Go to Top

Visitor Behavior Analysis: That darn Back Button

The browser back button is both a blessing and a curse. It is a welcome sight when navigating a poorly-designed site – imagine, for example, if every time you walked into a room, only to remember you’d forgotten to bring something important in with you before you’d taken even a full step, you had to pivot, re-orient yourself to the door (checking above it for that Exit sign to be sure) and then take proceed back from where you came from. Not so efficient. The back button saves us from futilely seeking poorly labeled or non-existent links.

But the back button is often the opposite of efficient for analysts – if you’re relying on logfile analysis and haven’t controlled caching on your site through HTTP headers in the server’s configuration, you won’t be able to find those areas of your site that have visitors looping til their heads spin.

(btw, I have to thank Jeremiah Owyang for his post that models search engines using dogs, which, given that I’m a huge dog lover, inspired me to continue to build the presence of dogs in search engine marketing with this post).

Most script-bug analytics programs negate the effects of browser caching, allowing you to get a sense of where these problem areas are in a site. But there’s a catch – these programs, like Google Analytics, make it difficult to quickly download the entire set of navigation patterns for your site. The Google Analytics navigation summary gives the percentages of visitors to a page who came to and continued to specific other pages in the site, but each page must be examined separately. This isn’t really a surprise. For a site with hundreds of pages, displaying the comprehensive range of navigation summaries for the site would be a long page indeed.

Thus for some questions, turning to the actual logfiles can be easier, assuming you’ve had caching turned off for a long enough period of time. If you’re like me, this isn’t such a bad thing after all. After all, the fun of web analytics lies in inferring the relative strengths and weaknesses of a site, as well as the feelings of users about the site, from data on a session basis.

Looping can be inferred by looking for just that – loops in the structure of the graphed navigation patterns. Statistics then identify which pairs of pages have higher than would be expected occurrences of loops, if you assumed that occasional back buttoning is a natural part of navigation.

facebooktwittergoogle_plusredditpinterestlinkedinmailby feather

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>