Go to Top

Decreased Bounce Rate – A Metric For Success?

In client meetings every month I share paid search results and Google Analytics findings. For Google Analytics I typically look around for interesting stories/findings, take some screen shots and then recommend things to tweak or change based on what we find. Within the screen shots clients see some of the metrics that Analytics tracks and the 2 metrics I get a lot of questions about are “Bounce Rate” and “Time on Page.” Questions about bounce rate and time on page are good because these numbers can be misinterpreted easily. Jessica had a good post about Visitor Engagement and she pointed out that depending on the type of site you have, time on page should be viewed differently. The same is true for bounce rates, and to make matters even more complicated I would say that different pages within a site should be viewed differently in terms of bounce rate.

A common way I’ve seen these metrics misinterpreted is by not looking at sufficient amounts of data. For example if you see an average time on page at 10 minutes for a certain page you may be tempted to jump up and down thinking it’s a success, but what if you’ve only had 5 visitors to that page in that time frame, and let’s say one of those five left there computer on and walked away. This one person would skew the data from showing the accurate average time on page which could be less than 2 minutes. This is why I always suggest looking at the larger time frame to reduce the level of inaccuracy. This strategy can apply for most of Analytics metrics including Bounce Rates although with bounce rates I see a different misinterpretation.

The bounce rate is the average percentage of initial visitors to a site who “bounce” away to a different site, rather than continue on to other pages within the same site. Basically, a bounce is someone who doesn’t find what they wanted on your site and leaves right away. The bounce rate metric can be a good measurement if it is studied in the right way. From my experience you really have to look at Bounce Rates on a page by page basis because different pages on your site probably have different goals. This is very apparent when you think about landing pages. Landing pages are pages that are used for a paid search campaign. It is a page designed with a goal of having that visitor take some form of action like buy a product, fill out a form, download a whitepaper etc. Since landing pages are in some ways a gate that qualifies and collects information from customers, bounce rates should be viewed with different expectations than your homepage. The landing page is supposed to make non ideal visitors leave and ideal visitors to take action, so a high bounce rate could actually be a good thing. However, if you are examining a page like your homepage, which is supposed to speak to anyone and invite them to explore more, then a high bounce rate would be a very bad thing.

My overall suggestion is to read Analytics metrics with caution and don’t jump to quick conclusions. Analytics is a great tool that allows us to see a lot of insight into a website, but sometimes you have to dig deeper than the surface level to determine if the numbers you are reading are “good” or “bad.”

facebooktwittergoogle_plusredditpinterestlinkedinmailby feather

4 Responses to "Decreased Bounce Rate – A Metric For Success?"

  • awhinston
    January 29, 2009 - 5:06 pm Reply

    Clouding the picture even more is the way analytics packages record time on page/site. The common approach is to take the timestamp of two subsequent pages and subtract them. Seems straightforward enough, but think about a user who visits one page for 30 seconds, visits a second page for 10 minutes, then closes their browser. The application will report the ‘time on site’ for that visit as 30 seconds and report ‘time on page’ for the second page at zero (not 10 minutes). Anyone understand this differently? Anyone know of an analytics package that can report true ‘time on page’ for all pages?

  • jhullman
    January 30, 2009 - 10:19 am Reply

    You are right that this is how Time on Site would show up given the scenario you describe (30 sec on page 1, 0 on page 2)

    I don’t know of any package that will by default get around this, but something we’ve done to track engagement with Flash includes setting a timer variable that periodically logs time if the page is still open. This way you get some information on how long the user was actually on the page watching the video, even if they leave without further navigation.

    So you could probably do something similar with javascript variables – initialize a timer when the page loads, and have it periodically (every 30 seconds or so) record a reading, and pass that as a user defined variable to your web analytics package.

  • paisley
    January 30, 2009 - 10:38 am Reply

    Bounce rate works better when you use it on a per-visit basis correlated with keyword.

  • mwilliams
    January 30, 2009 - 11:30 am Reply

    I agree, looking at bounce rates are more meaningful when you look at individual keywords. If you lump them together and look at the bounce rate by page you won’t be able to determine which words caused your good or bad bounce rate.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>