Selecting browsers for cross browser compatibility testing

One of the more common types of quality assurance testing we do is browser compatibility testing, where we make sure that a site looks the same or at least very similar across browsers.

In order to perform this type of testing, you need to come up with a list of browsers to test against. Ideally someone would have this in mind before they made a site, but most often quality assurance is brought in at the end of a project and answers to questions like these are open.

So where do you get the information you need to make a list like this?

One option is to look at log files or Google Analytics for the existing site, if there is such a thing. This will tell you what people are currently using. The problem with using data like this is maybe people with particular browsers aren't visiting your site because it just doesn't render properly in that browser. It's useful data, but it needs to be examined before being used.

Another option is to go with general data, like something you'd find at http://www.w3schools.com/browsers/browsers_stats.asp. The problem with blindly following this information is it's an aggregate of a lot of users, and you might attract particular niches that have different browser usages. You can assume that teenage gamers would likely have different preferences in browsers than senior citizens.

One more thing to keep in mind is if you will assume the burden of supporting particular browsers, namely older browsers. Choosing to do so might restrict your ability to use things like ajax and advanced javascript and CSS. Or it might mean that you essentially need to make two or more versions of your site. Making a site gracefully degrade, or continue to be usable for those without all the latest technologies in their browsers is always a good option, and an alternative to making your site for the lowest common denominator.

Try to be inclusive, but be wise as to the expense you'll take on, possibly just for a small number of users.