Is Linux market growth stagnated (at least)?

Is Linux market growth stagnated (at least)?

No matter how hard I’ll try, I will never be able to determine what is the real number of Linux users, and share it with you. This article is not about market shares directly, nor about judging or analyzing the reasons behind the users hesitation in installing/using/adopting Linux.

This article is meant to provide some food for thought, concerning the (probable) surcease of Linux growth. I am not saying I know something for sure here, I am just having this feeling that 2011 was the best year for the Linux desktop and will try to share a maybe somewhat premature picture, using multiple information sources. Note that when referring to Linux I mean GNU/Linux distributions and not the kernel alone so Android will be considered something different here.


I will begin with data from the popular web developer information website. What we see here is the Linux market share from 2003 to 2012. The website provides monthly statistics but I took the liberty in creating a diagram showing the average value for every year to get the whole picture.

As you can see, 2012 is lower than 2011 and will only get even lower till the end of this year as Linux users tend to be more during the summer (probably thanks to holidays/more home usage/more experimentation etc).

These statistics derive from and represent the visitors so in reality Linux users market share is way smaller as users that visit this website are technology enthusiasts, developers etc…


Continuing with, a website that allows us to get an idea about how many visitors a website has and how the numbers of the visitors changed from 2011 to 2012. On the below diagrams you can see how the most popular distributions and did.




What we see here is a general decrease of visitors for almost all popular distributions websites that actually shows the interest decline. What I believe to be extra characteristic is the reduced interest in newer releases compared to elder releases. What we certainly don’t see here is actual growth. There is no sign of growth of interest even for the most popular Linux distribution that is Ubuntu and even when the latest release is an LTS one.


Checking Google Trends only confirms the validity of the Alexa data. Trends is based on Google Search to show how often a particular search-term is entered. I entered various Linux-related keywords and they all gave a decline of interest from 2004 to 2012.




As a side note, I could say that Linux Mint and Mageia Linux are the only major distributions that show rise in both Alexa and Google Trends but are still away from the numbers of Ubuntu (especially Mageia).


Then I took a look on where Linux users were below 1% Worldwide so there isn’t even a line to see. Alternatively I checked Linux friendly nations where Linux counted 1% to 2.5%, and saw no real/significant trend for growth. In contrary, you can notice a small decrease here.

————————————————————————————————————————————– was my next destination and the results were not good once again. Linux is showing no real sings of market growth from 2011 to 2012. Instead, you can see a small decline here too.


What I really feel is that Linux is actually experiencing a small market decline at a time that our market share should be steadily increasing. Is this just a bad period, or the beginning of an even greater decline? Is this something that has to do with exogenous factors or with Linux itself? Is Linux market share really in fall or is this an illusion?

As statistics are really no proof of anything, I can draw no conclusions. I give you the above numbers, graphs and percentages as a motive for further discussion.

  • robert pogson

    I have no doubt the web stats are being manipulated. NetApplications, for instance, shows 1.59% for GNU/Linux in USA, but, if you subtract Sunnyvale, California, home of Google (~10K employees) the number drops to 0.71%. Clearly, the numbers do not reflect actual share of PCs running various operating systems but simply clicks. If one burrows down to communities where whole school divisions or governments (many thousands of seats of GNU/Linux) there’s not even a blip. The stats just are not counting GNU/Linux seats very well.

    Go to and see GNU/Linux is the best-selling of the best-selling desktop machines. Still Brazil with a government preferring GNU/Linux and having 500K seats in schools and there’s hardly a blip, just 1.48%. Cuba with an actual trade embargo gets only 3.93%. etc. Wherever you look in places with huge roll-outs of GNU/Linux there’s scarcely a hint in the web stats.

    One way this happens is if all the units in a school are NATed down to a single IP-address and counted as one machine. Home-owners/consumers buy on retail shelves Wintel and count nearly 1:1. Who knows how the web stats are accumulated? They are not a good indicator of actual share. They do show trends however, and you can see Google switching to GNU/Linux back in 2010 if you look. You can’t see any of the other big roll-outs however… so I suspect NetApplications counts GNU/Linux at Google because they are coming from a business domain during office hours or such. They don’t tell us.

    If you look at Wikipedia for stats, they are heavily biased to USA/Europe and do not represent South America, Asia etc. emerging economies with less lock-in. With the lock on retail shelves in many places, GNU/Linux is doing real well, but Canonical stated they sold x% of PCs with GNU/Linux last year and next year expect 5%, so someone is clearly wrong if both numbers are supposed to represent share.

