Why High Bounce Rate isn’t That Bad: Understanding Bounce Metrics

Your bounce rate is up around 60-90% and you are completely desperate to get it down? But do you know whether site visitors in fact leave your site or do you just rely on the default data in Google Analytics?

Bounce rate is one of those misunderstood metrics that drive webmasters completely crazy. A website should be engaging and fulfilling its purpose, meaning that site visitors should spend xx time on it. High bounce rate means that this doesn’t happen, likely due to poor quality of content or misleading titles and descriptions. Because high bounce rate signals poor content, it leads to lower rankings. Therefore, you must get in and fix whatever needs fixing, so people are more engaged on your pages and Google likes your site. Right?

All this sounds logical and this is the reason why most webmasters believe it religiously. However, the bounce rate metric is so misunderstood that webmasters end up wasting their time on what doesn’t really matter and end up not having time for the really important tasks.

Bounce Rate: The user left the site right away. Or… did they?

The very first thing you must understand is that there are different “types” of bounce rate: the bounce caused by a visitor leaving your site almost before having taken a look; and the bounce caused by the user not interacting with your website.

The first type of bounce rate is the real one, where a user landed on a page only to leave it right away. While the reason for leaving might be that they found what they were looking for (like contact information or another type of key content that they can use someplace else), most often bounces are caused by slow loading pages, irrelevant content, low quality content, spam content, security warning, etc.

This type of bounce rate is difficult for you to measure, due to the fact that it is usually hidden in your analytics data, where real bounces are mixed with the other type that we’ll discuss next.

Even though it requires some customizations and data mining, it is recommended that you uncover the real bounce rate and react on it, if necessary. Depending on your site and its purpose, a real bounce rate somewhere between 12-30 % can be expected and is nothing to worry about – there’ll always be somebody that accidentally clicked on your links in search or reloaded the site by mistake on program start.

The second type of bounce rate is a grey zone: the person might have left the site but might just as well have spent an hour on your page. The uncertainty comes from the fact that Google Analytics doesn’t by default register time spent on a page, if no interaction found place. As a result, every site visit that only consists of one-page view and no events (scrolling doesn’t get registered by default) is labeled as a bounce.

It is important to understand that no interaction doesn’t mean no engagement. A user can spend half an hour reading your content, without it being registered by Google Analytics, as long as they don’t click on a button or move to another page on your site. Keep in mind that even custom events aren’t taken in consideration!

It is critical to investigate all of these grey zone bounces for two reasons: so that you get the real picture in regard to user engagement; and uncover the actual bounce rate.

Cracking the Code of Bounce Rate

Webmasters and SEOs monitor bounces as one of the key performance metrics on any site they manage. The reason is that they hope to get a reliable picture of how good the site is and whether it lives up to its purpose.

Therefore, those, who really know what they are doing, don’t rely on bounce rate as an isolated metric but rather use it in combination with other data, such as timers or scroll monitoring.

Timers allow you to record the duration of page views for all pages, also exit pages. The scroll depth metric enables you to investigate, how much of your content actually is seen by the site visitor. While neither of these metrics alone can help you understand user behavior, combining the data can provide you with a better understanding of your site’s performance.

Using a mix of data is necessary, when it comes to understanding page view duration and bounces, as Google Analytics only measures time on non-exit pages. The time spent on the last page of each session is never recorded due to limitations in the way Google Analytics records activity on sites. These limitations mean specifically that a page view with no following event is recorded as a page view with duration of 0 seconds. Google records the “end” of the page view only in connection to a consequent event, such as viewing another page, interacting with a video, or similar. As long as no end of the page view can be registered, it is recorded as zero. For one-page sessions, this means that the session duration of 0 seconds and the bounce-label they get isn’t reliable.

Yet another implication of this major flaw of Google Analytics is that two other metrics get unreliable: user engagement and time on page. It is therefore not possible to debug bounce rate by using Analytics’ own time on page and user engagement metrics.

Bounce Rate is(n’t) Important for SEO

Bounce rate is often referred to as a major SEO factor, which is one of the reasons why SEO specialists get really nervous, when it goes up. However, it might all be empty worries.

While there never is 100% evidence to support such statements, Google announced back in 2016 that bounce rate isn’t a reliable ranking factor, which is why its use as such is quite limited. Their argument was that this metric can easily be manipulated and, in the same time, it doesn’t take into account user intention, which makes it irrelevant, when evaluating the quality of a web page.

While this statement was immediately challenged by MOZ, who performed a test, showing that massive bounces do in some cases affect page rankings, Google insisted on not using bounce rate as a major ranking factor.

No matter where the truth lays, it’s important to remember that we are talking about the actual bounces, not the mix of actual and grey-zone bounces that you usually can track in Google Analytics.

This is why you only rarely should be worried about a high bounce rate – and only after investigating user engagement, as described below.

