12 ms·

IE10 below 1% market share, Firefox back under 20%, Chrome recovers from losses

31 points by iProject 14y ago

mtgx 14y ago
How does Net Apps measure browser market share that they get such huge difference for IE, which they put at 54%, compared to StatCounter who puts it at around 30%? To me the StatCounter numbers seem more accurate.
- hashtree 14y ago
  "To us, it makes more sense to keep track of users than of page views. Nevertheless, for December 2012, StatCounter listed Chrome as first with 36.53 percent share, IE in second with 30.88 percent, Firefox in third with 21.93 percent, Safari with 7.89 percent, and Opera with 1.26 percent."
- dbaupp 14y ago
  (For anyone interested, statcounter's stats are here: http://gs.statcounter.com/ http://gs.statcounter.com/) "IE6 somehow managed to gain 0.06 percentage points", which may indicate the NetApps data isn't very reliable. (maybe?)
  - Drakim 14y ago
    Depends on the margin of error. If it's outside the margin of error it would indeed raise some eyebrows.
  - bdcravens 14y ago
    There will always be some testing of IE6 by developers on clean WinXP VMs. As testing becomes easier and easier (see apps like browserling.com) there may be increased usage of older browsers.
- vibrunazo 14y ago
  I always forget what the difference is, so I always go back to this same wikipedia page to remind me of the differences: http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#Differences_in_measurement http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#Dif... You can click around on each source, and they have an explanation of how each source get their numbers and adjust them. It's very insightful to understand the differences in measurement. For example: > StatCounter statistics are directly derived from hits (not unique visitors) from 3 million sites using StatCounter totaling more than 15 billion hits per month.[17] No weightings are used. > Net Applications bases its usage share on statistics from 40,000 websites having around 160 million unique visitors per month. The mean site has 1300 unique visitors per day. On 1 August 2009, Net Applications began weighting its raw data based on the number of internet users in the countries concerned, using data from the CIA ; the changes were applied retroactively to older data.
- azakai 14y ago
  > How does Net Apps measure browser market share that they get such huge difference for IE, which they put at 54%, compared to StatCounter who puts it at around 30%? To me the StatCounter numbers seem more accurate. Apples and oranges. Net Applications measures users, while StatCounter measures usage. It is entirely possible both are right, and the different numbers are both accurate about what they measure. It is also possible one or both is wrong, possibly by a large amount - we don't have any good data against which to validate them (that data exists, but it isn't public - Google and Facebook know the truth).
  - mtgx 14y ago
    Net Apps seems to be doing it the other way around for mobile browsers, though. Because iOS always comes out with something like 60% market share, and Android with 20%, even though there are more Android devices than iOS devices now, but the iPad browser usage is much higher than the browser usage of any phone, and Android doesn't have that many tablets in the market yet.
  - mddw 14y ago
    The numbers that matter are your websites' numbers and no others. I had to make a website for a brand with strong sales in China. The client asked for a full IE6 support "because chinese use IE6 according to X and Y." IE6 support is expensive. I was quite doubtful, the brand is a luxury one, kinda confidential, so the user had to belong to the new rich class which can afford a real computer. But they did not have any data and did not want me to gather data on their actual website to confirm or infirm my hypothesis. The IE6 friendly website lauched. The stats reports less than 4% IE7 and less usage (including funny browser derived from IE.)
- meaty 14y ago
  This is because you can't measure it accurately. It's akin to religious interpretation. Unfortunately there is plenty of room to divide people and draw up conversations for the analysts to sell. Browser market share is so incredibly off. I know of 50000 IE7 users who don't even put their browser near the internet.
- jlgreco 14y ago
  The StatCounter measure seems more accurate to me too, if only because their numbers are more in line with Wikimedia's numbers. And I mean, who doesn't use Wikipedia? I imagine few people are in a better position to measure browser usage than they are, possibly not even google (though possibly with their ad network?)
TazeTSchnitzel 14y ago
IE10 below 1%? I guess that will change when the Windows 7 upgrade comes through. And then finally we will have reasonable support for CSS3, WebSocket and ES5 Strict Mode for Windows 7 IE users.
- mtgx 14y ago
  But all of those should've been available in IE9. IE seems to be at least one development cycle (of their own) behind. I see Chrome 10 had better HTML5 support than IE10 has now, and it was released around the same time with IE9. We probably won't see support for WebGL, WebRTC, Opus, Web Audio, WebM, and other features that Chrome and Firefox have for another year and a half, until IE11. I'm going to assume they won't support the web crypto API at least until IE13.
  - TazeTSchnitzel 14y ago
