Note: This blog entry predates responsive design. All modern websites should be built using responsive design principles.

While working on a recent web project at work, I wondered if I should go for a fixed-width layout or stick with my preference for fluid layouts. Fixed-width layouts are certainly easier to manage, but they just feel so… rigid. With the boom in larger monitors, I also wondered if fluid sites start presenting a problem due to being too wide. I decided to check around the web to see what others are doing.

The search

In my (very quick and unscientific) research, I visited 150 popular sites, including news sites, shopping sites, software company sites, and the personal sites of well-regarded web designers and developers. I purposely avoided blogs that use canned themes.

When visiting the site, I determined the page’s width by reducing the browser window’s width until the horizontal scrollbar appeared. In some cases, I didn’t use the front page of the site, as these are sometimes standalone pages that use a completely different layout than the site’s actual content.

The results

Of the 150 sites I visited, only thirteen used fluid layouts — a whopping 8.6%.

A number of those thirteen sites weren’t completely fluid, as they either break at smaller sizes or have fixed-width sidebars, but I included them in the list since they aren’t using a fixed-width wrapper.

I was genuinely surprised about the number of fixed-width sites; a sizable chunk of these sites belong to (or were designed by) well-known web design gurus. I had assumed a number of them would use some sort of min-width/max-width CSS trickery, but I was wrong — only a very few sites (less than 3.5%) used this approach.

I was also surprised to see that the vast majority of the fixed-width sites were between 900 and 1000 pixels wide, a size that was unthinkable a mere 5 years ago.

Observations

  • Most of the fixed-width sites fell between 900 and 1000 pixels wide, and were usually centered in the browser
  • 20 sites were 800 pixels wide or less
  • 14 sites were between 800 and 900 pixels wide
  • 100 sites were between 900 and 1050 pixels wide
  • 3 sites were over 1050 pixels wide
  • the widest site was time.com at over 1100 pixels, although the layout uses a fixed-width wrapper wider than the content itself (the content was closer to 1000 pixels).

Sites visited

I picked sites somewhat randomly, using a combination of Alexa.com rankings, sites I frequent, sites my co-workers like to visit (hello, www.bebe.com), sites belonging to famous companies (McDonalds, Target, etc.) and sites belonging to well-known web designers and developers. I know I left out a number of good sites, but this was a very quick-and-dirty project.

Fluid layouts

  1. allinthehead.com (Drew McClellan)
  2. clearleft.com
  3. craigslist.org
  4. danwebb.net
  5. dean.edwards.name
  6. drupal.org (partially broken due to floating elements overlaying other elements)
  7. gmail.com
  8. htmldog.com (Patrick Griffiths)
  9. joeclark.org
  10. meyerweb.com (Eric Meyer)
  11. molly.com (Molly Holzschlag)
  12. people.opera.com/howcome (HÃ¥kon Wium Lie)
  13. wikipedia.org

