Skip to main content
Hero Background Image

Not all Bounce Rates are Created Equal

November 17, 2016

Bounces are often simply explained as a single page view and therefore bounce rate is the percentage of visitors that view a single page and then leave the site. It's one of the most popular metrics in user engagement and satisfaction reports because it can indicate whether or not users are finding what they need on a page. But it's often one of the most misunderstood metrics! The usual calculation for bounce rate equals total single page views divided by total visits – however, that is not actually the calculation for all analytics platforms. IBM’s Coremetrics bounce rate is made up of this calculation. However, if you are using Adobe’s Omniture Sitecatalyst or Google Analytics, what makes up a bounce and bounce rate is not simply based on viewing one page and leaving the site. Not all bounce rates are the same and it is important to know how it is calculated so that you know how your visitors are behaving.


Adobe Omniture Sitecatalyst

Adobe Omniture Sitecatalyst’s calculation is based on what they call ‘image requests’. Bounces are defined as visits to the website with a single image request. Image requests include pageviews and they also include custom links. Custom links are links that you have on your site that refers the visitor to a different site – perhaps you link to a partner site or offer additional resources on a different website. So that means if a visitor comes to your site, only views one page and leaves your site by clicking on a link, it is not a bounce because the visitor had two image requests. The first image request is the pageview and the second is the click on the link. To get to the calculation of bounce rate, there are two calculations you can use in Adobe. The first is a standard metric which takes the percentage of users that only had one image request. The second bounce rate can be created as a calculated metric which usually is created using bounces divided by entries.

Google Analytics

Google Analytics bounce rate is more similar to Adobe’s but is different and includes additional variables. GA’s bounces are based on what GA calls ‘hits’. Hits include pageviews and events. Events in Google Analytics can include a variety of things including but not limited to, outbound links, downloads, clicking on an email link and a variety of other custom events that you can set up. So if a user comes to your site and only looks at one page but completes any of the events like clicking on an outbound link, downloading a form, or clicking on an email address, they are not included in the bounce rate because they have more than one ‘hit’ during their session. So a bounce in Google Analytics is a single engagement hit. Bounce rate is the calculation of single hits divided by site entries.

One important thing to note is that you can exclude events if you do not want their interaction to impact bounce rate. You simply include a parameter known as ‘non-interaction’  on your page with the event code and this gives Google the signal that you do not want this tracked in bounce rate. Or if you are using Google Tag Manager, you will simply select ‘true’ within the event setup Non-Interaction Hit area and the event will be a non-interaction event and will not be included in your bounce rate.

Why is this important?

It is important to fully understand your bounce rate so that you avoid making the wrong assertions. For example, if I had multiple outbound links on my homepage, before knowing what was used in my bounce rate calculation, I might have thought that my bounce rate was even better than what was being presented in my analytics platform. I would think this because I would know that X number of visitors left through links I presented them so they did not actually just leave my site but they engaged with it which resulted in a single pageview.  I might have even tried to calculate the “true bounce rate” to exclude visitors that left through outbound links. But as we saw above, if you are in Google Analytics or Omniture, your data will most likely take those visitors into account and they will not be included as bounces, so no extra calculations are necessary.  

In closing, please remember the following three things; First, not all bounce rates are simply based on one pageview. Second, even if you understand your bounce rate on one platform and you switch to another, remember that the bounce rates may not be measuring the same way so it will be more difficult to trend. And finally, even within your own site, one page can have different metrics included in bounce rate than another page on your website (ie. if one page has outbound links and another does not).

Now that you have a better understanding of what can make up bounce rate, I encourage you to make sure you dig into your data and understand what elements on your site are included in your calculation.

Additional Resources
How to Find and Exclude Bots from your GA Account | Mediacurrent Blog Post
The Synergy of SEO and Web Accessibility | Mediacurrent Blog Post
SEO: Your Gateway to More Visitors | Mediacurrent Webinar