot of budget smartphones and PAYG users with more traditional phone usage patterns.
Personally, I own a cheap burner phone that runs Android and use PAYG, and one of the first things I did was disable data.
Symbian may have been given a date of execution but that doesn't mean the huge number of devices in circulation have suddenly stopped working. What I find really odd is that Statcounter seems to think that no one with a S40 phone ever used the web until March 2012 and that the decline in Symbian since that time matches almost exactly the increase in S40.
Something's not quite right with them there stats me thinks.
Stastically speaking, 94% of all statistics are incorrect with 4% being made up on the spot, although 48% of that is made up by marketing droids and PR people, the other 52% being made up by people other than the people who made up for 48%. However, using a unique weighting system we actually see that 154% of statistics are incorrect with 205% of them being made up by frogs who have nothing to do in their spare time.
Wasn't StatCounter also the basis for Google claiming Chrome as the world's most popular browser?
I find these measures irrelevant since there's no checks and balances. They can be manipulated from the inside by StatCounter admins, or outside by someone with a sufficiently large network bandwidth and IP range.
Exactly. If you have an iPad you are probably going to be browsing with it a lot of the time. If you have a browsing capable you phone it may be in your pocket or ,god forbid, actually making a call with it some of that time.
While the scale of the measurements are large and to some extent informative (taking into account all the issues you mention) I think your use of the phrase statistically significant is loose at best.
1) The hypothesis being evaluated for significance is not clear.
2) The sampling is not random (of sites never mind all requests).
While neither tells you very much (of course), neither of them makes me want to buy their more detailed stats either as their comments imply they just want to put the other down rather than produce a useful dataset.
Both of these stats are about web browsing, and not applications, which means the underlying OS doesn't actually matter, you only care about the browser and form factor.
(Ok, iOS, Android and WP7 only have one real browser. Not true of the others.)
Net Applications bundling of "tablet" with "phone" commingles completely different form factors, thus giving no guidance whatsoever. How do I tell from that whether I should put more effort into making my site look good on a Windows Phone or Symbian? It feels like a statistic specifically invented to make iOS look good.
StatCounter on the other hand is useful, because it tells you how much effort to put into making your "for phone" version(s) of the website website look nice on anything other than the Android and iOS big two.
Again, it doesn't tell you whether it's worth working on the various tablet versions, but at least it's up front about that.
The argument about page impressions against unique users... What exactly is wrong with tracking both? I'd rather like to know if a huge number of page impressions is really only a small number of users, or if I just get a host of uniques who never come back, as these two are indicators that I'm doing different things wrong.
How do you work that out?
Android has at least 3 real browsers in my opinion - Opera Mobile, Firefox, and all those using the stock Webkit rendering engine (so are all essentially the same). I'm not even certain that Chrome for Android doesn't use it's own "internal" version of Webkit, so it is possible that Chrome for Android is a "4th" browser.
"I fully believe it, most of my & my wifes mobile access show an iPhone or iPad user agent, so it must be true."
I suspect you & your wife represent an unimaginably small drop in the ocean when compared to the millions of people who will not change their user agent.
Yes and don't forget that story about how Android devices only need to load half the web page data of iPhones to get the same result, so everyone should know these stats can be safely ignored by all Fandroids because the figures unfairly misrepresent usage.
For these studies, number of page views or visitors are not as important as
1. The number of sites tracked, and
2. The nature of the sites tracked.
This is because the sites themselves will select for specific types of users, For example, one would expect a much larger percentage of Apple devices visiting Apple's site. But it's not that black and white. Shopping sites, for example, might attract a different mix than technical reference sites, which could attract a different mix than media-sharing sites, etc.
I have trouble believing 40,000 would be a broad enough sample of sites to balance site selection bias. By having so many more sites, StatCounter protects against this somewhat, but without knowing the nature of the sites they track, we don't know how well. Ironically, this is where weighting would actually be useful -- weighting the results of each site by the relative popularity and level of general appeal of that site would help offset the site selection bias.
Net Applications shows California has 9.64% GNU/Linux share. If you ask for California - Sunnyvale you get 2.93%. The difference? Google's 10K employees switching to GNU/Linux a couple of years ago. We know there are whole school divisions and others using GNU/Linux but they count as nothing because they are not businesses connecting from business domains to business sites during office hours. There are more than 30 million people in California. Is that bias or what?
Never mind iOS vs Android, what's with the StatCounter graph showing Series40 jumping from 0% to 15% in 4 months?
Of course, it is mirrored by a similar drop with SymbianOS. But why doesn't the jump happen during a single month, like it should if StatCounter was just about changing the way handsets are counted? WTF?
And how are they working out which device is what? Esp as most Android browsers offer an easy to change the user agent, and due to lazy web monkeys, you used to have to change to imitate the iPhone, as that was the only device they tested for.
So Android user could be inflating the iOS (and desktop numbers).
Oi Rik! Mark Twain didn't come up with all the best quotes!
Accounts of the origins of this quote have been grossly misreported.
See here (http://www.york.ac.uk/depts/maths/histstat/lies.htm) for details about the actual history of the phrase. Even Twain acknowledged that he himself didn't come up with it.
Perhaps your copy should be adjusted to say, "as Mark Twain was wont to say, 'as Disraeli was wont to say, "There are ..."'"
Grumble, grumble, can you believe the cultural imperialism coming from the colonies these days, grumble, grumble.
Unless I have "Mobile View" checked in the browser on my HTC Desire S (Android). My User Agent shows as:
"Mozill/5.0 Macintosh U Intel Mac OS X 10_6_3 en-us AppleWebKit/533.16, KHTML like Gecko Version/5.0 Safari/533.16"
It doesn't look very Androidy to me, and I suspect wouldn't count as Android in most web stats! As most mobile versions of websites are crippled, I tend to keep Mobile View unchecked most of the time!
iOS devices don't offer this flexibility they "just work" - *snigger*
So obviously these sort of charts confirm what we all suspect about the decline in Blackberry, but no comment about the non appearance of Microsoft Windows Phone 7.x?
Given the amount of money poured into WP7 by MS and Nokia, the fact that it is still pretty much flat lining (even when filtering to the most 'popular' market of North America) should be a clear indication just how badly it has failed. Ballmer and Elop most be proud.
... would be my HG motorbike jacket's back pocket.
There's a pocket on the back of the presumed driver's jacket, so the pillion passenger can take a map out and peruse it at whatever 70mph+ cruising speed is applied that day. Given the demise of map reading in the last few years, the pocket is going empty so she may as well be playing angry birds now.