Bounce rate is a unique metric within Google Analytics. It appears within many of the Google Analytics reports and is a popular obsession for managers trying to quickly ascertain the performance of a website or landing page. However, as is often the case in any analysis, the secret to understanding the bounce rate metric is context. Taking bounce rate at face value, or assuming that a high bounce rate indicates low engagement is unlikely to deliver any realistic insights into the true performance of your content.
The definition of bounce rate
A lot of misconceptions about bounce rate start with people having the wrong definition of what it is and how it’s calculated. We have come across various incorrect definitions of bounce rate, from “the percentage of people who leave a site” to “the exit rate of the page”.
So what is bounce rate then? Well, bounce rate is the percentage of visits with only one interaction.
Lets take a moment to break this down and look at visits and interactions. Firstly, a visit occurs when someone lands on your site and the tracking code fires and that person is logged into your Google Analytics reports. A visit can contain someone engaging with one page on your website or multiple pages on your website. Google Analytics can also log a visit (or session) if you are tracking your mobile app or even if you are using the measurement protocol with Universal Analytics. There are a number of other things that play into how a visit is tracked and reported, but lets keep it simple for now.
Secondly, our bounce rate definition talked about an interaction, so what is an interaction? Well, interactions can include:
- A pageview
- An event (unless you define an event as non-interactive)
- A social plugin hit
- An ecommerce transaction
Coming back to our definition, we are looking at the percentage of visits where only one of these types of interactions occurred. Typically, a bounce will mean that someone’s visit only included the original pageview (their landing page) and no other pages or recorded events on the landing page. So by this definition, a customer who lands on an efficient landing page that provides all the information they need to satisfy their visit could end up showing as a bounce.
Another important thing to note is that by default Google Analytics does not allow you to see how long people spend on the last page someone views. This is because Google Analytics uses the second page to calculate time for the first page. When we are talking about bounces, there is no second page, so we won’t see how long people spend on the page. You can get more clarity by using events (or triggering another hit) on the page, but this is a topic for another post. The other impact on time is that visits (or sessions) also reset if there is a period of inactivity of 30 minutes or more.
Here is an example that represents a bounce rate of 50%:
This means that the following scenarios will all lead to a bounce (assuming your landing page has no event, social or ecommerce tracking):
- Someone visited your site and left immediately
- Someone visited your site, spent 5 minutes reading an article and left
- Someone visited your site, spent 45 minutes reading an article and then continued to browse the site
- Someone visited your site, read and printed an article, clicked on an email link and sent you an email about how useful it was
On the other hand, the following scenarios will NOT lead to a bounce:
- Someone landed on a page that executed the GA code and redirects to another page, then left immediately
- Someone landed on a page that triggers an event 5 seconds later that the page has loaded, then left immediately
- Someone landed on a page where the same GA code was executed twice.
The important takeaway is that your website architecture and any Google Analytics customisations can affect the bounce rate you see in your reports. It is important to set up your Google Analytics implementation in line with your business objectives, so that bounce rate matches what you expect and consider a reasonable bounce rate for your industry and business.
FAQs on using bounce rate to drive action
Should I obsess about the site-wide bounce rate as a KPI?
No. A site-wide bounce rate is an average that hides a lot of detail. Focusing on the overall bounce rate would be a little like planning a trip to Russia based on the average temperature for the year. If you pack your suitcase based on an overall average of 15 degrees Celsius (59 Fahrenheit), then you will be in for a bit of a shock if you arrive in December! The same applies to bounce rate, you need to look at individual pages or sections of your website, instead of bounce rate for your whole website.
What is a good or bad bounce rate?
Once you understand how bounce rate is calculated, it becomes more clear how misleading it can sometimes be to compare different websites. Bounce rate depends (amongst other things) on your implementation, website architecture, industry, page type and traffic source. These things need to all be considered for a proper context. Typically a website where people complete actions on a single page, like a blog, will have a higher bounce rate on average, so you can expect anywhere from 70% to 90%. For a lead generation or ecommerce site, you want to aim for a lower bounce rate, typically below 50%.
Does a high bounce rate mean that my website is not delivering what people expect?
High bounce rate can be a red flag, but it also heavily depends on the specific page. If it’s a single-serve page where people are getting all their information without the need to navigate, a successful visit will probably count as a bounce.
My bounce rate has increased. Should I panic?
The most important thing would be to find the reason for the increase. Is this increase attributed to an additional stream of traffic (for example starting to get more social media traffic)? As a business grows it will typically attract more people who are a bit further than the target demographic – more casual visitors who will drive the bounce rate up. By finding the page, traffic stream, device type or other dimension responsible for the bounce increase, you’ll be able to tell if this is a genuine worry or just natural business growth contributing to a change in averages.