Bounce rate has become a popular metric lately. SiteCatalyst 15 includes it as one of the default reports in the Site Overview dashboard, it's a hit with anyone working on SEO/SEM strategies, and Avinash himself called it the "sexiest web metric ever" (see links below).
I find myself explaining the finer details of the bounce rate metric quite often and would like to compile a concise post that helps people understand how to interpret the numbers a little better.
Bounce rate is the percentage of single-page visits
or visits in which the person left your site from the entrance (landing) page. Put another way, it represents the ratio of visits that ended on the same page that they started, without clicking or viewing anything else, divided by total visits.
For me, and for the type of traffic we see at our organization, bounce rate is much more meaningful when we look at traffic coming from a search engine or social site. This is traffic that has some level of curiosity and has come to your site, maybe for the first time ever, looking for something, expecting to find what they are in search of, or anticipating a certain level of goodness (maybe trusting the source that shared the link via social media, for example). Yet, once they have clicked the search result or socially shared link, they quickly realize that the page is not what they were looking for, did not meet their expectation, and did not fulfill their anticipation for goodness, and they hastily leave your site.
We see a certain level of bounce from direct traffic (bookmarked, URLs typed in by hand, etc.), particularly on our homepage, and I am prone to shrug it off. Why? Analysis has shown that much of this bouncy traffic may have our homepage as the default homepage for their browser. Traffic coming direct to your site is more likely familiar with what they will find and represent a different portion of the audience that is less significant in my opinion, when it comes to bounce rate at least. Not worthless traffic, and by no means should it be ignored. But it is less meaningful and actionable to chase the bounce rate associated to direct traffic.
The metric feels especially confusing at our organization where we have dozens of report suites, many of which are for subsections of the large corporate website. So it will appear that you have a bounce on your page when you are looking at your specific report suite, when really the visit crossed through multiple pages within the organization. This is ultimately a fragmentation problem that we have to deal with, and will be able to address with the ability to segment in SiteCatalyst 15. It can also be viewed using a global report suite to get a broader context of the visit.
When it comes down to it, bounce rate is a confusing but very actionable metric. Changes can and should be made on key landing pages to decrease the number. Landing page may be ranking for keywords that are misleading to the visitor. Social network campaigns might not be targeting the correct audience or are giving the audience misconceptions of what they will find. The page might not have any other calls to action, useful navigation, or overall relevance, and most of those things can be fixed.
In a data filled world with no shortage of numbers, metrics, charts, and reports, it's nice to have one that you can do something about, once you understand it.
Other helpful blogs on the subject of bounce rate:
Great post! I had a question regarding how you make this actionable. Have you tried implementing A/B testing on some of these pages and looking at the bounce rate between those tests? Or are you making one change, keeping it for awhile and then comparing the bounce rate to the rate before you made the change?
ReplyDeleteA/B Testing would be a fantastic idea and a great way to tinker with the page. I'm not entirely clear with how A/B testing would affect SEO, so if that were the focus, you might get mixed results.
ReplyDeleteIf the goal is to get the visitor to do something on your page, then rearranging elements or promoting your calls to action would be very doable with A/B Testing.
I wonder if there is a way to tell the testing software to deliver the current page to the search engines every time? I'm not sure how sophisticated Omniture's A/B testing software is but maybe it's something to look into...
ReplyDelete