  • Aldi

    On Mediawiki stats Linux is very slowly growing (after deducting Android) and on Globalstats September is a good month, too:

    I think it is temporary issue, largely due to the “desktop-confusion” in Linux at the moment: Unity is becoming usable only now. Gnome 3.6 is becoming great, too. Cinnamon shapes up and KDE is again great, too. One year ago, it was hard to recommend anything.

    OSS and Linux development is stronger than ever and it will no slow down. I had people using Ubuntu and realized: I know VLC from “PC”, I know OpenOffice, I know Firefox, I know Thunderbird. Those applications’ success is Linux’ success in the long run.

    I see more and more proprietary software coming in: In the past months many games were released for Linux. Qoppa PDF Studio closes the PDF editor gap. There is now only a speech recognition gap in Linux (=there is none really usable). Wine is now running MS Office 2010 nicely which is currently the most recent version; Wine was never that complete and fast in being compatible to MS Office!

    There are some killer “applications” coming: One is IMHO “Ubuntu for Android” ( If Canonical manages to get it on the smartphones of a big smartphone producer, Linux desktop share will dramatically increase. I hope that this will be the case. Honestly, for many tasks “Ubuntu for Android” can replace my notebook and desktop quite nicely.

    I look forward to Linux’ future.

  • Bruce_Mc

    It seems clear that the iPad and other tablets have taken sales growth from the traditional PC industry. The numbers you are seeing may just be another way of saying that there are no tablets that run Linux – other than Android, which you are excluding from your count.

    I think traditional desktop/laptop LInux will continue to decline as tablets become more popular. I’d like to see some good reference designs for Linux tablets and netbook sized notebooks: open source hardware to match Linux software. I think that would promote the growth of Linux.

  • Bill_Toulas

    I thought the same too about tablets market and how they affect the statistics from 2011 to 2012. I think this is a strong possibility.
    I hope we’ll see tablets using Linux soon. Gnome Shell and Unity are on the way :)

  • blackbelt_jones

    I hesitate Gnome 3 to speculate Gnome 3 about any possible causes Gnome 3 for a hypothetical decline in the Linux Desktop Gnome 3, as anything I could come up with Gnome 3 would be pure speculation Unity.

    Beats me, but the important thing to remember is that we’ll still be here. The Linux market share has always been tiny, but development exploded. Linux is not a business, so it can’t be put out of business.

  • MichaelADeBose

    This is temporary and only reflects a change in user habits with the introduction of tablets to the general user base. Your methodology however in my mind demonstrates the Linux community’s inability to articulate a unified message to consumers. Why does the Linux community need to be so specific when purism only hurts what could be some of Linux’s success in the minds of consumers? Remember Apple really didn’t succeed until they gave up the RISC vs. CISC type arguments, so much so that they ended up on basically the exact same hardware as Windows.. They did that because consumers don’t understand those kinds of arguments and in the face of confusion they will go with what they think is easiest to understand. Red Hat didn’t try to change the way corporate customers think. They espoused instead, certifications and acronyms. These identifiers while needed and clearly demonstrate true degrees of sysop capability, when it comes to technology in the enterprise whether you’re talking about Windows, Red Hat or Solaris, its all just acronyms that managers use without actually understanding most of the time what they’re even talking about. The only division in any corporation that is tasked to understand the tech is IT, the other divisions just talk without a clue so who could change that and why try? Linux on the other hand is constantly trying.

    Unix and unix inspired OSs like Linux are vastly successful. The way you sidelined Android’s numbers is the same way Apple acts as though OS X is somehow different. OS X correlates directly to KDE & Gnome. The rest is BSD like any other BSD or even Solaris. Apple nor Red Hat nor Ellison (Solaris & Unbreakable) are trying to make arguments about kernels, boot partitions, boot managers or anything else as byzantine and likely to confuse the public. Biggest picture terms, other than technological nuances (ZFS, jails, systemd), the biggest difference between most of the OSs is the licensing which again most people don’t read and free apparently isn’t as big a persuasion, when you look at Apple’s and Red Hat’s pricing models. The linux community has to meet general users where they are. Red Hat is baddass in the enterprise and Suse rules HPC, but you can’t sale that. Opensource doesn’t have to be a business but it does have to be business minded if it wants higher market share.

    When the Linux community is able to accept that it has much more in common with the high priced ecosystems that Apple, IBM, Oracle, Red Hat, and HP have built around Unix and unix like OSs than differences, It may have an easier time understanding what it needs to do to make its case to consumers. Besides commercial interest use opensource advancements all the time and despite having the marketing budgets to tell the world about it, they never do. Apple did get its OS Unix not UNIX certified, but that was before they realized the world doesn’t actually care. Just us.

