How market share surveys actually work/fail? (Linux vs Win)

Today I stumbled upon several surveys and Linux is hovering about 2% of the desktop market share.

Well, I know it is one of those "never ending" topics and yes it is as far as Linux exists besides Windows and MacOS. But I was questioning myself because it is kinda strange to me, if the surveys are actually made in a proper way.
Many Distros and their used browsers are masking Linux for many good reasons, if I go on
https://whatsmyos.com/ as an example, with my Garuda Gamer Dr460nized KDE super sweet wet dream OS the website tells me
Your OS is Windows 10 possibly Windows Server 2016 or Windows 11
:rofl:

Which is funny to me but in regards to surveys, and I don't know how they are made, they could be wrong possibly by large numbers. And don't get me started with those DistroHub ratings, the sheer amount of downloads doesn't tell you sh..t about the actual usage and market share imho.

So, my next thought was about other surveys like the Steam Survey which, at least can overcome any masking made by web browsers. But then there is the not uncommon habit or at least possibility to use the Windows based Steam client on Linux. So, this survey too, has at least that point of weakness. Props for the guys for Valve to separate their SteamDeck from Linux desktop users, though.

My next point which made the whole thing poking me was, how come Linux desktop market share kinda sticks around 2% for years. It is kind of an oddity by any means of statistics and in a way Proton and DXVK and Valve made a certain change within Linux and many new users appeared, objectively. Not mentioning further bad moves by M$ which even gave their hard core fan boys a hard time LOL.

I tried to not vacate in my Linux bubble and looked around on any forums and communities.
Companies like Tuxedo or System76 are working fine and are growing, hence the count of Linux users is growing as well. So there are hard facts numbers to proof that one.
Like those reports which count them 'fast growing' Annual Report on System 76's Revenue, Growth, SWOT Analysis & Competitor Intelligence - IncFact

How come Linux is actually considered about the 2 to 2.3% range while several developments in the past 3 years are moving more and more ppl to Linux.

Is this a masking issue?
It is an honest question from my side, I'd love to hear/read your opinions (or even facts) about it.

Cheers

Bruce :shark:

P.S.: Well, I still don't have any issues with my Garuda, so why not some Community topics? :slight_smile:

5 Likes

Even when I connect to whatsapp web then it identifies my OS as windows instead of linux , and my google accounts too ,


I believe that I have some identification issues on my device , that's why on many places my device is recognised as windows , I remember that gave I an exam recently held by HCL company and interview (just for fun , I never accepted their offer after clearing the exams , the salary they promised me after my 4 years of college and training which would be organized by that company was too low :stuck_out_tongue_winking_eye: ) , all of their interviews and exams where online , the first condition of my exams was that my OS should be windows and it always verified my OS as Windows but it was garuda linux :rofl: , no matter which ever linux distro i use most of the time it is recognised as window :grin:

4 Likes

I use Firefox, and

image

Probably you people use Firedragon, and it has some masking features.

I, too, was thinking the same after the 3% news. Your speculation seems plausible.

3 Likes

Oh that happens with me too, :joy: I thought I was the only ugly duckling but it seems I am not alone.

5 Likes

Interesting o_o good to hear that many seem to get masked as Windows users without having to install any additional stuff!

With my Garuda, I have to use a user-agent switcher to make it look like I'm your average Windows user. If I don't mask it then I'll appear as a Linux 64-bit user D: I use Librewolf and Ungoogled Chromium. Happens on both.
That said, the average computer literacy of a Linux user is bound to be higher than that of Windows and MacOS users, so even if they aren't automatically masked, there is a higher chance that they're using user agent switchers too. :slight_smile:

If they aren't sending out surveys to people and are finding peoples OS with this method, the results will definitely be skewed.

3 Likes

Thank you for validating my impression.

Yep, that could be a serious reason for some wonky survey data, as far as I don’t know how they are obtained.

You’re never alone :wink:

yep and this isn’t a Garuda thing alone. Personally I love that masking but I have some concerns about the actually market share values which got presented by several sources. I wonder if they got tricked by masking methods!

Exactly, that way!

Additionally: https://www.youtube.com/watch?v=1QkoA5fcrQw

Thanks for your answers I’m curious what we’ll find out further on that one,

Cheers,

Bruce :shark:

4 Likes

This is only happening in Fire-dragon Browser…and Tor Browser
Not in case of Firefox /opera/chrome…

So I think this is for privacy aspects only…(as regarded to tor browser)

firstly … i was also surprised :melting_face:…when my whatsapp web shows Windows computer in “connected devices” section… i thought is my device hacked lol :sweat_smile: :rofl:

Not in whatsapp … Even google takes it as windows computer. :rofl:

4 Likes

The reason we have been stuck at 2% for years and years is…

…wait, hang on a second…

:tada: :partying_face: :tada:

2 Likes

I know @BluishHumility I've added an YouTube Link regarding that.

We're on record levels (and so is Mac btw). But my question is still: Are this even true numbers?
B/c of reason I've written down in this thread.

Cheers o/

1 Like

There is definitely a margin of error with these estimations. In the case of StatCounter, who is cited as the source in the "After 30 Years" article above, they have tracking code on 1.5 million sites that log the browser useragent activity to get their stats. A couple obvious ways that gives skewed information:

  • They are tracking 1.5 million sites. If there is a bias in the users who visit these 1.5 million sites (like they are tracking sites where more Windows users visit than Linux users), that will skew the result.

For example, let's say we have 1000 Windows users and 100 Linux users--so Linux is 10%. This company puts their tracker on google.com, and let's say 90% of Windows users use Google for web searches, but only 80% of Linux users do. So the web tracker reports 900 Windows users, and 80 Linux users. This tracker will report Linux use at about 8.89%, not 10%.

That is some wildly over-simplified logic there, but the point is the numbers are bound to be a little off.

  • The actual site list matters. If they put trackers on i-love-windows.com and windows-is-the-best.com, but do not have trackers on i-love-linux.com and linux-is-the-best.com, that will skew the results. I actually do not see the site list published anywhere, although honestly I'm not about to go scrolling through 1.5 million URLs anyway. :smirk:

  • The tracker will report the OS as whatever the browser says it is. If Firedragon announces the OS is Windows, that will actually count as Windows. Similarly, browsers where the OS is obscured or otherwise not able to be detected are probably not able to be counted at all. This may include people using proxies or front ends to visit sites (Whoogle, for example)--I'm actually not sure on that.

Maybe it winds up being reasonably accurate, but it definitely will be a little bit off. If you want to read more about specifically how they do the counting, they describe it a little in their FAQ: FAQ | Statcounter Global Stats

3 Likes

Thank you, it literally proves my worries.

So even with conservative and super cautious estimations, Linux might be way above 3% market share?!
Is this another "big player move" from M$?
As you've been describing the way of recognition of the OS it is blatantly rigged.

Thank you for your effort @BluishHumility :smiling_face_with_three_hearts:

I think so! I mean, look at the chart:

When I see “Unknown” for the OS, all I can think is “That’s gotta be Linux!” :joy:

4 Likes

Same thought but then you have 3rd world countries with super old Windows computers as well, as some homebrew OS and (maybe) failed masking. I couldn’t even guess how much of them might be Linux driven devices tbh…(then there are routers as well, which sometimes access websites, perhaps…)

EDIT: So it could be go up to even 4 or 5% :face_with_peeking_eye: :thinking:

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.