Website effectiveness (part 2) – using stats and analytical tools

By Matt Cornock

This is the second part of a three-part series looking at approaches to judging website effectiveness, based upon my approaches and prep notes for a meeting I had with a colleague on this matter. See introduction post.

Using stats

The problem with using web stats alone to judge how successful a website is, is that there are many other factors which may influence them. For example, coverage in a newspaper on a topic similar to your site may increase overall search traffic, which may increase overall website traffic. How do you know that the improvements you’ve made to your site have had a direct impact on increasing hits? Without clever tracking (see later post) you can’t for certain.

In addition, stats alone are useless unless you have a form of benchmarking or context, to see whether the numbers you’re getting are on a par with competition. You might use Webmaster Tools provided by Google, Bing and Yahoo! to see where your site is placed on search terms. However the recent push for personalised search results pages seems to make this a somewhat nonsensical report. Instead, look at the search terms which provide the hits (rather than its placement on the results page). Do these terms reflect your site? Are there some terms which you want to appear on but don’t? You can also use the Webmaster Tools to get a sense of where your site features on other sites. Both these combined gives you an indication of the perception of your content by the rest of the web. It’s then up to you to determine if you are being as effective as you can.

Therefore, I tend to rule out actual overall page hit rates as a meaningful indication of effectiveness.

If you’re familiar with web stats you will know about ‘bounce rate’. This is the percentage of visitors to a page who decide immediately to leave the page (i.e. bounced off). A high bounce rate is generally bad, but I can’t provide you with specific figures as ‘it depends’. For example, on a targeted, specific page coming from a targeted, specific advert, a high bounce rate is bad. However, for a page which provides one thing, which a search engine has mistakenly linked to, a high bounce rate should be expected.

There are two things which counteract a high bounce rate: clearly meaningful content and a scent trail to meaningful content. Primarily bounce rate can be used to judge effectiveness of content (is the page clearly indicating what it’s about, and for targeted pages does it meet the users’ intended need?), and to a lesser extent the navigation (when the immediate content is not useful, are there obvious leads using terminology appropriate for your audience?).

The final element to look at with your web stats is to do with the route your users take through your site. In ecommerce sites, the intended route will be for a user to look at an item, add it to their basket and purchase. In information sites, this route isn’t so clear cut. However, if you know that there are certain pieces of information which are collected on a theme, you might lead the user to view multiple pages, or undertake an action (such as sharing on Facebook). These routes can be tracked with your statistics and used to judge how effective your site navigation is. If people break off the route halfway through, then it may be indicating a problem with navigation or navigation which doesn’t seem to relate to the user (i.e. with obscure terminology).

The final post will look at user surveys and the role of interactions in judging website effectiveness.

Leave a Reply

Your email address will not be published. Required fields are marked *