Bounce rates? Are we getting too caught up in calculation rather than application?

By: Garry Lee | November 24th, 2009

I’ve just finished reading an excellent little article about bounce rates (http://bit.ly/2msnIV). The guys from padicode show how to hack Google Analytics so you can record page visits greater than 10 seconds enabling you to create your own ‘bounce rate’ based on page visits lasting less than 10 seconds.

For those that don’t know this is the method of choice from the guru himself Avinash Kaushik. Avinash’s theory is outlined in his first book (Web Analytics An Hour A Day) and I believe it’s as good an example as any on what the calculation should be.

However, some have argued if the aim of your marketing is to drive people through your website, it doesn’t matter how long they spend on the first page, if they don’t go any further, then this is a failure and should be tracked as a bounce rate. (This is the industry standard calculation for web analytics vendors.) In his new book (which I recommend to all) Avinash comments on using bounce rates for email tracking based on this method!

Now, personally I’m not that bothered which method you choose (or if you have a third method I’ve not seen before). The thing that’s important about bounce rates is not the calculation, but how you analyse the data. And this means getting past looking purely at data and instead focusing on the meaning / goals of the page in question.

You cannot hope to analyse the effectiveness of a web page by numbers alone. The best way is to group your pages based on type. For example, is it a doorway page (a page aimed at getting you through to more important pages)? An information page (where engagement is the key measure)? Or maybe an instant conversion page (rare but they do exist and can be very successful)?

By looking at pages in groups you can use a combination of metrics to decide how a page is performing; bounce rates, engagement factors (number of views, time per visit),% leading to a sale. From this you can prioritise pages needing the most attention.

This last point is really the most important for me. If we are not taking this data and actually doing something with the individual pages, what’s the point of playing around with the metric in the first place? So let’s focus on application rather than spending too long on calculation. Too often it’s the media agencies buying the placements that ignore the application of the metric and focus on the number. It is then left for the client or analytics consultant to point out its importance. My point? Numbers need to be driven by the person paying the bill rather than the one being monitored!

In the meantime, if you want help or advice to analyse your site and turn data into action, please drop me an email at blog@redeye.com and please do follow my random musings on twitter.