Fixed-width layouts

  1. 24ways.org
  2. 37signals.com
  3. 456bereastreet.com (Roger Johansson)
  4. about.com
  5. adactio.com (Jeremy Keith)
  6. adobe.com
  7. alexa.com
  8. alistapart.com
  9. amazon.com
  10. americanexpress.com
  11. andybudd.com
  12. anthropologie.com
  13. aol.com
  14. apartmenttherapy.com
  15. apple.com
  16. authenticjobs.com
  17. bankofamerica.com
  18. barackobama.com
  19. barnesandnoble.com
  20. bbc.co.uk
  21. bebe.com
  22. bestbuy.com
  23. blogger.com
  24. borders.com
  25. boxofchocolates.ca (Derek Featherstone)
  26. burgerking.com
  27. cameronmoll.com
  28. cartoonnetwork.com
  29. cbs.com
  30. chase.com
  31. clagnut.com (Richard Rutter)
  32. cnet.com
  33. cnn.com
  34. comcast.com
  35. comcast.net
  36. crateandbarrel.com
  37. danbenjamin.com
  38. dean.edwards.name
  39. dell.com
  40. deviantart.com
  41. dictionary.com
  42. digg.com
  43. directv.com
  44. dojotoolkit.com
  45. dustindiaz.com
  46. ebay.com
  47. espn.com (MLB page)
  48. facebook.com
  49. fastcompany.com
  50. fedex.com
  51. flickr.com
  52. fox.com
  53. friendster.com
  54. gamespot.com
  55. go.com
  56. guardian.co.uk
  57. guitarhero.com
  58. happycog.com
  59. haveamint.com
  60. hicksdesign.co.uk
  61. home.live.com
  62. hulu.com
  63. iht.com
  64. ikea.com
  65. imdb.org
  66. jasonsantamaria.com
  67. jeffcroft.com
  68. jetblue.com
  69. jquery.com
  70. kaiserpermanente.com
  71. latimes.com
  72. linkedin.com
  73. livejournal.com
  74. m-w.com
  75. macys.com
  76. markboulton.com
  77. mcdonalds.com
  78. mediatemple.net
  79. mezzoblue.com (Dave Shea)
  80. microsoft.com
  81. mootools.net
  82. mozilla.com
  83. msn.com
  84. myspace.com
  85. nbc.com
  86. neopets.com
  87. netflix.com
  88. newegg.com
  89. nfl.com
  90. ning.com
  91. nintendo.com
  92. npr.org
  93. nytimes.com
  94. opera.com
  95. oreilly.com
  96. paypal.com
  97. pbs.org
  98. quirksmode.com (Peter-Paul Koch)
  99. reuters.com
  100. rockband.com
  101. rollingstone.com
  102. secondlife.com
  103. sfgate.com
  104. shauninman.com
  105. si.com (MLB page)
  106. simonwillison.net
  107. simplebits.com (Dan Cederholm)
  108. sitepoint.com
  109. skype.com
  110. snook.ca (Jonathan Snook)
  111. sony.com
  112. stuffandnonsense.co.uk (Andy Clarke)
  113. target.com
  114. techcrunch.com
  115. theonion.com
  116. time.com
  117. tivo.com
  118. twitter.com
  119. typepad.com
  120. ups.com
  121. urbanoutfitters.com
  122. usaa.com
  123. usps.com
  124. veerle.duoh.com (Veerle Pieters)
  125. virgin.com
  126. wait-til-i.com (Christian Heilmann)
  127. wamu.com
  128. washingtonpost.com
  129. weather.com
  130. wellsfargo.com
  131. whitehouse.gov
  132. williamssonoma.com
  133. wordpress.com
  134. yahoo.com
  135. yelp.com
  136. youtube.com
  137. zeldman.com (Jeffrey Zeldman)

Comments

Tom wrote on December 17, 2008 at 1:51 am:

Quite an interesting sample of sites. What did you decide for the project in the end – fluid or fixed?

We're in the process of trying to convince management of the benefits of fluid layouts for e-learning modules, rather than the traditional fixed window.

One thing that would be very interesting to know is whether the boom in large monitors necessarily corresponds to larger browser windows. My personal preference on a large monitor (anything over 19 inches IMHO) is to have the browser window sized sensibly, so that i've more space on the screen for other applications as well as the browser – IM, Email, etc.

I wonder if any web analytics packages could the visible window size of the users browser, as well as detecting the native screen resolution? It would probably be more relevant than screen res as it would take into account varying sizes of chrome from users having multiple toolbars installed. Any patterns of window size as the screen res gets larger could be very useful.

Steve Howard wrote on December 17, 2008 at 9:05 am:

I used to like totally liquid layout. But now I have a 1920 wide screen and some dumb designers still think I need to view their text-heavy sites in full screen mode. No I effin don't.

Ultimately I prefer a fixed-width site that looks roughly similar in width to a little less than half my screen width. This matches your findings and matches this site 🙂

I think the reason is pretty simple. We cannot scan wide lines of text as easily as narrow ones.

