Net Marketshare Shows Highly Misleading Stats. Again

January 10, 2012

I’m getting really tired of arguing against Net Marketshare stats. Every month they post the same type of hugely misleading stats. But this month they’ve really surpassed themselves in how much they are misleading people. There are so many things wrong with their stats, I barely know where to start.

Let’s start with the misleading tactic they keep using over and over again.

1) Browser Usage vs Browser Units

If we were comparing just PC browsers, or just phone browsers, maybe I wouldn’t argue much against this, but even then, counting “browser usage’ or better said “time spent in a browser”, is really a pretty meaningless and useless statistic. Who really cares about time spent in a browser? Do developers care? No. Why would they? They care much more about what type of browsers there are out there, and how many “units” of that browser are out there, to see if it’s even worth his time to make the website optimized for it.

So now that we’ve established that browser usage is pretty useless to begin with, here’s how they mislead people into thinking there are many more iOS devices than Android devices, because this is exactly what people think when they see their charts. I know it. They know it. And they still do it.

So because they are counting time spent in a browser, this gives iOS a huge advantage, that is also very unfair. Why is it unfair? Because again, it misleads people into thinking it’s about device units, when it’s not. The iPad has a much larger screen than a phone, and therefore people spend a lot more time browsing from site to site, than they do on their phone, on a daily basis. iPad browser usage is actually a lot closer to a laptop/PC than a smartphone, and even much higher than a feature phone, and I’ll get to this soon, because this is another misleading tactic they’ve used.

 2) Feature phone vs Smartphone

Most of the world’s phones are still feature phones, like 70% of them. And Android is only about 50% of the other 30%, so about 15% of total. And yet they compare the user base of the feature phones with that of Android smartphones, and they point out that they are about equal. How is that even possible?

It’s possible because, again, we’re not talking real units here, we’re talking “browser usage”, and in total the time spent in a browser on a feature phone is also a lot less than the time spent on a smartphone’s browser. So even though there are still a lot more J2ME feature phones than Android smartphones, the time spent in total, is about equal for each right now. Of course, they make it look like Android is “losing” to J2ME phones, which is pretty ridiculous, because there are already several times more feature phones than Android phones.

3) Not counting real browsers

They are counting J2ME vs Android vs Opera Mini vs iOS (which is just Safari Mobile). This is completely ridiculous, because J2ME is not a browser and it’s not an OS either. It’s a Java layer put on many proprietary phones, with different operating systems, and different browsers. And the reason I’m mentioning this is because Opera Mini is actually one of those browsers on J2ME phones, and it’s by far the most popular one. So a lot of the Opera Mini marketshare overlaps with J2ME, which makes their numbers very flawed once again.

Another “surprise” flaw is that they are counting “stock Android browser”, even though many users use other browsers, including Dolphin HD, Firefox Mobile, Opera Mobile, and yes even Opera Mini.

With all these flaws, I’m surprised Android doesn’t come out even shorter in their stats. I don’t know why Net Marketshare is doing this, but I would attribute it with bias for Apple and against Android. If this would’ve happened just once, maybe I would’ve given them the benefit of the doubt, but this keeps happening for at least 6 months since I first noticed them doing this, and I’ve noticed the Apple bias in other similar stats of theirs, too. So when seeing the Net Marketshare stats from now on, I would suggest immediately disregarding them.


  • Demonwall74

    ishit paid for them to make these state.

  • Leif

    Another point is that the Android Browser often declares itself as Webkit Android Mobile Safari. Wouldn’t wonder if some stats count Mobile Safari as iOS Device.

  • Someone

    Don’t forget these statistics also don’t count native application usage.

    If I use, for example, the cracked reader, it wouldn’t count.

  • Guest

    Here’s you: “Wahhh!!! Wahhhh!!! Life’s not fair!!!”

  • Abc

    IMO, unless the device is running Android 2.2 or higher, it might not report what it is with sufficient clarity to the web host. The host would CERTAINLY be unable to differentiate a phone from a tablet at that level. I would personally discount any such survey that doesn’t isolate numbers for 2.2 or higher.