From Bounce Rate to Actionable Insights

Let’s discuss bounce rate from practical perspective. When looking at your Google Analytics account, you might see something like this:

Very high bounce rate, very low session duration, and excruciating engagement levels. While the situation might seem desperate, don’t give up just yet.

As discussed above, the high bounce rate isn’t a very useful metric, due to the fact that it covers two types of bounces – actual bounces and single-page sessions that might just as well have lasted more than a half an hour. Therefore, before starting the work on a strategy to lower the bounce rate, it’s worthwhile investigating, how high the actual bounce rate is and how engaged site visitors actually are.

The reason why you can’t trust your Analytics’ engagement report to understand session duration is the same as described before: Google Analytics doesn’t register correctly one-page sessions and last-page time on page. Therefore, EACH session duration as recorded by Google Analytics is shortened with the time, spent on the last page of the session. In addition, all one-page sessions are registered with 0 sec. duration.

To solve this quite critical issue, you can set timers triggering an event that is in turn recorded by Google Analytics. By setting timers you enable GA to record session duration, regardless of whether or not an interaction with the page found place.

To set timers that measure the same time intervals as Google Analytics’ Engagement report, you should add the following code to the header your site, just before the closing HEAD tag. In WordPress, you can simply add the code into the custom JS/CSS field provided by your theme. It adds the code to the header of your site.

You can choose to give the events another label (the time intervals), action (the numbers from 0 to 6), and category (TimeOnPage) that make more sense for you. You can also choose to set other intervals to measure. It is though recommended that you leave the timers and the labels as they are in order to get results compatible with the Engagement report in Google Analytics.

Save your settings after adding the code and wait until GA starts registering the new custom events. It can take up to a few hours.

Then, you can access the new data under Behavior->Events->Overview. The report will look something like this:

Take a look at the Event report. It can be a bit confusing to understand, if you don’t know how it works but it is in reality quite simple: the timers register elapsed time and record an event for each time the limit is reached. That is, for each session that lasted 12 min. for example, you get recorded the following events (one of each): 0-10 sec, 11-30 sec., 31-60 sec., 61-180 sec., 181-600 sec., 601-1800 sec. To figure out the exact number of sessions in each time interval, you must start from the longest sessions and deduct them from the lower levers, one level at a time.

In the example above, we have 35 sessions with session duration of 30 min.+ and 55 sessions with session duration of 10-30 min. This means that the actual count of 10-30 min. sessions is 55-35=20 sessions. To figure out, how many sessions lasted between 3 and 10 min., you should deduct 55 from 85, meaning that we had 30 sessions lasting between 3 and 10 min.

Now, take a look at the combined screenshot below. At the top, you can see Google Analytics’ own Engagement report. At the bottom, you can see, which custom events were recorded during the very same period.

Notice the difference in the number of high engagement events (601-1801+) between the custom event counter and the standard Engagement report. This difference is due to the fact that a large portion of sessions on this site are one-page events, where no further interaction is recorded.

Therefore, Google Analytics records very high bounce rate, which luckily seems to not be precise.

Now, when you have activated the timer-triggered events, you are ready to start tracking user behavior and get a new perspective on how to work with user engagement or content.

My Bounce Rate is Still High?

If the custom events also record very high bounce rate, that is 30-40+, you might want to go through this checklist, covering the most common reasons, why people bounce from a website:

1. Site speed: Slow loading pages is the nr. 1 reason, why site visitors leave a page right away – they simply don’t want to wait. In order to minimize the negative impact of low speed, you can optimize your images, optimize CSS and JavaScript, enable cache and gzip, and finally – check whether your hosting is fit to handle the traffic and the content of your site.
2. Meta data: Often, very high bounce rate is caused by misleading or confusing titles or description in search. Once the visitor opens the page and realizes that the content isn’t as expected, they leave. Some sites use such deceptive practices on purpose but, as long as you honestly try to provide and market quality, do check, whether your meta titles and descriptions match the pages content. It’s recommended that the meta title and on-page title are either identical or close to each other to not confuse site visitors.
3. Spam looking page: Another common reason for high bounce rate is a page that looks cheap or spammy. If you have many ads on your pages, bright colors or too many graphic elements, it might be a very good idea to consider redesign, where you focus on simplicity and delivering focused, quality content.
4. Wrong headers on your pages: Most site visitors skim the page, before they start reading. As long as your title (h1) or headers (h2-h6) don’t seem relevant to the searcher’s intent, they’ll leave. Make sure that you break your texts in paragraphs, where you focus on certain aspects of the topic. Make likewise sure that the headers are well formulated and help the site visitor find, what they are looking for. Remember also to write in logical order.
5. Poor content: Poor content is actually not one of the most common reasons for bounces, as readers can only evaluate the content as poor, after spending some time reading it. Therefore, here you can’t expect immediate bounce. You can, however, expect that the visitor will bounce, if they accidentally land on your site again.