Linux 46% Market Share, Windows 43.5% Market Share

After a study of operating system usage of thousands of people I have discovered that Linux has a 46% market share. Linux now surpasses Windows which is shown to have a 43.5% market share. Overall GNU/Linux distributions have taken the lead from Microsoft based on this study. Honest.

Okay, in reality I looked at my Advanced Web Statistics (awstats) web log today and saw this:

Linux has 46% market share!
Linux has 46% market share!

What I am demonstrating is that anyone can take the statistics that favors their desired outcome and use those to “prove” something. That is the great thing about statistics, one can make them “prove” pretty much what one wants. I was thinking about this after I read Steven J. Vaughan-Nichols article How Many Linux Users Are There (Really)? at today. Any honest person admits there is really no way to get a hard figure for the number of Linux users, meaning the kernel, nor the number of GNU/Linux desktop users, meaning distributions.

Companies like Microsoft can get relatively solid figures from their sales departments. Microsoft can put up the numbers to “prove” they have large sales figures and “market share”. Microsoft can buy the “analysts” that do the “studies” that “prove” Microsoft’s dominance. Of course they leave off the number of systems purchased with the Microsoft tax applied that get wiped and reinstalled with a GNU/Linux distribution. Why? More than likely even Microsoft, with its’ billions, cannot get hard figures about that. I doubt Microsoft would reveal those figures even should Microsoft have them.

Then there are sites like my own SOHO office. Who counts us? We have four custom built systems here that were built in-house from components. Three of them run Mandriva Linux and one, the server, runs FreeBSD Unix. We have one second hand Compaq laptop that also has Mandriva installed. None of the operating systems were purchased. Therefore there are no sales figures for the operating systems. Only two of the systems are used actively on the WWW. How exactly are these systems going to be counted in a “market share study”? The answer is, they are not going to be counted. Nor are hundreds of other sites just like mine. One can see, based on these simple facts, that no “study” to get a “market share” figure for Linux can ever be close to accurate.

To finish up I will leave you with this. In my office GNU/Linux has an 80% market share and FreeBSD has a 20% market share. That is all that really counts for me. It is all that really counts for anyone that has found GNU/Linux, learned its usefulness and uses it every day.

NOTICE: All comments are approved by a moderator before they show here. Please only post comments once, they will show once they are approved. It may take up to a few hours for comments to be approved depending on the time of day based on UTC -0600.

Edit Thu Feb 19 01:03:40 UTC 2009: Fix a typographical error.

Edit Thu Feb 19 20:30:05 UTC 2009: Fix a typographical error.

This article has been accessed this many times:

free hit counter
hit counter code download

Published by

Gene A.

Gene is a "Unix Guy", network technologist, system trouble-shooter and IT generalist with over 20 years experience in the SOHO and SMB markets. He is familiar with and conversant in eComStation (a.k.a. OS/2), DOS (PC, MS and Free), Unix, Linux and those GUI based systems from Microsoft. Gene is also a follower of Jesus (forgiven, not perfect), and this does inform his world view.

13 thoughts on “Linux 46% Market Share, Windows 43.5% Market Share”

  1. “learned its? usefulness” should be “learned its usefulness”
    You only use an apostrophe with “its” if you are combining “it” and “is”.

  2. Of course any stats can be used to prove anything. The key word here is “any”. That is why there are 3rd party research companies that produce market statistics with no bias.

    I think we all understand that credibility to any given statement or article is directly tied the sources of statistics.

  3. crabsncancer (comment #2), I question the “no bias” in any research for market share. But primarily the point here is no one can really get hard figures for GNU/Linux. Any claim to have such figures is seriously questionable.

  4. This is a common problem in any kind of statistical analysis. The real problem here lies in the inherent misunderstanding of how true statistics (academic) work. One can calculate values such as: alpha, chi-square, r-square, etc. These values (which are never published outside of the academic realm) can give one a gauge for the validity and accuracy of the data. In any kind of market research, you can take into account “the unknown” along with other outliers, such as those who would eventually uninstall Windows. Once taken into account the outliers will modify the ‘strength’ values that can give us a more honest depiction for what is really happening. Unfortunately, this will not be happening anytime soon in lay statistics, ho hum.

  5. All, I removed some of the approved comments as they had no real merit. Please try to at least have a point when you comment. Thanks.

  6. Daniel (comment #4), I think you may be seeing the results of faked user agent strings. Note these are hits in the image, not unique visits.

  7. Jim (comment #5), thank you for that comment. While I studied Statistics and Probability in college, way back when, I have forgotten nearly all of it. I appreciate your input.

  8. According to this statistic 100% of computers on the internet are running Linux:

    I only use Linux computers, if there’s not one there I’ll break out my laptop and connect to my home network…. I am the only one that accesses my computers from outside, if I decide to track my stats I will find that only MY COMPUTERS on MY IP ADDRESSES are accessing MY NETWORK. Since only Linux boxes access my website then only Linux computers must exist.

    Case closed, M$ no longer exists… I cant believe this crap gets published.

  9. daybringer (comment #9), I appreciate your sarcasm. However, it is humorously ironic that even though you appear to have missed the point of the article you are assisting me with making that point.

    That point being, statistics are often skewed to show what one wants to show. That is something to keep in mind when reading statistics where some “studies” say Linux has ~0.88% market share. That is highly likely to be biased on the low side based on who is doing these “studies” and who is funding them. That sort of “crap” is published all the time.

  10. Benjamin (comment #12), thank you for the comment.

    Okay, I will commission a phone poll. Do you know of any “free” services out there for this? After all, that is not usually done for “free”. Or perhaps you were going to donate the money for the phone poll? I know my small business could not cover the cost alone. Also, who are we calling? USA residents and businesses only? USA government agencies? USA Military? All the above?

    Or are you thinking larger than that? Major individual “first world” governments like USA, China, India, France, England, and so on? I would think that unless we could get a very large, world wide sample of several million people any such poll would be skewed.

    The gist of this message still stands. There is no real way to get an accurate figure for the number of GNU/Linux users. All I know for sure is it is more than one.

Comments are closed.