Goran Candrlic   Jan 18 2013   Flag

Not a really great article IMHO. First of all, bounce rate is sometimes interpreted poorly - this article explains it far better: http://www.analytics-ninja.com/blog/2012/06/google-analytics-bounce-rate-demystified.html

I've added the snippet that calculates at least 30 seconds of visit duration before user leaves the website and imagine, overall bounce rate on the site and blog is 20%, with search traffic 11%. Now, this is acquisition and aligning content with user intention (and we know that if user stays more than 8 seconds on the website, we've done a great job).

Other metrics can help you define the success of your site better than bounce rate.

... and the part of long vs short pages - well, now that we have Panda, Google wants longer, better articles. So it's out of the question to short it down just to have better "user experience" if the content is long - it must be better, more informative and so on.

Nick Eubanks   Jan 19 2013   Flag

Hey Goran - Not really sure on your correlation here... my article is a list of considerations, that I clearly spell out in the beginning of the article are not perfect for every situation, hence "bounce rates need to be looked at subjectively."

The article you link to is about exploring some common calculation and interpretation issues when trying to interpret bounce rate... and doesn't seem to have any thing to do with ways one might reduce bounce rate?

In terms of your snippet, I'm not really sure what you're talking about; how are you calculating at least 30 seconds - wouldn't you just be 'tracking' or 'measuring' this? 

If you re-read #20 on splitting up long posts, it's an approach to potentially optimize engagement, not to thin out the pages.. if you've ever read a multi-page article they're still the same article URL just with pagination in either an anchor (#page2) or a pagination directory (/page/2/) which makes it far from <em>out of the question</em> to cut down your individual page lengths... as Google is still going to attribute all of the content to the article URL.

Thanks for the comment.

Goran Candrlic   Jan 20 2013   Flag

Hey Nick, 

I was refering to the overall "bounce rate" misconceptions - especially within GA. If I type in "Skype Free Download", I download it directly and I visit only one page - GA will say I bounced (if there's no additional trigger on the link), but did I finish my funnel? Did I finished the goal in the query?

So in my sense, bounce rate is badly interpreted in the first place (one page visit) in GA. Improving the overall metric and understanding/correlation between the query and what the user did on the website is more important.

Example: I want to link to rel="canonical" blog post on Webmaster central blog. I know where it is. I type navigational query. I go there, maybe I do a screenshot. I copy/paste the link. Never visit a single page afterwards.

Is this bounce? How is my engagement?

So, in terms of reducing bounce rate, we should first "define" a bounce (something in email marketing is better - soft&hard bounce). So, maybe we should ask Avinash to define it better?

I wrote a blog post on engagement and how to do it for blog posts: http://www.webiny.com/blog/2012/06/26/how-to-reduce-bounce-rate-engage-your-audience-and-drive-conversions-with-your-blog/

Maybe than you will see my "line of thought". Otherwise - your tips are ok, we know most of them already, but I believe that metrics/analytics is not giving us right direction here.

Please comment. I would love to hear your opinion :)

Nick Eubanks   Jan 20 2013   Flag

I think we're in the same camp here; I agree that bounce rate is not a great KPI for most sites - that was the point I tried to make in the intro to the post (but apparently did not do so effectively :) ).

In terms of defining a bounce, I don't think we need a new definition but better education for people to understand that KPI's are not so straight forward as GA presents them.

In your blog post (which is well done btw) you talk about reducing bounce rate and increasing engagement, citing examples like 'creating the perfect layout' - would you say that your post is very different from mine? Aren't we talking about an over-lapping school of thought here?