Update: The final report is ready.
For years we have seen statistics telling us the size of people's screens. We know that 1024x768 is the dominant screen size, and that large resolution screens are coming.
But, this is useless information. We do not want to know the size of a person's screen. We want to know the size of the browser's content area. A person might not use the browser maximized, and other elements - taskbar, toolbars and extensions - takes up valuable screen space.
I have measured the size of the browser's content area from different sites, and this is the preliminary results (based on app. 15% of the total data collected - duration 1 month+).
Note: A final report will be published in November - based on about 8 times as much data and in much greater details (more graphs and schematics).
The size of the browser's content area. The most common browser width is 1008px, and the most common height is 600px.
The % of the screen used for the browser.
The actual browser size and percentage graphs clearly show that a vast majority of people use they browser maximized (or very close to maximized).
98% of the width is more or less the same as the full width of the browser without the scrollbars. 78% of the height is about the size of the full height without the toolbars, taskbar etc.
Updated September 5, 2006
We can also see a different behavior by the few people who use high-res screens - above 1400px (10%) and above 1680 (1%).
This pretty much destroys the myth that many people browse the net in a smaller window. But, it does illustrate a possible trend for the years to come.
Updated September 16, 2006
Over at Authentic Boredom there is a discussion going specifically about 1024x768 - and what width you can use for that resolution.
This is what my analysis shows:
The size you need to design for (based on the percentages of users you want to support):
| % | Minimum width(px) |
| 60% | 1008 |
| 70% | 1001 |
| 80% | 992 |
| 90% | 840 |
| 95% | 792 |
The above data is not based on the preliminary data, but the full set (about 82% of the data used for the report in November)
The gap illustrates two things:
The size you need to design for when making fluid or smart layout
| % | From (px) | To /px) |
| 60% | 888x544 | 1256x768 |
| 70% | 880x504 | 1264x824 |
| 80% | 864x472 | 1280x856 |
| 90% | 784x432 | 1368x896 |
| 95% | 736x408 | 1424x920 |
The minimum size you need to design for when making fixed layout sites.
| % | Minimum size (px) |
| 60% | 1008x600 |
| 70% | 1008x592 |
| 80% | 984x568 |
| 90% | 800x512 |
| 95% | 776x432 |
e.g. if you want to support 90% of you visitors with a fluid layout, you need to make it work between 784x432px and 1268x896px. Similarly, supporting 90% of your visitors with fixed layout means fitting the design within 800x512px.
A more detailed report will follow in November. The collection of data will continue until October 15, 2006 (and then I need time analyzing the data).
Note: Since this is only a preliminary finding, the figures are likely to change.
Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
CM, Yes, you are right about the Mac vs. PC behavior... In the final report I am planning to look at that specific behavior (from a platform and resolution perspective).
This analysis was just a simple sneak-peak.
Aleksandersen,
Take a look at this page:
Hi Thomas,
Yes, your findings match mine almost perfectly (thank you for writing them up - I've been meaning to for ages!).
Its shocking (to me, anyway) that such a high percentage surf maximised, or close, but I definitely see the same stats.
Funny, i ve conducted exactly the same study and i ve already treated datas of may 2006 over 130 000 users.
My study is in french but you can find some figures here:
http://www.gwix.net/web_gwix/article.asp?filID=303
http://www.gwix.net/web_gwix/article.asp?filID=300
http://www.gwix.net/web_gwix/article.asp?filID=286
Contact me, would be funny to format our figures the same way and compare them ;-)
Hi Thomas,
Great to see you're already publishing this so soon.
Just a thought: I'd love to see your %height and %with graphs on a logarihtmic y-axis. This would look a lot like what Jakob Nielsen uses to show the importance - in this case of supporting other browser sizes.
Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
Lea, Interesting
Laurent, Good article you have posted. I really like the way you made your graphs.
Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
Jesper, Thanks.
Could you provide me with an example if Jakob Nielsen logarihtmic graphs?
Heh. It is ironic that I have to do a horizontal scroll to read this article. It is a common issue for me because I often browse on my Tablet PC in portrait mode at 768x1024, maximised.
In this AJAX world I'd hoped to be able to slide closed the left-nav column. That would be just enough space to get rid of the horizontal scroll-bar.
Cool data to collect, I'm looking forward to your final results. I'm new to your site but am eager to read more!
Regards!
thanks a lot thomas Baekdal, dont hesitate to contact me if you want more figures.
I dont think a logarithmic graph will be usefull because window size follow a dispersion. I dont think u'll manage to obtain any lograithmic pattern in this case.
Hi Thomas,
Really interesting results. What was the methodolody and profile of users?
Best regards
Simon
Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
Simon, Thanks.
I will provide info about the methodolody and user profiles in the full report - coming in November.
I can't see how this "destroys the myth". The myth, and specifically what Jacob Nielsen says, concerns users **with resolutions at or above 1600x1200 pixels**. Since this kind of resolution is hardly noticeable in your results, what you really have shown (if I understand correctly) is that users with 1024x768 and 1280x1024 almost always browse with a maximized window. That's kind of obvious - for a PC user like myself, that is.
