Bounce Rates – some facts

Recently I have had a fair amount of clients asking about the Bounce Rates they see in Google Analytics.
This usually occurs after changing links, creating new content, and/or  using customized landing pages.
All of these things can of course effect the Bonce Rate, so in an effort to clarify some of the misconceptions of the Importance and meaning of the  often misunderstood metric of Bounce Rates I offer the following blog post written by Mike Potts from a post written by Cristina Mataix.

This originally appeared on http://www.elisa-dbi.co.uk blog. in June

One of the metrics that generates a great deal of attention (and anxiety) among web analytics professionals is “Bounce Rate”. More simply know as “Bounce” it is often the metric against which everyone tries to fight when optimizing a site.

It seems that everyone wants to succeed in reducing this metric, trying to find out the optimal or ideal level. Of all the metrics on our web analytics scorecard it is often the one we have the least affection for.  It could almost be seen as the ugly duckling of metrics!

But is a high page bounce rate always bad? Absolutely NOT! And the proof provides some insight into how we need to take a segmented view of our website traffic…

Imagine I have bought a stay at a luxury hotel. I arrive at the airport and realize that I’ve forgotten to write down the directions to the hotel. Damn! So I open the browser on my iPhone, search for the hotel address in Google, and the page www.my-luxury-hotel.com/address is displayed first.

2 things here….the first is that, if you’re really doing your job as an online marketer for the hotel, you might want the address to be in the advert itself, meaning that I, as your customer, don’t even make it to the website.

The second thing is that success in this case, once I arrive on the website, is that you give me the information that I need and want, and then that I LEAVE. As your customer I am completely satisfied, I arrive at the hotel and I give you a bigger tip. I’ve bounced, and you’re happy.

So, in order to provide further explanation about truths and myths of bounce rates, I’ve listed some, errr, truths and myths about bounce rates! Enjoy (and please please comment).

Bounce Rate Truths

Visitors who see only one page count as a bounce – TRUE!

This metric counts all those visits which request only one page of a website, regardless of the page location or the time spent on the page. Other interactions such as scrolling, mouse movements also do not affect this. Therefore, if someone comes to visit the page, reads all the content, moves around the page, watches a video, uses an in-page form to contact you, (doesn’t puke) and then goes to another website, this visit will be counted as a bounce.

It’s a good to note from this that as web technology moves on and interactivity within a single page becomes easier to develop, the bounce rate metric becomes kind of defunct. Better to measure “interactions” than bounce rate don’t you think?

Bounces mean that the ‘Average Time on Site’ metric is definitely wrong! – TRUE!

A visit which bounces, in some analytics tools (such as Google Analytics), is recorded with a time on site of zero. This is because Time on Site is calculated by comparing the timestamps from ENTRY to the first page of the visit to the timestamp on ENTRY to the last page visited a. If a user does not move to a second page then the entry timestamp is the same for entry and exit pages, hence a time on site of zero seconds.

To look beyond this we always recommend to use the “Length of Visit” report (in the case of Google Analytics) that shows a graph of distribution of visits according to the time spent on the site (and therefore allows us to eliminate “noise” from the visits of zero seconds).

Bounce Rate is not the right metric for assessing all content – TRUE!

As discussed earlier, I often see people making the same mistake – the assumption that if page bounce rate is high then their content is not working. For content from news sites, blogs or other informative content such as contact pages we would expect a high bounce rate.

One suggestion to better understand how good or bad your content actually is, is to group all “high bounce rate” pages together in a profile so you can analyze this type of traffic separately. Consequently, by having a profile for all content where you would expect navigation to continue beyond the entry page you will be able to more effectively assess your site’s bounce rate.

Bounce Rate Myths

The % Exit metric has a direct relationship with Bounce Rate – False!

All visitors who access a website end up out of it at some point (% Exit), but not all visits will bounce. 
So, given any specific page on your website, bounce rate will always be lower than the % Exit, since all those visitors that bounce will be counted also as exits as well.

Having a low Bounce Rate is a sign of good website ‘health’ – False!

We talked about this at the start of the post and beyond, however there is a further point to be made here. As with all discrete metrics, looking at them is isolation is a very bad idea. Better looking at metrics across time to see if they are getting better or worse.

A few questions I’d recommend asking yourself.

1. Which pages, specifically, need to have low bounce rates (think products, home, transactional, results)?

2. Are the bounce rates on these pages going up (bad) or down (good) over time? Why?

3. When you segment bounce rate data based on the characteristics of your visitors (are these new visitors?, have they bought before?, did they click on a PPC ad?…..), what effect does this have? Do you have different bounces rates for different segments? Why?

When you’ve answered these questions perhaps then you’ll have a better than good view of your websites’ health!

You should focus on the Bounce Rate for the site as a whole – False!

But why not? Surely a lower number is better than a higher one, right?

Yes, probably, yet at a site wide level there is no way that you can make an analysis, or make decisions, that will change your site for the better, thereby reducing bounce rates further over time. In fact I would suggest that you don’t even look at site wide bounce rates as an important metric at all.

In summary, it doesn’t look like bounce rate is as important as it’s made out to be! It might help you do an initial identification of problem pages, but without a lot more information, and without using a segmented view of your traffic, there is no way that measuring bounce rate is going to help you improve your site for the better.

Perhaps we should think about bounce rate as a pre-analysis, pre-optimization metric, as a “table of contents” pointing to the pages that you really want to start looking at……