Stats > Browser Trends

This discusses browser trends, resolution trends, colour-depth trends, and Java and JavaScript trends, for designers who want to decide what to support.

Caution : browser stats are of limited use to designers. First, stats will vary from site to site, so it is only the stats for your sites which matter. Second, stats can be skewed by many factors, so the true numbers may be higher or lower than the numbers reported. Third, and most important, browser stats are really only useful to designers when deciding which browsers are so little used that they need no longer be supported: it may be entertaining to know, for example, how well Chrome, Firefox, and Safari are competing with Internet Explorer, but so long as the numbers are large enough that the browsers must be supported, the exact numbers are irrelevant to the design.

Browser Trends 

This discusses trends in the usage of the major families of browsers.

Summary

1% or more use each of the following: Gecko-based browsers (mainly Firefox); KHTML-based browsers (mainly Chrome and Safari); Internet Explorer 6, 7, and 8; and Opera.

Note : the stats sources used here primarily attract users from the United States and Canada. Browser usage varies by country: for example, there is strong evidence that Firefox and Opera are used more in Europe than in North America.

Note : total IE numbers started dropping in mid-2004 as users switched to the Gecko based browsers.

Note : data from several sources suggest that ~0.5% of page accesses are by users with mobile devices, e.g. cellphones, with the numbers growing fairly quickly.

The Curse of Yesteryear’s Browsers

Many people use browsers which are no longer supported, and which therefore have security defects which have not been fixed. This encourages the spread of malware. Some reasons for this are:

• Inertia : some people are content with what they have, and do not fully appreciate the benefits of using a browser which is up to date.
• Ignorance : some people are ignorant about computers, often depending on others to occasionally maintain the PCs and keep them up to date. This factor becomes greater when the update is viperware, since special skills may be required to prevent or undo the unwanted effects of viperware.
• Old Operating Systems : browser makers often don’t support older operating systems, which prevents users from upgrading to the safest, most capable browsers. For example, IE 7-8 are not available for anything older than Windows XP SP2.
• Pirated Operating Systems : browser makers may also not supply browser updates, including security updates, to those who have pirated software. For example, Microsoft does not offer security updates for Windows XP older than SP2, but people with older, pirated versions of Windows XP cannot upgrade to SP2, so they cannot get security updates.

For example, a study of one stats source in Nov 2008 revealed that 97% of Opera users were using Opera 9, but only only 39% of these were using the latest version of Opera 9.

The use of older browsers not only complicates site design and testing, but also maintains a pool of vulnerable PCs which can act as a reservoir for malware.

Recommendation

A good way to ensure that sites will work for as many users as possible is to (a) design sites to the HTML, CSS, DOM, and other standards, (b) to test sites with common browsers that implement these standards well, and (c) to tweak the sites so that they work well enough for antique browsers still in common use.

This author has found that it is easy to make sites which only require special coding for Internet Explorer.

Details

As the Stats page shows, the percentage of users using each browser depends a lot on the site. The percentages listed here are, at best, guesses as to which browsers ‘typical’ users are using, and in any case, what matters most is not the exact percentage using a browser, but rather which browsers are used by enough people that sites should support these people. You should use the stats for your site, and remember that, if few people are using a particular browser, it may be that the browser is so obsolete that few use it; or it may be that you aren’t adequately supporting the browser, and its users are going elsewhere.

• Gecko-Based Browsers (Camino, Firefox, Flock, Mozilla, SeaMonkey, etc.): • The reported percentage of users varies a lot, largely because different sites attract different types of users. • I suggest that ~18% typically use Gecko browsers, but many more use Gecko browsers on sites that attract people who are more aware of the alternatives. The percentage of users is increasing slowly.
• KHTML-Based Browsers (Chrome, Konqueror, OmniWeb 4.5+, Safari, etc.): • I suggest that ~5% typically use KHTML browsers — the great majority being Safari users — with the number increasing slowly.
• Microsoft Internet Explorer: Roughly 74% use IE-based browsers, down from a high of ~94% as users switched to other browser families — mainly Gecko and KHTML. • About 42% use IE7, with its percentage growing slowly. About 32% use IE6, with its percentage decreasing slowly as users switch to IE 7-8 and other browsers. • Many still use IE5, with the number shrinking slowly as users upgrade or switch; IE5 will remain a factor for several years, especially since IE 7-8 cannot be installed on Windows older than XP SP2, and since many companies using Windows 2000, which came with IE5, are refusing to update to IE6. • Browsers older than IE5 may generally be considered extinct.
• Opera: • The reported percentage of users varies a lot, largely because different sites attract different types of users. • I suggest that ~1% typically use Opera browsers, with the number changing little, because people looking for alternatives to IE prefer Gecko- or Safari-based browsers. • Browsers older than Opera 8 may generally be considered to be extinct, since Opera users keep fairly up to date.
• Mobile Browsers: • Few people use mobile browsers, e.g. with cellphones, but the numbers are growing fairly quickly. It is possible that, in 2009, site designers may have to seriously consider making their sites friendly to these browsers, even for sites not made specifically for users of mobile browsers. • One factor making it hard to support mobile browsers is that many mobile browsers support the HTML, CSS, and DOM standards poorly and inconsistently, making it difficult or impossible to make a normal site that can adapt gracefully to the various mobile browsers. For example, few mobile browsers support the CSS @media handheld. In many cases it may be necessary to serve completely different, crippled pages to mobile devices.

