Bounce rate is defined by Google as “the percentage of single-page visits or visits in which the person left your site from the entrance (landing) page.”
What this means is it measures the number of visitors who arrived at a page and went no further into your site.
A bounce need not mean something bad. The visitor may have landed on a page that completely answers their question, may have clicked on an advertisment or perhaps just got up to go make a cup of coffee and forgot to come back.
It can though also reflect a poor reader experience — people arriving at your site and finding the content just doesn’t answer their question, or perhaps your website doesn’t look right in their phone or browser, or stories are overly paginated, or the colour scheme is awful or any number of other reasons that would generally connotate a “thumbs down” experience.
Some site styles, such as blogs, often have a higher bounce rate due in part to a combination of reader behaviour and poor site architecture, with others though, for example affiliate landing pages, the whole purpose may be to send the reader elsewhere.
So keep that in mind.
For a general content site like Travelfish.org, bounce rate matters. It matters because reading a single page on Travelfish is not going to answer every question you have about travel in Southeast Asia! There is always another question that can be answered or further information that can be imparted.
With that in mind I’m going to write a bit about three areas where we see a bounce rate that is significantly higher than the overall site bounce. I’m slowly making changes to address some of the worst offenders!
But first, how to measure it. We use Google Analytics for most of our statistics needs and, with a bit of refinement you can get some useful bounce data out of it.
You can also use this data to pick your battles. There is less value in spending 12 hours rewriting a page that gets 4 visitors a day and has a bounce of 80% than a page that gets 4,000 visitors a day and has the same bounce rate.
Note also that not all bounces are equal. A bounce with an average time on site of five seconds suggests “Recoil, evacuate!!!” while a bounce with an average time on site of five minutes suggests “Thanks, that really answered my question.”
So pump up GA and click on the “Pages” tab under “Content”. This will give you a chart showing you your ten most popular pages with time on site, bounce, exit etc.
Then click on the advanced filter, make the average time on page less than say 90 seconds and the bounce rate over say 60%. (Obviously select variables that fit with your readership).
This will give you a list of pages that people spend an average of under 90 seconds on, with over 60% of them leaving your site. Order the result by pageviews, increase the number of results to 50 and you’ve got your homework for the week!
Each of the pages may be bouncing for a different reason, but in Travelfish’s case, three of the main reasons (in no particular order) are:
1) Question is answered
2) Unanswered forum questions
3) Blog pages
Question is answered
Just one example area for this, we have hundreds of FAQ questions on the site. Stuff like “Are the ferries in Thailand safe?” or “Can I use ATMS in Thailand”, or “Can I use drugs in Cambodia?”
Each of these can be answered with about three words. (in the above cases, Mostly. Yes. Yes, but don’t.) What this means is if the person Googles “Are the ferries in Thailand safe?” and they get sent to our page, then they’re going to get the answer in about 6 seconds, and unless we do something about it, they’ll leave.
So what can you do?
We’re in the process of expanding pages like this to try to second guess other questions people may have. Again staying with the above, we could suggest “Extra reading” along the lines of:
Popular Thai ferry routes
Photos of Thai ferries
What ATM fees are there in Thailand?
What cards do ATMs accept?
Are drug overdoses common?
What penalties are there for drug use?
And so on.
Obviously this can be a bit hit and miss and not all topics can easily have more topics associated with them, but you’d be surprised just how many can.
Assumming the extra answers you’re providing are accurate, useful and value adding, then you should start to see bounce rate reducing.
Unanswered forum questions
We’ve got around 15,000 questions and 70,000 answers on the Travelfish forum. And while most questions now eventually pick up an answer or three, some don’t, and this was especially the case with the older questions, when we didn’t have quite as many active readers as we do now.
The “problem” is that many of these unanswered questions have been well indexed by Google, and visitors continue to Google “How do I get from A to B on a Tuesday by horsecart”, and get sent to us, where there is a question that exactly mimics the search query, but unfortunately has not been answered.
This is a poor reader experience. I find it annoying when I Google something only to be sent to a page that asks the same question but has no answer — and I find it doubly annoying to be subjecting others to that same problem.
There are two obvious approaches to this:
a) Mark as NoIndex/NoFollow any forum thread that doesn’t have at least X number of answers. This means Google will not index the question until someone answers it.
b) Go back through and answer the unanswered questions.
I’ve slowly been following the second option. Adding a post at the bottom saying something along the lines of “This is an old thread but if you end up here, here are some links for further reference. If they don’t help. please ask the question again on the main forum” And then I lock the thread, meaning no more comments can be added.
With the first method, the bounce should drop off severly as you’ll stop getting any search traffic to these pages, with the second you’re trying to steer readers to other parts of the site. If we don’t see a significant drop in bounce we’ll switch to the first method, but for now, answering and locking seems to be helping.
Trying to drag blog readers onto other entries or further into the Travelfish site, has been somewhat of a struggle, but we’ve been lucky in that with ten nearly identical looking blogs layout wise, we were able to test different features and plug-ins to see what worked and what didn’t.
Currently we’re doing the following:
Interlinking blog posts with in-context links
We do this both within each blog but also across to the greater Travelfish site.
We use YARRP and it is excellent. Easy to configure and doesn’t use images. We position it at the end of the story and in early testing, saw an 8-10% clickthrough rate.
We use search-meter with it positioned after the related posts and before the Facebook dialogue. Search Meter allows you to easily track what people are searching for which in turn can be very helpful for new story ideas and highlighting gaps in your coverage.
A simple Top 10 plug-in in the left column that displays, you guessed it, popular posts. If others liked them, perhaps the next reader will too!
The above are but a handful of means to approach bounce rate. The “correct” approach will depend on your style of site and what you’re trying to do with it.
Every reader counts and if it is raining outside and it’s tackling bounce rate or doing month-end accounts, I know which I’d rather be doing.
Got a suggestion? I’m all ears!
PS I know I said I was going to be writing about sprites, but I put my back out through the week and wasn’t up to figuring all that out. Will do soon though!