    >But all of those should've been available in IE9. IE seems to be at least one development cycle (of their own) behind. I see Chrome 10 had better HTML5 support than IE10 has now, and it was released around the same time with IE9. You have to consider that WebKit has supported many of these things for years, which have just been very recently written from scratch for IE9 and IE10. Come on, give them some credit, they're making a good effort to catch up.
    - Drakim 14y ago
      I am still pretty confused as to why a company so big as Microsoft keeps shipping a browser that scores a weak forth place in terms of these things. It also doesn't boost my confidence that they aren't releasing the newer Internet Explorer versions on older versions of Windows (where browsers such as Firefox and Chrome run just fine) making it harder to get a larger part of their userbase to upgrade to the newest (while it's rare to find people who use old versions of Chrome).
      - TazeTSchnitzel 14y ago
        They do release new versions on older versions, but only those which are still supported. XP isn't, so that's how it is.
Toshio 14y ago
It always leaves me dissatisfied to see the word "popular" in the context of web browsers, when people actually mean to say "ubiquitous". You know what would be a very interesting statistic? Take all windows PCs, look at usage of non-IE browsers, then take all Macs, look at usage of non-Safari browsers, then take all Ubuntu PCs, look at usage of non-Firefox browsers. Tally it all up and you'll have a reasonable gauge of popularity, where "popular" is defined as "people actively choosing this browser over the default one". I strongly suspect Chrome would utterly dominate that statistic.
- nolok 14y ago
  I agree with your premise but not with your solution. For exemple, IE would never have any point in your "popular" statistic.
  - recuter 14y ago
    Of course it would in the cases where you run a newer version then the one that shipped originally on your windows install.
    - tjoff 14y ago
      Since newer versions are available in windows update it hardly presents a challenge or signals an active choice to update IE any more than trying to keep your system up to date / secure.
- glass- 14y ago
  > I strongly suspect Chrome would utterly dominate that statistic. Of course it would, because you would be ignoring a chunk of the other browser's users but counting each and every Chrome user.
  - Toshio 14y ago
    Interesting point, but I have to disagree. Notice I have counted windows, Mac OS and Ubuntu but I left out Chromebooks since they are browser-locked. As such there are instances of Chrome I'm excluding.
nohorse 14y ago
Luckily, all the latest browsers are good. The real effort should be to upgrade older browsers. The latest round of "Browser wars" is less about capabilities and more about brand loyalty. rise above.
xpose2000 14y ago
Another useful source for browser marketshare: http://clicky.com/marketshare/global/web-browsers/ http://clicky.com/marketshare/global/web-browsers/
melling 14y ago
Microsoft is suppose to push IE10 to consumers, thus replacing IE9 on Win7. I would expect IE10 to shoot to at least 10% sometime in 2013, hopefully, closer to 15%.
pixl97 14y ago
What percentage of Firefox users use adblock or something similar to filter themselves out of stats like this? I know that just about everything stat is filtered out by noscript or adblock on my installs.
- eknkc 14y ago
  Adblock filters stat services? Like analytics? I did not know that.
  - pixl97 14y ago
    http://easylist.adblockplus.org/en/ http://easylist.adblockplus.org/en/ Easylist+Easyprivacy cuts out the vast majority of analytics sites.
- nolok 14y ago
  Same thing for chrome too, I remember seeing a stat saying the two ad block extension were the two most popular extension on the web store. And then add extensions like ghostery and its ~500k users on chrome only, etc ... Those stats are never meant to be taken to the %, or even 5%. It only gives a raw idea of who uses what. You don't know for sure if firefox/chrome is 15% or 25%, all you know is that it is massive but not overly dominant.
btipling 14y ago
I've been working on SVG animations with d3.js and have to say that the Firefox performance is very poor. I don't use Firefox so I don't know how Firefox performs in everyday use, but I know when it comes to animating SVGs Firefox crawls and chugs along. It is so bad I'm actually going to disable SVG animation by default for Firefox users (there will be a big switch to turn it back on for people who I guess have a better computer than my new 15" MBP Retina). But maybe the performance problems are why people are abandoning Firefox. It certainly doesn't seem to eat memory anymore the way it used to, and that's good.
- pestaa 14y ago
  In a way I too feel Firefox being slower compared to light Webkit browsers, but you must be writing a resource intensive animation if a modern laptop fails to render it fluently. Perhaps not only in terms of pixels and assets, but full page reflows instead. Can you share some details of your work and requirements?
- azakai 14y ago
  > I've been working on SVG animations with d3.js and have to say that the Firefox performance is very poor. Can you perhaps make a testcase and provide a link? We can file a bug on that.
eknkc 14y ago
When IE8 goes down like IE7 and 6, I'll light a fire on my living room and jump over it to celebrate. Other than that, I don't really care. All modern browsers have great JavaScript engines, has decent standards compliance and if you can ignore IE8, it's almost a unified platform to develop for.

