Linux Reaches 1% Usage Share

je ne sais quoi writes "The April data is out for the Net Applications 'market share' survey of operating systems (more accurately referred to as a usage share). For the first time, Linux has reached 1%. This past month the Linux share increased by 0.12% which is well above the average monthly increase of 0.02%. Historically, the Net Applications estimate of market share has been lower than that of other organizations who measure this, but the abnormally large increase reported this month brings it closer to the median estimate of 1.11%. For other operating systems, Windows XP continued its slow decline by 0.64% to 62.21%, whereas Vista use is still increasing to 23.90%, but its rate of adoption is slowing. That is, this month's increase of 0.48% is well below the 12-month average increase of 0.78% and down from the peak rate of increase of 1.00% per month on average in January-February 2008. The total Windows share dropped to 87.90%. Mac OS use decreased slightly to 9.73% from 9.77%, but usage share of the iPhone and iPod Touch combined increased by 0.1%."

Comments

But seriously folks

[BadAnalogyGuy]

Should we really be including both Windows and iPhone in the same OS usage chart?

My John Deere riding mower does a bang-up job cutting my lawn (get the fuck off it), but it's not quite built for the same purpose as my around-town Escalade.

Re:But seriously folks

[TheModelEskimo]

Windows is indeed a lot like an Escalade. An overpriced, bloated, and inefficient showcase of false beauty.

And I guess the iPhone is a lot like a John Deere riding mower, too. People buy it for the brand prestige, then get angry when their neighbor goes out and buys one the next day. Because everybody knows your neighbor is a jerk.

GIGO?

[AliasMarlowe]

How do they come up with these numbers anyway? The jump from 0.90 to 1.02 is relatively large, as was the drop from 0.91 to 0.71 a few months ago. Do they have uncertainty estimates? Inquiring minds want to know.

I wonder how it breaks down...

[fuzzyfuzzyfungus]

When you look at browser numbers, it is reasonably doable to get a sense of where the browsers are being used. IE6 spikes during working hours, while FF and friends increase on nights and weekends. Fairly obviously, there is a huge difference in usage rates between workplaces, especially big ones, and the home/small business market.

I'd be curious to know how Linux's market share breaks down in those terms. Is the 1% growth assimilation of the more or less geeky home/school user? Is it j. average user with a netbook or machine set up for them by somebody else? Did a few large corporations shift 250,000 call-center seats in order to save a few bucks on what are basically just terminal emulators?

I'd be curious to know what the data actually say; because you can tell the story either way: You can say "Linux will make it in the home setting first" and argue that the home has relatively fast app turnover, few critical legacy apps, and tends to suffer from viruses/spyware/malware because it lacks professional admins. On the other hand, you could argue "Linux will make it on the corporate side first" because they have highly standardized hardware and software needs, so there are fewer driver issues and "why isn't aunt maybell's scrapbooking shareware working" issues, and professional admins can handle the tricky configuration bits. Whenever something can be argued either way, that is a sign that you need actual data.

Re:Boy oh boy!

[levell]

I can't wait! At this rate, 2024 will be the year of Linux on the Desktop!

If it increases at 1 percentage point per year (which is what is has increased in the whole of its life so far), we'll reach 100% a lot later than 2024

Re:Boy oh boy!

[orkybash]

I often wonder how easy a time people who are new to computing can have with Linux. It seems to me that re-learning can in many cases be a harder barrier to cross than learning.

Methodology? How do they measure that?

[wowbagger]

I get very suspicious of any site that doesn't go into detail on their methodology for making a claim like this.

Especially when the site seems to be a web advertiser.

Have they corrected for the fact that Linux users are more likely to be able to use a variety of ad blocking and filtering tools, and thus may not be showing up in their statistics?

I always try to be clear about exactly what I am measuring - what are these guys measuring? When they say "market share", what "market" are they referring to? "Users who see our ads?" "Users visiting this set of sites (many of which refuse to work with That Which Is Not Internet Explorer)?"

Absent a statement of exactly of WHAT this is 1%, and a statement of methodologies used to make that measurement, this is a very questionable number.

the power of the pre-loading

[Locutus]

back in the old days( ~1994 ), IBM was fortunate enough to find one or two top OEMs in Germany who couldn't be paid off by Microsoft and accepted the technically superior IBM OS/2 as their primary preloaded OS. In one short year, OS/2 had 25% marketshare in Germany.

Preloading is the game and Microsoft knows this and is willing to pay out millions in marketing kickbacks to make sure a Microsoft OS is what is preloaded instead of a Linux distro. Remember the ClassmatePC deal in Nigeria? Microsoft got caught purchasing the favor of replacing the preloaded Mandriva with Windows XP once they were delivered. Egypt took tens of millions and became a Windows-only government at the expense of the OLPC MOU for a million units. And that is just the tip of the iceberg. Microsoft just redirected billions of "R&D" funds and you know where those will likely end up? Most likely place is in the pockets of companies looking to preload Android, Ubuntu Netbook Remix, or other Linux products. IMO.

It's the preloads. So when you hear the press complaining about Linux as it came from the OEM and not about installation problems, it's game-on and most likely game-over for MSFT.

LoB

Re:Inexplicable statistical variations

[je ne sais quoi]

Hi there, submitter here. One certainly wonders what the statistical variability is, it's probably pretty high for month to month data. That's what I was trying to do by reporting the 12 month average increase or decrease. I posted a chart of that data here. Rather than look at percent usage share, this is the percent change in usage share for a given month. If it's positive, it means the OS grew, if it's negative it means it shrank.

Ultimately this is one of those things like political polling data, nobody can really know what the actual answer is. What's interesting here is that there are big bumps in all the OSes, which is the random error, but if you look at the averages, they follow what you might expect. That is, XP stopped increasing a long time ago, but didn't start to shrink (go negative) until Vista was released. Vista really is slowing in its growth, you can clearly see the peak in the average data right at Jan or Feb 2008. For linux, the latest little uptick is this newest data, which in itself is probably insignificant (as is the arbitrary 1% mark), but what is significant is that linux on average is enjoying positive growth as there's more upticks than downticks, as is OS X.

Re:Boy oh boy!

[TooMuchToDo]

That's like saying "I don't drive a semi to work, therefore no one is using them", while neglecting to understand that almost all goods are shipped via semi trailer.

Re:Boy oh boy!

[nine-times]

There are distros like ubuntu that try to make as much as possible accessible from the surface, but when you have to do something not exposed by pretty control panels you need a level of understanding far beyond that of the average user.

That's true of Windows and OSX too. Don't tell me you'd advise normal low-level users to open up the Windows Registry.

So here's the real question: how much of the functionality required by new/unknowledgeable users is not exposed by pretty control panels?

Re:Boy oh boy!

[TheLinuxSRC]

Which is *exactly* the reason why he mentions it to his clients. If someone is made aware of an issue before it becomes a problem (thus giving them all the information they need to avert the potential problem), in their mind there is no problem.

Re:Boy oh boy!

[Dunkirk]

Sounds like someone hasn't had very good luck with using Linux on the desktop. Yeah, well, me neither, but it hasn't stopped me from using it on my desktop machines (both at home and at work) for over 10 years now. You can say it sucks, but, in my experience, it just sucks DIFFERENTLY than the alternatives.