  • a-non-e-mouse

    Why would a linux distro on a tablet make any difference? Will it have an array of apps/app store comparable to Apple and Google? The same factors which make Windows 8 adoption unlikely goes tenfold for Linux.

  • a-non-e-mouse

    But a lot of financing and technical help for Linux comes from business, and if businesses decide it isn’t worth it, Linux is going to get hurt.

  • tuqui

    at the DESKTOP!?, Yes, Linux failed to grab any significant share at the DESKTOP without doubt. But on the servers, mobiles, clouds & supercomputers. The share is Growing.

  • Chris Kari

    “tablets have taken sales growth from the traditional PC”
    “desktop/laptop LInux will continue to decline as tablets become more popular”


  • Bruce_Mc

    “Will it have an array of apps/app store comparable to Apple and Google?”

    That’s a very important question. An ARM based Linux netbook would be relatively easy to port existing Linux apps to – as sort of an interim step. With an ARM based Linux tablet, I suspect it would take a lot of work to get an app ecosystem going.

  • blackbelt_jones

    Happens all the time. Individual businesses have been “deciding that Linux isn’t worth it” for 20 years. Has it hurt Linux? Well, it’s been bad for the market share. But we’re still here.

    There’s no doubt that having a tiny market share is bad for the market share, but Linux is supposed to be technology driven, not market-driven, and the technology has been developing to the point where it’s kind of annoying.

  • Pingback: Is Linux market growth stagnated (at least)? | Unixmen - cPanel Knowledge()

  • Zombie Killer

    Web stats are really not all that meaningful. You have two kinds of stats from the web; user-agent stats, and search term stats. User-agent stats are meaningless, because they’re skewed by many factors, such as false user-agents, and the type of user who visits particular sites generating the stats. Search terms WILL decline because Linux is offering a far far stronger “out of the box experience”. It used to be a struggle to set up a Linux box, try installing it and the installer doesn’t load, search for solution. Install it and most of the hardware doesn’t work, lots of searching to find the solutions for all that. Eventually muddle your way through to a usable system. A modern Linux distro, however, is basically a matter of sticking in the disk and following the prompts, the end result is a system that 90% of the time, works fully, 10% of the time requires *minor* tweaking.

    The decline in the stats is probably due to the improvement in the Linux experience, as well as the decline in the desktop. Because you insist on counting Android separately from MS, you should know that Android devices are outselling MS and apple **COMBINED**. This really speaks of a fundamental shift in computing. Desktop computers (and laptops) may not actually be declining in numbers, but they most certainly are in PROPORTION.

    Now to get a truly accurate count of Linux installations, unfortunately, requires that you have access to MS and apple OS registration and update checking statistics. MS more than apple, of course. I forget where I saw the reference, but there was some documentation from MS that actually placed Linux market and user share ABOVE that of apple. How do they know? Well every new computer that gets sold with an MS license but never checks in… they can calculate to 10 decimal places how many new computers get their MS wiped clean.

  • John Shepherd

    Several Linux distros – such as Ubuntu and Fedora, already run on ARM, and several more – such as OpenSUSE and Sabayon – are in beta or release candidate stage. Linux kernel 3.7 will see one kernel being able to support several different ARM architectures, which will make ARM distro ports even easier.

    The question about app stores seems odd. An ARM distro would use the same package manager the x86 distro uses. Tools like OpenSUSE’s Open Build Service (OBS) have long been able to cross compile packages for ARM (and has been used by Meego, Tizen, etc. for just that). When I say OpenSUSE is readying an ARM release, I don’t just mean a kernel and desktop – I mean the vast majority of apps in its repositories. Ideally, running Linux on ARM should be little different from running on x86, something Windows can’t promise (at least in terms of program cross-compatibility).

  • John Shepherd

    Open source is the opposite of the walled garden though. The more users enter a walled garden, the worse it is for the users past a certain point because vendor lock-in and monopolistic practices begin to take hold (just as in a real walled garden, more people is nicer to a point and then it starts to get crowded and more restricted). With open source, since it’s people-powered, the more users there are the better things get as the more folks are contributing in every way from coding and translations to bug reports, testing, and brainstorming new features. Open source may not be able to die off so long as one interested (and capable) user remains, but it can be hurt by a decline in users.

  • Bruce_Mc

    “Ideally, running Linux on ARM should be little different from running on x86, …”

    For an ARM notebook or netbook, I agree. I suspect that more work will be required, on both the OS and apps, for a good experience on tablets. If the work can be automated or easily done, that would be great.