Bounce Rate was once a metric to be respected, maybe even a little feared. By definition it tells you the percentage of users that leave your site without interacting. As analysts, we would anxiously look in Google Analytics and watch that percentage go up or down. This once revered metric has become, at best, a metric unintentionally influenced; and at worst, a metric intentionally manipulated.
Let’s set the stage for our discussion
Many of our clients participate in Dealer 20 Groups. In these groups, non-competing dealers come together to talk about best practices, marketing strategies, experiences. Inevitably, website performance is discussed. Many of the metrics are straightforward — sessions, users, pageviews — they are what they are. Bounce rate, however, needs some interpretation. One dealer’s bounce rate is not comparable to another’s. Within a 20 group, bounce rates may range from below 20% to above 40%. So why the variance?
The first thing we need to do is eliminate bounce rate by channel as an explanation. Certain channels or sources will have higher bounce rates. For example, a source that is driving brand awareness, as opposed to user engagement. A campaign designed to drive brand awareness will have a higher bounce rate. But this isn’t what we are talking about here.
How event tracking affects bounce rate
A bounce occurs when a user leaves a page without interacting with the page. In other words, only the initial page load occurred, and the session contained only one interaction hit. This is important to understand when we talk about event tracking.
Event tracking is a way to collect data about user interaction. Events trigger when a condition is set and met. For example, you may want to know when users click a video play button. Event tracking can measure this interaction. By default, an event hit is considered an interaction hit. This means that the event hit is included in the bounce rate calculation.
Let’s see our example in action. A user lands on your page and clicks play on the video. After watching the video the user leaves your site. Without event tracking, the user is counted as a bounce. But we have an event set to track video clicks. Therefore the user is not counted as a bounce. His interaction with the page is tracked as an interaction hit. The session now contains more than one hit.
Unintentional vs. Intentional
Here is where the unintentionally influenced versus intentionally manipulated behavior comes in. At Dealer Inspire and Launch Digital Marketing, we gather data on hundreds of events. We use this data to drive strategy for our clients and make improvements to our platform. Because we have so many click events tracking, the bounce rate will naturally be lower than a site without event tracking. If we wanted to quickly and drastically lower the bounce rate even more, we could do this by adding events that fire continuously. To illustrate, we added a scroll event to a site and took the bounce rate from 38% to 15%, almost overnight! This scroll event can fire as users scroll down the page.
In addition, within event tracking, you have the ability to set events to be interactive or non-interactive. We set all click events to interactive, because they are (meaning, a person clicks their mouse button in an effort to interact with the website). However, we would set an event that is tracking a session timer to non-interactive, which would not influence bounce rate. This is our philosophy. We want as much useful data as possible. We don’t set up event tracking simply for the sake of lowering bounce rate. We have seen this several times though. A pageview event set to interactive will quickly lower the bounce rate to a very low number, as will certain scroll and timer events.
Moral of the story
When looking at bounce rate and comparing your site to another dealer site, you are probably not comparing apples to apples. Know what is influencing your bounce rate. Understand what events your website provider, third party vendors, and marketing agency have set up. And understand the intention behind it. Make sure you are gathering useful data, and not just manipulating the numbers.