Silicon Glen, Scotland > Web usability Valid XHTML 1.0 Transitional Valid CSS!

Cross browser compatibility and website design


Contents

The beginning

Netscape's downfall

The situation today

Internet Explorer left far behind

Broken websites

Throwing away customers

Cost effectiveness

Measure the costs yourself

Why your site should work on multiple browsers

This article explains some problems encountered when designing exclusively for Internet Explorer. Can you be sure that your visitors and customers are not going to migrate away from IE when it isn't going to be updated until 2006, will cost money for future updates, is already significantly behind other browsers in terms of functionality, is poor at supporting HTML standards and offers inferior facilities for disability access? For those who already use a different browser or are considering migrating from IE, are you sure your site won't annoy and frustrate them needlessly?

The Beginning

Many years ago when browsers first became widely available, Mosaic and Netscape dominated the market - Microsoft and Internet Explorer were nowhere. Those of us who started using the web in those days became used to having to download browsers and became used to Mosaic and Netscape. They might have been primitive compared to today's browsers, but in 1994 the web was pretty primitive too yet companies founded then such as Netscape and Yahoo went on to be worth billions of dollars on just the traffic from those primitive browsers.

When Microsoft eventually woke up to the Internet and launched Internet Explorer it took until IE 4 before it posed any sort of technical competition. However, the business advantage of pre-installing it on a computer gave them a big advantage over having to download 10 or 20Mb of a competitor's browser. This tactic however, was found to be illegal. I preferred using Netscape and continued to do so through the four versions of IE it took before IE posed any sort of challenge. From a moral point of view I felt it was better to support Netscape rather than the tactics of Microsoft that were illegally destroying the Netscape market share. From a technical point of view I was also able (via fortify.net) to legitimately enable my Netscape with 128-bit encryption, a feature not available in Internet Explorer to Europeans due to US export regulations in force at the time.

Netscape's Downfall

Through an agreement, AOL (which subsequently bought Netscape), made their default browser based on IE rather than their own in house browser. AOL basing their preferred browser on IE rather than Netscape was in return for AOL getting preinstalled by default on Windows. The logic of this is that there would be a better return on AOL subscribers than on a browser that had to be given away - as Microsoft was giving away IE for free, it made charging for a browser difficult.

With Microsoft pre-installing IE on Windows, and AOL using IE by default, Netscape lost market share rapidly. This was compounded by the fact that Netscape 4 was becoming hopelessly overtaken by IE in terms of support for things such as style sheets which many web designers preferred to use. The user generally had a better browsing experience in IE and Netscape's V6 replacement was not only delayed but was hopelessly buggy when it first appeared. Even Netscape fans like myself were switching to IE.

The situation today

The situation now is that IE dominates the browser space with the majority of market share although reliable statistics are fairly difficult to come by. As a result, of this domination, many websites (especially corporate ones) state "best viewed with Internet Explorer" and are not designed for other leading browser. Alternatively, they may "never" test on newer versions of Netscape, Mozilla or FireFox and ban access, for instance the Royal Bank of Scotland's Direct Banking site bans you from using any part of the site if your are using a version of Netscape newer than 4.78. Netscape 6 was released in November 2000. Just how long does it take Britain's biggest bank to test? Powergen ban Firefox 1.0 users (released November 2004), believing them to be using Netscape 1.0 (released December 1994).

You are similarly banned outright from National Lottery Subscriptions if you are using Netscape 7.1 or Mozilla 1.5 and prohibited from using the site. This is because these browsers allegedly don't support 128-bit encryption (laugh). Again an embarrassing lack of testing here, since 128 bit compatibility is built in as standard in these browsers and a reliable test for it is freely available. You are also repeatedly annoyed by alert popups when using Mailblocks, they tell you every time you log in with a non-IE browser that some features might not work. Surely once is enough guys, do customers really need told this 10 times a day?

This browser hostile behaviour is odd since all the above sites only look at the browser ident string. This string is trivial to change on Opera or with the prefbar or Multizilla plug-in for Netscape/Mozilla. A more reliable way to do browser detection is via JavaScript (assuming the browser has this capability and it is enabled). If reading the browser was important enough to ban or annoy a large number of customers, surely it is important enough to do it correctly and ensure it is the right customers you are excluding from your site? Fundamentally, however, the question needs to be asked why you really need browser detection at all? None is used on this site.

Internet Explorer left far behind

Some narrow thinking sites ban you or limit your functionality or usability if you use a non-Microsoft browser. Yet Internet Explorer has now been left well behind in terms of functionality and Microsoft has publicly stated that Internet Explorer Service Pack 1 will be the final standalone installation and there will be no update of the browser until the release of Windows known as Longhorn comes out in 2006. If you use a Mac, you won't be getting any more IE releases either. Website designers take note - the Mac community has no choice but to use a non-Microsoft browser and they have an excellent, highly standards compliant browser Safari which kicks IE out of the picture. By designing for standards rather than a browser you will allow Mac users to visit your site. More generally, designing for HTML standards and having valid HTML, such as on this page, is the basis for cross-browser compatibility. There are many free HTML validators to help you achieve this.

