and so should you...
Answer: none. Using Visual Website Optimizer on my blog, I tested the layout of sidebar to see if it has any effect on bounce rate. Particularly, I tested whether the sidebar in right (default on many blogs) works better than the sidebar on left. The answer for my case turned out to be that it doesn’t really matter. (Though it may differ for you – you should A/B test it, it is really quite simple). Here are the versions I tested:
Control – Sidebar on the right (click to expand)
Variation – Sidebar on the left (click to expand)
Both versions had surprisingly similar engagement rate of around 22%. Engagement rate is inverse of bounce rate, so in a way this means , irrespective of sidebar positioning, bounce rate of my blog remains 78% (quite high, by the way). If you want to see actual results, here they are (click to expand):
Since the test involved changing the layout of blog across all posts, if you are curious that this would have involved digging into Wordpress or PHP code – the answer is no. In fact, the took just 5 minutes to implement. Using the advanced code mode of VWO, all I did was to define the Left Sidebar variation with this CSS: #sidebar { float: left; }. Yes, just that! VWO took care of the rest.
Related posts:
RSS feed for comments on this post. TrackBack URL
Subscribe blog via RSS
Actually, this isn’t very conclusive. All it says is that *your* website’s layout doesn’t matter left or right weighted.
Jakob Nielsen tested more than one website and came to a very different conclusion: Left weighted sites are preferable.
http://www.useit.com/alertbox/horizontal-attention.html
Comment by Dave Rodenbaugh — April 23, 2010 @ 8:01 pm
Yep, 100% agree. That is why I wrote in the post:
>(Though it may differ for you – you should A/B test it, it is really quite simple).
A/B testing results (even Jakob’s) should never be implemented without testing on your own site. Though A/B testing case studies give ideas and set a good benchmark using which you can compare your own results.
Comment by Paras Chopra — April 23, 2010 @ 8:07 pm
Informative, thanks!
Comment by Guy At HockeyBias dot com — April 23, 2010 @ 8:29 pm
Nice work on the A/B testing
Comment by steve — April 24, 2010 @ 12:51 am
> bounce rate of my blog remains 78% (quite high, by the way)
A bounce rate of 78% seems to be quite high for starters. But when I consider my own behaviour when reading blogs, it’s all about reading one post and leaving (or bouncing) when I’m done with it. So I guess that engagement or bounce rate is not the most suitable parameter to be tested for blogs – unless you test for some very intrusive alterations, like Smashing Magazine for example does by oftenly inserting a paragraph which is actually off-topic, beginning with “by the way:” and referring to other topics within their site.
Much more interesting for your blog might be, whether your post has been read up to the end and people didn’t bounce after reading the first paragraph. Hence one apparently more suitable parameter to be tested seems to be the time users spend on reading your post (which might be hard to measure).
Another interesting conversion goal might be to convince people to subscribe to your RSS-Feed or to follow your Twitter feed. I assume that tests for those parameters were much more telling.
Bounce rate oftentimes seems to be quite overrated.
Keep up your impressive work, Paras!
Comment by Rainer — May 1, 2010 @ 4:07 am
Hi Rainer,
I totally agree that bounce rate is overrated. Better metrics are conversion rate, which for a blog could be subscriptions to RSS and comments on a post. Your idea of reading the full post can also make a powerful conversion goal!
-Paras
Comment by Paras Chopra — May 1, 2010 @ 10:30 am
GetClicky.com recently changed their definition of a non-bounce to include people who stayed on the site for longer than 30 seconds. When you take that into account, the engagement rate goes up a lot.
It would interesting to be able to test against something like that. No, it’s not a feature request. Just food for thought.
Comment by Melvin Ram — May 2, 2010 @ 1:25 am
We’re an online publication that currently gets about 10 million page views/month and a key initiative for 2010 is finding ways to create more engagement with our users to increase our page view/visit number.
I’m looking at several different A/B testing platforms to help us with this initiative and it’s nice to see that you can measure engagement rate with your tool. So many A/B testing tools are focused on specific goals the user much reach, that’s great for an e-commerce site but as an online publication we don’t define our goals the same way and it’s refreshing to see you that you offer engagement as a goal.
Thanks for this post.
Comment by Nick — May 3, 2010 @ 8:58 pm
Thanks for commenting Nick! Yes, we realized that visitor engagement is a great metric to optimize for a lot of websites where no direct conversion goal stands out.
Comment by Paras Chopra — May 3, 2010 @ 9:02 pm