The exception to my page-width preferences are sites that are design-based – images, video etc. rather than text based.

Philip Hutchison wrote on December 17, 2008 at 11:39 am:

@tom i'll probably go with a 1000px fixed-width, but try and work in an alternate stylesheet for mobile devices. ultimately, the fixed-width approach makes the site easier to design and maintain. for instance, you always know what image dimensions will work, and it's less confusing for others who may need to collaborate with me on the site.

@steve i hear ya. my monitor at work is 24", and i *never* maximize the browser. i have to manually size it so sites remain usable. i usually make my browser about 1/2 the monitor's width.

JF wrote on January 6, 2009 at 8:14 pm:

I'm not sure it's a realy good idea to adopt a fluid layout. I know there is an issue concerning the respect of the user how visit the page, but there is an issue to in reading. I'm most a teacher (actually e-learning dev) than web developer, but I know some reader can have difficulties with long lines of text… if it's not with the text they read, it's with the motivation to read it…

Nice research Philip. Thanks

Darren Taylor wrote on January 26, 2009 at 10:40 am:

The thing is with fluid layouts, forgetting about web standards etc is that aesthetically, they look vastly inferior. Looking at the 13 above, none of them inspire me, if I were a paying customer I wouldn't accept any. I've been working with fluid layouts for over 5 years and fact remains that fixed width just look better. Now take that alone, is it not a very good reason in this endless debate to sway towards fixed, thinking of our end users who at the end of the day, are fickle and want pleasing?!

Similar Posts

5 Comments

  1. Quite an interesting sample of sites. What did you decide for the project in the end – fluid or fixed?

    We’re in the process of trying to convince management of the benefits of fluid layouts for e-learning modules, rather than the traditional fixed window.

    One thing that would be very interesting to know is whether the boom in large monitors necessarily corresponds to larger browser windows. My personal preference on a large monitor (anything over 19 inches IMHO) is to have the browser window sized sensibly, so that i’ve more space on the screen for other applications as well as the browser – IM, Email, etc.

    I wonder if any web analytics packages could the visible window size of the users browser, as well as detecting the native screen resolution? It would probably be more relevant than screen res as it would take into account varying sizes of chrome from users having multiple toolbars installed. Any patterns of window size as the screen res gets larger could be very useful.

  2. I used to like totally liquid layout. But now I have a 1920 wide screen and some dumb designers still think I need to view their text-heavy sites in full screen mode. No I effin don’t.

    Ultimately I prefer a fixed-width site that looks roughly similar in width to a little less than half my screen width. This matches your findings and matches this site 🙂

    I think the reason is pretty simple. We cannot scan wide lines of text as easily as narrow ones.

    The exception to my page-width preferences are sites that are design-based – images, video etc. rather than text based.

  3. @tom i’ll probably go with a 1000px fixed-width, but try and work in an alternate stylesheet for mobile devices. ultimately, the fixed-width approach makes the site easier to design and maintain. for instance, you always know what image dimensions will work, and it’s less confusing for others who may need to collaborate with me on the site.

    @steve i hear ya. my monitor at work is 24″, and i *never* maximize the browser. i have to manually size it so sites remain usable. i usually make my browser about 1/2 the monitor’s width.

  4. I’m not sure it’s a realy good idea to adopt a fluid layout. I know there is an issue concerning the respect of the user how visit the page, but there is an issue to in reading. I’m most a teacher (actually e-learning dev) than web developer, but I know some reader can have difficulties with long lines of text… if it’s not with the text they read, it’s with the motivation to read it…

    Nice research Philip. Thanks

  5. The thing is with fluid layouts, forgetting about web standards etc is that aesthetically, they look vastly inferior. Looking at the 13 above, none of them inspire me, if I were a paying customer I wouldn’t accept any. I’ve been working with fluid layouts for over 5 years and fact remains that fixed width just look better. Now take that alone, is it not a very good reason in this endless debate to sway towards fixed, thinking of our end users who at the end of the day, are fickle and want pleasing?!

Comments are closed.