寮居酒屋

Blog | About me | Scams | The Fack | The COCk | DMCA | FALI DB | Webring | Glossary | Donate

RSS Feeds | Soycial Media | PeerTube-chan


Marketshare ANALysis is inaccurate

Return to home
寮、 2022/07/23 — blog, technology, linux, smartphone, bsd, internet

No, this is not me just being biased towards certain operating systems, browsers, search engines, and so on.
I mean the way marketshare is being measured is inaccurate.
The most popular ANALytics soyte for that purpose was Net Marketshare, which they killed off 2 years ago apparently due to a change in user agent detection, but if that was the reason, then the soyte should have seized 20+ years ago to be honest.

How are these statistics measured?
Net Marketshare already gave it away in their rather beefy notification: using user agents.
And this is no different from every other similar soyte.
In fact, I even remember 2 decades ago how a computer focussed magazine claimed that Linux had a 10% marketshare, a reader emailed them saying that this was ridiculous, because the real marketshare is only 1%, so the writers responded by showing a screenshot in the magazine showing that 10% of all traffic to their own website came from Linux users.
Perhaps just me spamming F5 or some shit?
Because I legit have no idea how a Japanese website would otherwise get this much Linux traffic from a historically extremely pro-Microshaft civilization.

Normies don’t know what a user agent is, but everyone who has ever tinkered around with browser settings before does.
In order to spoof user agent, you simply write down something else, and a very commonly recommended user agent for anonymity is making it seem like you’re using Furryfox 68 on WinDOS 10, even though you’re actually using qutebrowser on Artix Linux btw.
A statistics websoyte can’t see through that, and will count up to Furryfox and WinDOS users.
Better yet, unless you have installed some questionable rootkit, this user agent detection is limited to just visiting that soyte.
The biggest users of these soytes are probably business people checking for which market they should aim for, which are most likely using WinDOS, so while thinking they’ll find out what the market consists of, they’ll actually find out what their own peers are using to access that websoyte, so it’s kind of creates a pseodo-echo chamber, so trusting a marketshare websoyte on what people are using is like trusting bots on Shitter on what products people want to see from you.

Comment in SNS discussion here.