Browser Statistics

Browser Statistics As a departure from our usual blog topics of software testing concerns and methodology, let’s talk a little bit about where we get some of the statistics that we use in our work. One of the more common topics that comes up in our labs when scoping out website testing is, “What browsers are people using the most?” While on the surface this seems like a simple question, the reality is a bit more complicated. While there are numerous sources on the web that provide browser usage statistics, they all share one attribute in common: they all differ greatly.

There are a number of reasons for this: some pages count hits, while others count only unique visitors. Some pages lump together desktop and mobile browsers, some break them out. Some pages have better browser detection code than others, and all sites attract different audiences with different degrees of technical know-how and interests. None of these reasons (except perhaps detection code quality) are bad or unreliable, but they all contribute to make what should be a simple question more complicated.

The bottom line is that, for browser usage statistics, the absolute best numbers are ones that you pull off of your own servers – you know the quality of your own detection code, you know what type of data you are collecting, and best of all, you know that these numbers apply to your unique user base rather than being an extrapolation from someone else’s data. However, there are always times where you need a place to start – maybe you are in the process of creating a new website, or are working on a project for which no current usage statistics are available. Where do you start?

When we’re in the dark and asked to provide input, we usually start with (what else?) Wikipedia: http://en.wikipedia.org/wiki/Usage_share_of_web_browsers. The best part about the Wiki page is that it incorporates data from a large number of major sources, making it easy to compare different sources at a glance, and also to load them directly. It is a great jumping-off point.

Another source that has served us well for many years is Browser News: http://www.upsdell.com/BrowserNews/stat.htm. This site composites several different data sources and breaks down browser share by version as well. It also has some great discussion on design and the risks of taking statistics at face value.

While it is best to gather your own browser usage statistics, the above sites are excellent resources to use for planning, as well as providing a comparison for your own data. Just remember to know how your statistics are derived, and maybe be a little conservative in your design – the exceptions are always what turn into support calls.