Resolution Trends 

This discusses trends in the resolutions of browser displays.

Summary

Most users have 1024x768 or higher, but a large minority have 800x600. These numbers do not include those who browse using web appliances.

It is important to note that (a) the display resolution says little about the size of the browser window, and (b) users can normally resize the browser window. Consequently no particular browser window size should be assumed.

Recommendation

A good way to ensure that sites will work for as many resolutions as possible is to design sites to be resolution-independent, i.e. not to specify font sizes in absolute units (e.g. pixels), and not to specify widths in absolute units unless a width is that of a fixed-width object, e.g. a GIF, JPG, or PNG image.

Many modern browsers have zoom features which may be used, among other things, to resize pages which are designed for specific resolutions, in order to fit the entire width of the page within the browser window. When, however, a browser resizes images, image quality suffers, and therefore the user experience suffers. This makes it all the more important not to design sites for specific resolutions.

Details

At this point in time, precise percentages for the resolution of displays are not known: the stats source for these data reports that 17% of users have displays of undetermined resolutions. The reason for this is not known. It may be that many PCs have resolutions different from what the stats source expects: but this is entirely speculative. The percentages below ignore the displays of undetermined sizes, so the numbers should be taken with a copious helping of salt.

These are resolutions of special interest to website designers:

• 800x600: this accounts for ~7% of page accesses. Some users surely have old PCs, with little video memory; the rest likely have PCs whose resolutions are set lower because many new PCs default to a lower resolution: lower resolution displays are faster. The percentage will decrease steadily as higher resolution displays grow more common, but will remain comon for many years.

  Note: a PC display’s resolution is limited both by video memory and display size: more memory permits higher resolution; a larger display makes higher resolution text more readable. Typical resolutions increase, therefore, as larger memories and displays become more affordable.

  Note: the area of the browser window available for content is almost always smaller than the display. The area is reduced by the space taken up by the browser frame: the title, function bars, scroll bars, etc. The area is also reduced if the user resizes the browser, which they can do at any time, and will likely do for high resolution displays. Designers, therefore, should not assume any particular browser window size.

• 1024x768 and higher: this accounts for ~93% of page accesses. Most new PCs have enough video memory for high resolution displays, but many are set to a lower resolution because many new PCs default to a lower resolution: lower resolution displays are faster. The percentage of high resolution users will continue to grow steadily.

• Other: other resolutions are found among web appliances and some PCs. For example, my video resolution is normally 1280x1024, however, I can tilt my display 90°, making the resolution 1024x1280.

Colour-Depth Trends 

This discusses trends in the colour-depths of browser displays.

Summary

Most users have 24-bit (or more) displays (16,777,216 or more colours), but many have 16-bit displays (65 536 colours). Very few have fewer colours, and most of these likely have mobile devices.

Recommendation

You can generally design sites assuming 16,777,216 colours or more: the major thing you have to watch out for is that colours with 16-bit displays may be a bit off.

Details

These are colour-depths of special interest to website designers:

• 16-bit (65 536 colours): this accounts for ~7% of page accesses. Some users surely have old PCs, with little video memory; some may have mobile devices; the rest likely have PCs whose colour-depths are set lower because many new PCs default to a lower colour-depth. For PCs the percentage will likely drop very slowly.

• 24-bit and more (16 777 216 colours, and more): this accounts for ~93% of page accesses. Most new PCs have enough video memory for high colour-depths, but many are set lower because many new PCs default to a lower colour-depth. The percentage will likely rise very slowly.

Java and JavaScript Trends 

This discusses trends in Java and JavaScript usage.

Summary

A significant number of users have browsers in which Java or JavaScript are not enabled. The percentage has decreased somewhat over the years, but it will likely remain significant for years to come.

Recommendation

Java and JavaScript should generally not be used for things which affect a site’s functionality. They are best used to make a site look more attractive.

Details

You can find information about JavaScript, including which browsers support which versions of JavaScript, in Wikipedia.

TheCounter (Apr 2009) reports these levels of support:

Users	Java Support
13 %	none/unknown
87 %	enabled

Users	JavaScript Support
7½ %	none/disabled
92½ %	enabled

Sites which need Java or JavaScript to function will deter users who have disabled them, therefore such sites will have more visitors with scripting enabled, but exact percentages likely will depend a lot on the site, and such percentages are not available.

Note : the above numbers must be taken with more than a grain of salt, since both include large percentages in which Java and JavaScript support could not be determined.