Analysing Analytics

well some of them anyway...

[giggle]

Hi

I've just been investigating the GA Bounce Rate (BR) figure.

Currently our (eCommerce) website is running at about 40% bounce rate and the average time spent on our site is about 3 minutes.

I know that this figure is difficult to analyse as various types of websites would return different BR's.

A couple of websites that I read recently have said that lowering your BR could result in Google driving more traffic to your website as because your BR is low then you're obviously providing the correct content for the search phrase used. Is this true? Can a more 'sticky' site result in more traffic?

Also, I've noticed that a higher percentage of 'direct' visitors to our site (i.e. no referral information) result in a bounce so I was thinking of sticking some code into our GA tracking that would only load the GA tracking code if there was referral information in a effort to reduce the bounce rate even further. Any thoughts on the wisdom (or not!) of this?

Thanks

Mick

[James_WV]

Hi Giggle,

"A couple of websites that I read recently have said that lowering your BR could result in Google driving more traffic to your website as because your BR is low then you're obviously providing the correct content for the search phrase used. Is this true? Can a more 'sticky' site result in more traffic?"

Yes, in my experience a more sticky site can equate to more traffic. I haven't found any official google information this, but if you reason it through a lower bounce rate means the site is more relevant - therefore more likely to get more traffic (especially true if you're using adwords as it can help improve your quality score in my experience)..

"Also, I've noticed that a higher percentage of 'direct' visitors to our site (i.e. no referral information) result in a bounce so I was thinking of sticking some code into our GA tracking that would only load the GA tracking code if there was referral information in a effort to reduce the bounce rate even further. Any thoughts on the wisdom (or not!) of this?"

No idea on this one!

[giggle]

Hi James

I found this link (not sure if you're allowed to post links like this, but I'll give it a try) where they tried to analyse the effect of the Bounce Rate:

www*1stsearchenginerankings.com/2007/06/08/google-bounce-factor-research-data-is-in/

Interesting reading...

[edited by: Receptional at 9:57 am (utc) on Dec. 9, 2008]
[edit reason] No, links not allowed, so de-linked as halfway house [/edit]

[James_WV]

Interesting- I had a vague idea on this from my own experience but have never had the opportunity to really do a controlled experiment on it. Looks like Bounce rate is a factor...

[cgrantski]

If that effect is happening, and I believe it is, I don't think Google Analytics has anything to do with it. Long ago Google (I think) said they were paying attention to what searchers did --- i.e. did they come right back to the search page after clicking through to a site. A site would be deemed more valuable if the searcher never came back to the search page. That's a bounce rate of sorts, but inferred from search engine results page behavior, NOT from the site's own stats.

[giggle]

But surely having the GA tracking code in there helps Google to see that the searcher actually uses the site and proves a certain amount of 'stickyness' related to the search term used proving that the content of your site is useful and appropriate.

What is slightly worrying is that if you have a high bounce rate then Google could use your own site to impose some sort of penalty.

Personally, our bounce rate is running at about 50%. I made a mistake at when I first created the GA account and put the tracking code in - I forgot to add it into our 'rates' page, resulting in a higher BR as customers would land on one page, go through to look at the rates and then move on. I fixed that and the daily BR is around 40% and slowly bringing down the overall average.

[cgrantski]

Yes, it WOULD help if Google wanted to do that and was willing to take that risk. That's different from the question of whether they actually do it. Everybody seems to want to believe they do although I should think that one or two Google employees or former employees, being pretty cool people, would have leaked that fact and also raised a hue and cry by now. I know a lot of Google people and former Google people and it's not even a rumor over there as far as I can tell.

Bounce rate for hits from Google Search results is likely to be extremely different from bounce rate for hits from huge untargeted shotgun-type email blasts and other things. Anybody knows that. You can be extremely relevant to a search and still do some crazy things with your other marketing. If I were Google, trying to do a good job of providing relevant search results (which is more or less my business), and if I were willing to use EITHER the site's complete bounce rate secretly obtained from Google Analytics being used as a site spy, OR the bounce rate I was able to calculate from just behavior from the search results page in the context of a specific search term, I would definitely choose the latter and not bother with the risks and effort of the former.

Just my opinion.

[Receptional Andy]

Click back to search result pages and on to other pages is far more reliable than directly using Google Analytics data as cgrantski suggests. They have this data for every site, and have had such data for years. GA is at best a partial sample, is not especially reliable except in aggregate. Added to Google's clear message that they are not using it in this way, I'm not sure why the Google Analytics thing is still doing the rounds.

[analyticism]

Bounce rate = single page visits/entry pages (in most tools)

It is a measure of stickiness meant to help you determine the efficacy of your landing pages. If users aren't getting past landing pages then they aren't engaging with your site content and not converting.