Since you're doing a more thorough analysis later, it would be lovely to see the width and height in percent of the screen size, **divided into groups, depending on resolution**. That could prove Nielsen wrong, but I have to say I think he's right.
Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
Christian, I am not trying to prove Jakob Nielsen wrong. When I mentioned the "myth". I am not referring to specifc "above" size. In general terms it seems that browsing maximized is not what the majority do.
This preliminary test does not analyse what people at different screen-sizes do - it only looks at the general picture.
I actually think that Jakob is right when he states:
Starting at 1600x1200, users rarely stretch their browser windows to the full screen.- Jakob Nielsen, Screen Resolution and Page Layout
I will take a much closer look at this in thefinal report.
Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
Updated: I made a quick analysis of the behavior Christian wrote about - specific screen resolutions vs. percentages maximized (see graph above).
It shows that Jakob Nielsen was indeed right when he said that people with very high-res screens browse in a smaller window. But, the the general picture still illustrates that this is not a common behavior.
The new graph is based on the preliminary data.
Thomas,
Nielsen uses logarithmic graphs here:
http://www.useit.com/alertbox/980405.html
http://www.useit.com/alertbox/visualizing-traffic-analysis.html
http://www.useit.com/alertbox/zipf.html
What's interesting is that you might get easier to read if you have a logarithmic y-axis on the graphs.
Nice article. Always wondered why people didn't talk about it much.
I've a small suggestion for your graphs. Drawing the curve over the peaks might show the range of sizes well.
CM Harrington - I'm a Mac user who maximises his web browser! This is, I guess, because I use a 12-inch screen iBook. It is also, because most web designers make sites that don't display well unless you do maximise your browser, at least on a 12-inch 1024x768 display.
You also say "Mac users are much more likely to drag/drop/copy/paste into more than one document in more than one application, and this behaviour requires multiple windows to be visible on the screen."
Well, yes, I drag/drop/copy/paste as you say, but it absolutely doesn't require multiple windows to be visible, and they usually aren't when I do my dragging and dropping. Thanks to Expose, what windows are visible is irrelevant. I drag and hold, I Expose all my windows, then, object still under mouse, I select the target window from the Expose-d selection, and drop it there.
Whether Mac or PC, a 1024x768 screen just ain't big enough, IMHO, for any kind of multiple-small-windows-visible type of setup.
I actually hate the way MacOS tries to 'dissuade' you from maximising windows! i.e. press the green 'maximise' button and it makes the window bigger, but doesn't fill the screen. This p*ss*s me off! Damn it, the screen is small, when I hit maximise I want the whole damn screen filled up, not part of it, thank you very much, Mr Jobs!
And I say all this as a very pro-Mac user/web designer. I love many things about it but little windows are only a good thing if you have one of those gorgeous 20-inch Mac screens to go with the concept.
Another full-screen Mac-user here. Depends on the software though, but the applications I use most (e-mail, web-browser) are almost always in full-screen mode. And the lack of proper window maximization in OS X pisses me off too on a regular basis.
Having only one application window visible at once doesn't equal single-tasking either. Fast switching between applications enables you to work effectively even without a big screen. It's rather an issue of what you prefer, and thinking that "all Mac-users do it like this" just is too narrow-minded for me to comply.
Great study nevertheless, I find this information extremely useful. Keep up the good work!
fantastic! helped bunches! your work is much appreciated.
I realise this thing is kind of old, it's 2009 now but if anyone is still following, I'd like to ask, do you have any information grouped by screen ratio as follows: 5:4, 4:3, 16:10 (8:5), 16:9, other
Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
Roberto, I don't have any specific research. But I do know that widescreen is now the dominant form of screen. But, it is important to remember that it doesn't mean that people will use the full screen for browsing.
Baekdal.com is (nearly) everywhere. Check out our social profiles:
Published: Tuesday, August 22, 2006 
FaceBook
Twitter
LinkedIn
YouTube
FriendFeed
Delicious
CM Harrington
visit my website
I don't think it's a myth per se, but rather, a fundamental difference in how people interact with their computer. I don't know *any* mac user that maximises their window, whereas I know most PC users are in a constant state of single-tasking. It's also been shown that (professional) Mac users are much more likely to drag/drop/copy/paste into more than one document in more than one application, and this behaviour requires multiple windows to be visible on the screen.
For me (anecdotes don't equal data, so take it for what it is), web browsing is much less important, than, say, working in OmniGraffle, so I will frequently have my browser window *much* smaller than the window I use for my real work, but I want both on-screen at once. Same thing with IM. It's not the most important, but I need to see a sliver of the chat window, so I can see which person just replied.
Of course, people with *very* large monitors (actually high resolution —physical size means nothing in this case), once they get out of the habit of maximising everything, may be more productive in a multi-window workspace.