Already there are free browsers such as Netscape, my personal favourite Firefox, Mozilla, as well as Opera. These browsers now offer significantly better standards compliance, functionality and enhanced usability than Internet Explorer, e.g. tabbed browsing, popup blocking, cookie management, multithreading, speed of rendering and being able to set Google as my default search engine. These other browsers are also regularly updated, widening the gap between them and IE still further. With Mozilla 1.2 (now superseded by 1.7) it had 101 things you couldn't do in Internet Explorer, and this list is growing all the time. If you are desperate to hang onto that dated Microsoft browser, you'll need to wait until 2006 to get an upgrade to IE7 and it'll cost you because it'll only be bundled with the operating system. No more free Internet Explorers - it will be competing with free and superior products that are already available to download. If you can't wait until 2006, here's a preview of IE7.

Insufficient testing and broken websites

People with non-IE browsers normally have the correct security and facilities to view and use a website as intended. However, many sites with broken browser detection demonstrate a lack of sufficient testing and by default outsource this testing to their customers. This looks highly unprofessional and incorrect error messages simply compound the embarrassment to the companies in question. By leaving testing to customers rather than tackling it professionally in house these companies are seen as amateur, not the sort of image a quality company wants to project. All the more surprising considering the large number of people who use non-IE browsers, and some support for non-IE browsers may be a legal requirement due to disability access legislation.

Throwing away customers

Look at your web site statistics. If you use Webtrends it shows you the breakdown of hits by browser type. Typically this is about 4-5% for Netscape. Why bother then? It's only 4% of the market isn't it? Wrong. Measuring traffic by hits is fairly meaningless - visits and page views are far more meaningful. If you measure browser activity by these standards, you will find that of the actual visits to your site by Netscape/Mozilla users are much higher. Here are the statistics from this site (dated October 2003).

Listing browsers with at least 0.1% of the requests, sorted by the number of requests.

no. reqs %reqs browser
1 84905 68.74%: MSIE
61175 49.53%: MSIE/6
22965 18.59%: MSIE/5
660 0.53%: MSIE/4
2 8486 6.87%: Mozilla
2596 2.10%: Mozilla/1
3 6111 4.95%: Netscape (compatible)
4 5510 4.46%: Googlebot
5510 4.46%: Googlebot/2
5 3174 2.57%: Scooter
3174 2.57%: Scooter/3
6 2248 1.82%: Netscape
1982 1.60%: Netscape/4
211 0.17%: Netscape/6

From the above, Netscape, Netscape compatible and Mozilla account for (6.87+4.95+1.82) or 13.64% of all traffic including robots. Of the browser total, that's 13.64/(68.74+13.64) or 16.56%. This of course doesn't include any Mozilla and Netscape users who have altered their browser ident string to report the browser type as Internet Explorer in order to gain access to sites that try to ban Netscape and Mozilla.

Cost effectiveness

Here's the maths. Supposing you just spent $1m developing a website. If it has only been tested and only works in IE 5 and above, you may be losing a lot of your business. Using the above figures, nearly 32% of traffic is coming from sources besides IE, so you may be putting off customers with other browsers, or you may be putting off search engine robots from ensuring your site is correctly spidered.

Supposing your site attracts 20,000 visitors a day. The cost equivalent is $1,000,000/20,000 or $50 per daily visitor. If 8% of these people have a poor experience because you didn't design or test for their browser, that's $80,000 equivalent being thrown away. Developing and testing for Netscape and Mozilla even for a large site, usually amounts to less than 15 person days of effort (about $3000) and much of that effort is correcting genuine bugs, such as failure to follow HTML standards, like incorrectly using ALT instead of TITLE for image rollover text. So by spending $3000 you not only fix genuine bugs but you win back all of that $80,000 that was previously lost. Cost per customer 3000/(20000*0.08)= $1.88, a bit cheaper than $50. It's a complete no-brainer! You increase your potential market, fix genuine bugs and end up with a site that is more standards compliant. Netscape only has to account for more than 0.3% of your traffic for it to be cost effective to make your site compatible (3000/20000*0.003) = $50, your IE development cost equivalent.

Measure the costs yourself

Look at your browser stats again. If you haven't designed for non-IE browsers, chances are that people using Netscape and Mozilla are not getting the most from your site and are not staying as long. They may even think your site is poorly designed and buggy and are leaving it altogether. You can measure this yourself.

Look at the number of visits and the number of hits for each browser type. An overall trend is that people with newer browsers stay longer. You'll probably find the hits per visit for IE 6 is higher than IE 5.5 which will be higher than IE 5 and so on. Now compare these to Netscape. With Netscape 7, even though the number of visits is smaller, you ought to be seeing comparable numbers of hits per visit to IE 6. If it is significantly lower you can see that those Netscape visitors using a modern browser who should be able to get as much from your site as an IE user are simply not staying as long, probably through bad design and bugs. Surely support for any browser makes sense?

It's maybe less than $2 per daily visitor to cater for them, are you sure you want to turn them away? Using a phone analogy, would you want to slam the phone down on 8% of people who called you in the hope that they phoned back? What percentage would you slam the phone down on? Hopefully none - so why shut down your business to certain browsers? It might only take one reporter with Netscape to have a bad experience viewing your site, or one academic with Netscape doing research on your company having problems for you to miss out on a positive review or a valuable business opportunity.

 

Author: Craig Cockburn. Updated 1-November-2005
Further reading in our other Web usability articles