31 points by iProject 14y ago

mtgx 14y ago

How does Net Apps measure browser market share that they get such huge difference for IE, which they put at 54%, compared to StatCounter who puts it at around 30%? To me the StatCounter numbers seem more accurate.

hashtree 14y ago
"To us, it makes more sense to keep track of users than of page views. Nevertheless, for December 2012, StatCounter listed Chrome as first with 36.53 percent share, IE in second with 30.88 percent, Firefox in third with 21.93 percent, Safari with 7.89 percent, and Opera with 1.26 percent."
dbaupp 14y ago
(For anyone interested, statcounter's stats are here: http://gs.statcounter.com/ http://gs.statcounter.com/) "IE6 somehow managed to gain 0.06 percentage points", which may indicate the NetApps data isn't very reliable. (maybe?)
- Drakim 14y ago
  Depends on the margin of error. If it's outside the margin of error it would indeed raise some eyebrows.
- bdcravens 14y ago
  There will always be some testing of IE6 by developers on clean WinXP VMs. As testing becomes easier and easier (see apps like browserling.com) there may be increased usage of older browsers.
vibrunazo 14y ago
I always forget what the difference is, so I always go back to this same wikipedia page to remind me of the differences: http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#Differences_in_measurement http://en.wikipedia.org/wiki/Usage_share_of_web_browsers#Dif... You can click around on each source, and they have an explanation of how each source get their numbers and adjust them. It's very insightful to understand the differences in measurement. For example: > StatCounter statistics are directly derived from hits (not unique visitors) from 3 million sites using StatCounter totaling more than 15 billion hits per month.[17] No weightings are used. > Net Applications bases its usage share on statistics from 40,000 websites having around 160 million unique visitors per month. The mean site has 1300 unique visitors per day. On 1 August 2009, Net Applications began weighting its raw data based on the number of internet users in the countries concerned, using data from the CIA ; the changes were applied retroactively to older data.
azakai 14y ago
> How does Net Apps measure browser market share that they get such huge difference for IE, which they put at 54%, compared to StatCounter who puts it at around 30%? To me the StatCounter numbers seem more accurate. Apples and oranges. Net Applications measures users, while StatCounter measures usage. It is entirely possible both are right, and the different numbers are both accurate about what they measure. It is also possible one or both is wrong, possibly by a large amount - we don't have any good data against which to validate them (that data exists, but it isn't public - Google and Facebook know the truth).
- mtgx 14y ago
  Net Apps seems to be doing it the other way around for mobile browsers, though. Because iOS always comes out with something like 60% market share, and Android with 20%, even though there are more Android devices than iOS devices now, but the iPad browser usage is much higher than the browser usage of any phone, and Android doesn't have that many tablets in the market yet.
- mddw 14y ago
  The numbers that matter are your websites' numbers and no others. I had to make a website for a brand with strong sales in China. The client asked for a full IE6 support "because chinese use IE6 according to X and Y." IE6 support is expensive. I was quite doubtful, the brand is a luxury one, kinda confidential, so the user had to belong to the new rich class which can afford a real computer. But they did not have any data and did not want me to gather data on their actual website to confirm or infirm my hypothesis. The IE6 friendly website lauched. The stats reports less than 4% IE7 and less usage (including funny browser derived from IE.)
meaty 14y ago
This is because you can't measure it accurately. It's akin to religious interpretation. Unfortunately there is plenty of room to divide people and draw up conversations for the analysts to sell. Browser market share is so incredibly off. I know of 50000 IE7 users who don't even put their browser near the internet.
jlgreco 14y ago
The StatCounter measure seems more accurate to me too, if only because their numbers are more in line with Wikimedia's numbers. And I mean, who doesn't use Wikipedia? I imagine few people are in a better position to measure browser usage than they are, possibly not even google (though possibly with their ad network?)

TazeTSchnitzel 14y ago

IE10 below 1%? I guess that will change when the Windows 7 upgrade comes through. And then finally we will have reasonable support for CSS3, WebSocket and ES5 Strict Mode for Windows 7 IE users.

