I get a lot of statistics thrown in my face by various marketing groups trying to make a point one way or another. As an endless critic, I can’t help but question every number or “fact” presented to me without any background data.

Last time around, I mentioned why the carrier reviews from Consumer Reports seemed fatally flawed. Today, I stumbled upon something new that I find almost as questionable. The iPhone blog posted a pie chart from AdMob, which apparently displays the percentage of web traffic coming from various mobile devices.

According to the chart (click the image to view full size), the iPhone makes up a very healthy 55% of all web requests from mobile devices here in the United States.
Android follows with 20%, RIM Blackberries with 12%, and moving down to the bottom of the list are WebOS at 5% and Windows Mobile accounting for a measly 4%.

I am not surprised one bit that Apple’s iPhone is in the lead. It is clearly the most popular consumer oriented smartphone in the country. What DOES surprise me is the Android numbers, especially compared to Windows Mobile and WebOS.
In fact, speaking of WebOS and WM, the very notion that Palm’s new system has a higher percentage of traffic than Windows Mobile indicates that something is awry with these “facts”.

These statistics represent WebOS at a time when it is on a single branded device, the Palm Pre on Sprint. Sprint, meanwhile, is lagging behind in third place with national carriers as far as subscribers. Windows Mobile, meanwhile, may not be as media spotlight friendly as the other mobile OSs, however we’re dealing with an OS that exists on every one of the 4 major carriers, and in multiple handset styles by multiple manufacturers.

So, yes, it would seem strange to me that WebOS, as great as it may be, is reportedly being used for more web traffic than Windows Mobile. That got me thinking… how is this test being performed? What are the details behind these numbers? What sites are being monitored and what determines the the data?

(Click to read on…)


The first thing that stuck out at me is that iPhone, RIM, Android and WebOS have one thing in common… they all have only one choice in web browsers.
That makes traffic tracking very simple- if the browser’s user agent said iPhone Safari, you know the request came from an iPhone. If it says Blackberry, you know its RIM.
The thing is, Windows Mobile browsers aren’t all the same. The pocket Internet Explorer that has traditionally shipped with Windows Mobile is severely limited and archaic. Much like desktop Windows users who prefer third party browsers like Firefox and Chrome, Windows Mobile users have shifted away from using pocket IE, and all current HTC devices from the last few years have shipped with the Opera browser as the default instead. That means many users of current HTC devices are likely never to use Pocket Internet Explorer at all!
Not only does Opera not identify itself as the standard WM browser, it actually pretends to be a desktop browser and identifies itself as Windows NT! Look at the following examples of browser ID tags, taken from the same WM device:
Modzilla/4.0 (compatible; MSIE 6.0; Windows CE; IEMobile 7.11) 480x640; XV6850; Window Mobile 6.1 Professional;

and…

XV6850 Opera/9.50 (Windows NT 5.1; U; en)

That’s just the difference between the new mobile IE and Opera. There’s also netfront, Torch/Iris, and server-side compressed browsers such as Opera Mini and Skyfire, which has requests coming from outside the mobile device entirely and surely can not be traced to a mobile platform.
In fact, Torch/Iris (no longer being produced, but one heck of a great webkit-based browser) identifies itself as an iPhone running Safari, so that you can you load the iPhone versions of pages on your WM. Could this be artificially inflating the iPhone’s percentages? Probably not enough to effect the numbers, but the point is clear- it is difficult to tell whether or not these numbers are accurate, especially with an OS as modifiable as WM.

To me, what this pie chart shows is that 4% of web requests to their sites are coming from pocket Internet Explorer, most likely. The rest of the Windows Mobile requests are either getting mixed up in that 3% of Other, or not identifying themselves as mobile devices at all (this chart only shows mobile device requests)!

I, for one, would like more information about how these tests are being conducted before we base our market perspective on it…

This entry was posted on Monday, November 23rd, 2009 at 10:07 pm and is filed under Mobile Culture, Mobile Web, Windows Mobile, iphone, WebOS. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.