How is Bounce Rate calculated?

What is bounce rate and how is it calculated

Bounce rate is one of the most commonly used metrics in web analytics, but it is also one of the most misunderstood. In this article HammaJack Co-Founder, Jacob Moran, explains what it is and how to interpret it.  

A quick intro into how Web Analytics works

Before we get into Bounce Rates and how they are calculated it’s important to know how web analytics tracks things on your website/app.

First things first

When your website loads its Google Analytics tag it does a few things:

Firstly, it fires, by default, a Pageview tag to Google Analytics.  This tag has four really important parts for our purposes.

  1. It has a Hit Type of “Pageview”

  2. It has my Client ID (different for every user/device by default)

  3. It declares our Property ID (different for every website)

  4. It does not declare whether this is a non-Interaction Hit (by default all pageviews are)

Page Load Hit Google Analytics.png

On the Google Analytics Side

Google Analytics receives this and starts a new “Session” for that client ID on that Property ID and the clock starts ticking on the Session.

How is Bounce Rate calculated by Google Analytics?

Now that’s out of the way, Bounce Rate will hopefully make a bit more sense.

Bounce Rate is a fairly simple metric. It is the percentage of Sessions that Bounce compared to the total number of Sessions. The calculation looks something like this;

Bounce Sessions/Total Sessions

What is a Bounce then?

That sounds a little too simple, right? It’s because we defined a percentage. The more complicated metric is Bounces and understanding Bounces is key to understanding why Bounce Rate is so misunderstood.

A bounce is calculated as a session that fires only a single request to the Google Analytics server. This can occur when a person only looks at a single page on your website and then exits without triggering any other interaction hits to the Google Analytics server.

To complicate matters more there is a time limit on these hits. By default, Google Analytics will declare a Session over if no hits have been sent in the last 30 minutes. So if someone sees one page and then loads a new one 31 minutes later that will be a new Session. This might sound like it causes issues, and you can change it, but practically it doesn’t matter that much.

Interaction Hits

Interactions hits, sound confusing, but they are a fairly straightforward concept.

On your website, your users do things all the time but some of these things are active and some of these are passive and this is what is defined as Interaction Hits or Non-Interaction Hits.

For example, a user viewing a new page is always a new interaction because the users have actively done something that has triggered a new page. However, if a banner is shown, we might want to tell Google Analytics that something happened, but the user didn’t do anything and therefore this is a Non-Interaction Hit.

This is super important in measuring Bounces because, remember, a Bounce is defined as a Session with only one interaction hit.

The line of what is, and what is not, an interaction is up to the individual organisation. At HammaJack we determine that if someone scrolls pass 75% of a page, especially a blog, they have interacted with us more than once. As a consequence, we send the following hit to Google Analytics and that session is not considered a bounce, even if they only see one page.

Scroll Event Hit Google Analytics.png

As a default Google Analytics events are interactions, but you can and should declare them Non-Interaction Hits (like we do when they scroll to 50%) if they aren’t an active interaction.

A few examples to make sure we’ve got it

Let’s just quickly touch on a few examples to make sure we get it.

Scenario: a user comes to the site, sees a page and then another 2 in the next 5 minutes
Bounce: Not a Bounce

Scenario: a user comes to the site, sees a page and then leaves
Bounce: Bounce

Scenario: a user comes to the site, sees a page then clicks a popup that triggers a non-interaction hit event
Bounce: Bounce

Scenario: a user comes to the site, sees a page then clicks a popup that triggers an interaction hit event
Bounce: Not a Bounce

Why is Bounce Rate a misleading statistic?

Bounce Rate is such a misleading statistic for a number of reasons, but the main one is that people don’t understand how it’s calculated and therefore reference it in scenarios that it doesn’t apply too. For example if we are having a discussion about landing page performance, or advertising performance, then the Bounce Rate is a very important metric to discuss. If, however, we are talking about a middle of the funnel page then Bounce Rate is far less significant.

Let’s go through an example, and remember Bounce Rate is calculated as the number of single interaction hit sessions/total sessions.

Entrance Metric Google Analytics.png

The above screenshot shows a few of our pages’ performance over time. As you can see they all have pretty varied Bounce Rates. However, Bounce Rate is important to talk about for the first three pages but not for the /who-we-are/ page. The reason for this is the Entrances column. Even though there has been 48 Unique Pageviews for that page there has only been 9 Entrances on this page meaning that the Bounces on this page (to get a Bounce Rate of 44.44%) was a grand total of 4.

This is the most common mistake people make when reporting on Bounce Rate.

Consequently, if your Entrances don’t make up a large enough percentage of your total Unique Page Views then your Bounce Rate is pretty meaningless.

What’s a normal Bounce Rate

A lot of clients ask “so, what’s a normal bounce rate?” but they may have well asked, “how long is a piece of string?”.

The truth is it really does differ depending on your industry, your site’s content and what kind of marketing you do. For example, if you run a blog site (check out our article about reactive content planning) that does a lot of Display ads (you should really talk to us) and just have a standard Google Analytics implementation then a bounce rate of 65% or higher might be normal. On the other hand, this number may be far too high for another kind of site.

The key is to know what too low of a bounce rate is. I.e. if you have a bounce rate of under 10% it’s usually a good indicator to a Web Analyst that something is broken with your tracking. Most likely you have more than one interaction hit (like two pageview tags) on, or just after, page load.

Most Important Point

Bounce Rate is a useful statistic when understood. But, you have to understand how Bounces are calculated (single-session hits) to really unlock their value. If you’ve found this article interesting, please subscribe to our regular Get Digital updates.