mtgx 14y ago
But all of those should've been available in IE9. IE seems to be at least one development cycle (of their own) behind. I see Chrome 10 had better HTML5 support than IE10 has now, and it was released around the same time with IE9. We probably won't see support for WebGL, WebRTC, Opus, Web Audio, WebM, and other features that Chrome and Firefox have for another year and a half, until IE11. I'm going to assume they won't support the web crypto API at least until IE13.
- TazeTSchnitzel 14y ago
  >But all of those should've been available in IE9. IE seems to be at least one development cycle (of their own) behind. I see Chrome 10 had better HTML5 support than IE10 has now, and it was released around the same time with IE9. You have to consider that WebKit has supported many of these things for years, which have just been very recently written from scratch for IE9 and IE10. Come on, give them some credit, they're making a good effort to catch up.
  - Drakim 14y ago
    I am still pretty confused as to why a company so big as Microsoft keeps shipping a browser that scores a weak forth place in terms of these things. It also doesn't boost my confidence that they aren't releasing the newer Internet Explorer versions on older versions of Windows (where browsers such as Firefox and Chrome run just fine) making it harder to get a larger part of their userbase to upgrade to the newest (while it's rare to find people who use old versions of Chrome).
    - TazeTSchnitzel 14y ago
      They do release new versions on older versions, but only those which are still supported. XP isn't, so that's how it is.

Toshio 14y ago

It always leaves me dissatisfied to see the word "popular" in the context of web browsers, when people actually mean to say "ubiquitous". You know what would be a very interesting statistic? Take all windows PCs, look at usage of non-IE browsers, then take all Macs, look at usage of non-Safari browsers, then take all Ubuntu PCs, look at usage of non-Firefox browsers. Tally it all up and you'll have a reasonable gauge of popularity, where "popular" is defined as "people actively choosing this browser over the default one". I strongly suspect Chrome would utterly dominate that statistic.

nolok 14y ago
I agree with your premise but not with your solution. For exemple, IE would never have any point in your "popular" statistic.
- recuter 14y ago
  Of course it would in the cases where you run a newer version then the one that shipped originally on your windows install.
  - tjoff 14y ago
    Since newer versions are available in windows update it hardly presents a challenge or signals an active choice to update IE any more than trying to keep your system up to date / secure.
glass- 14y ago
> I strongly suspect Chrome would utterly dominate that statistic. Of course it would, because you would be ignoring a chunk of the other browser's users but counting each and every Chrome user.
- Toshio 14y ago
  Interesting point, but I have to disagree. Notice I have counted windows, Mac OS and Ubuntu but I left out Chromebooks since they are browser-locked. As such there are instances of Chrome I'm excluding.

nohorse 14y ago

Luckily, all the latest browsers are good. The real effort should be to upgrade older browsers. The latest round of "Browser wars" is less about capabilities and more about brand loyalty. rise above.

xpose2000 14y ago

Another useful source for browser marketshare: http://clicky.com/marketshare/global/web-browsers/ http://clicky.com/marketshare/global/web-browsers/

melling 14y ago

Microsoft is suppose to push IE10 to consumers, thus replacing IE9 on Win7. I would expect IE10 to shoot to at least 10% sometime in 2013, hopefully, closer to 15%.

pixl97 14y ago

What percentage of Firefox users use adblock or something similar to filter themselves out of stats like this? I know that just about everything stat is filtered out by noscript or adblock on my installs.

eknkc 14y ago
Adblock filters stat services? Like analytics? I did not know that.
- pixl97 14y ago
  http://easylist.adblockplus.org/en/ http://easylist.adblockplus.org/en/ Easylist+Easyprivacy cuts out the vast majority of analytics sites.
nolok 14y ago
Same thing for chrome too, I remember seeing a stat saying the two ad block extension were the two most popular extension on the web store. And then add extensions like ghostery and its ~500k users on chrome only, etc ... Those stats are never meant to be taken to the %, or even 5%. It only gives a raw idea of who uses what. You don't know for sure if firefox/chrome is 15% or 25%, all you know is that it is massive but not overly dominant.

btipling 14y ago

I've been working on SVG animations with d3.js and have to say that the Firefox performance is very poor. I don't use Firefox so I don't know how Firefox performs in everyday use, but I know when it comes to animating SVGs Firefox crawls and chugs along. It is so bad I'm actually going to disable SVG animation by default for Firefox users (there will be a big switch to turn it back on for people who I guess have a better computer than my new 15" MBP Retina). But maybe the performance problems are why people are abandoning Firefox. It certainly doesn't seem to eat memory anymore the way it used to, and that's good.

pestaa 14y ago
In a way I too feel Firefox being slower compared to light Webkit browsers, but you must be writing a resource intensive animation if a modern laptop fails to render it fluently. Perhaps not only in terms of pixels and assets, but full page reflows instead. Can you share some details of your work and requirements?
azakai 14y ago
> I've been working on SVG animations with d3.js and have to say that the Firefox performance is very poor. Can you perhaps make a testcase and provide a link? We can file a bug on that.

eknkc 14y ago

When IE8 goes down like IE7 and 6, I'll light a fire on my living room and jump over it to celebrate. Other than that, I don't really care. All modern browsers have great JavaScript engines, has decent standards compliance and if you can ignore IE8, it's almost a unified platform to develop for.