Linked by Thom Holwerda on Tue 24th Feb 2009 13:42 UTC
Apple Whenever we're talking market share and Macs, it'll inevitably get late. There are different means of measuring market share, and different ways to interpret the resulting data, usually leading to heated debates about who is right and who isn't. Ars decided to take a look at the different methods of measurement and see what they mean.
E-mail Print r 0   · Read More · 15 Comment(s)
Thread beginning with comment 350398
To read all comments associated with this story, please click here.
Net applications sucks
by averycfay on Tue 24th Feb 2009 14:16 UTC
Member since:

Even if they are not good for measuring market share itself, you can still use them to investigate where the market is going: sure, Linux may only have 0.5% according to these figures, but if a year ago this was stuck at 0.25, that still indicates a growing trend.

Not really, at least if we're talking about Linux. The numbers change way too much month-to-month to put any kind of weight in them even if we only care about trends.

Check out Linux's share between September and October of 2008. You can't seriously believe that ~20% of Linux users suddenly switched operating systems over a one month period. Much more likely they changed which websites they were tracking (or how much weight they were putting behind certain websites) complete throwing off the numbers. That might more accurately reflect reality if you only care about absolute market share, but it makes tracking trends completely worthless. It's like switching your procedures half-way through a scientific experiment. It completely invalidates the results.

Reply Score: 2

RE: Net applications sucks
by B12 Simon on Tue 24th Feb 2009 14:35 in reply to "Net applications sucks"
B12 Simon Member since:

I'd agree that a large part of the change was some change in their methods but there was a genuine big jump to linux around that time with the increased popularity of the netbook.

In this case I'd argue it's a little from column A, a little from column B
/grampa Simpson

Reply Parent Score: 2

RE: Net applications sucks
by spiderman on Tue 24th Feb 2009 15:40 in reply to "Net applications sucks"
spiderman Member since:

If I'm not mistaken, net applications only monitors web sites in english. Is that right?
Most linux browsers hide their user agent or report wrong informations.
So what kind of trend does it report? how many english speacking people upgraded their browser, how many of them compiled it with another user agent, or how many people became paranoid enough to hide their user agent?
Many people use linux because they want security and many people just won't let net applications know which system they are running.
Moreover, it all depends on how many web sites people visit with their machine. Plus, there are many people using a linux NAT but their browser is on a Windows machine.
So what kind of market are you measuring? Linux is free, it is not in the market, at all. I multiboot 5 different linux distros and I have a linux router, does it count as 5 linux installs, 1 linux user or 2 linux machines?

I believe the real question is, what purpose those number serve? If this is some kind of market analysys for software vendors to port their software to different OSes, there is really no point. It's a no brainer. Just make your code portable and get all the market, no matter how it is shared.

Reply Parent Score: 1

RE[2]: Net applications sucks
by h3rman on Tue 24th Feb 2009 16:17 in reply to "RE: Net applications sucks"
h3rman Member since:

I multiboot 5 different linux distros ...

Lemme innerduce..
Spiderman, virtualisation, virtualisation, spiderman. :-)

Reply Parent Score: 2

RE[2]: Net applications sucks
by Kroc on Tue 24th Feb 2009 16:18 in reply to "RE: Net applications sucks"
Kroc Member since:

"Most linux browsers hide their user agent or report wrong informations."

Most sentences starting with “most”, are wrong.

Reply Parent Score: 3