Showing posts with label Web Analytics. Show all posts
Showing posts with label Web Analytics. Show all posts

2012-08-15

Understand the super low bounce rate of your site

I read an article on Sidey’s Blog (China Web Analytics) some days ago, about bounce rate, there are some parts to analysis super low bounce rate, i translate it here:

Super low bounce rate implies monitoring technology problem 
However, if the website (or some major landing page) bounce rate less than 5% or even less than 10%, are likely to contain a monitoring technology problem.There are at least two situation will cause the abnormal low bounce rate. 
  • Firstly, there both are added tracking code in frames page‘s parent pages and sub-pages(i.e., page through iframe or frame nested) ,in most of time , when the page is loaded in the browser, the tracking code run two times in a short period of time. For web site analysis tools, such as Google Analytics, will think that two different pages are opened (because the parent page and page URI is definitely different), so this will obviously is not considered to be a bounce, even if the user do not click on  any links of the page to visit other pages. I have seen an example, in this case, bounce rate is below 5%.
  • Secondly, there are two tracking code to be added in the same page or because of some reason, the same tracking code run two times. So when the page is loaded, as the page is refreshed once, Google Analytics do not calculate this as bounce. Bounce rate  at this circumstances even lower than bounce rate of the first case. But for Adobe SiteCatalyst, because of different definitions of bounce rate, this problem does not exist. 
As an analyst, we must never to hear the result, we'd better test by ourselves and speak according to the test data. So i test the second reason that will cause low bounce rate. This test started on 1st Aug. and ended on 13th. I added two GA tracking code on one of my site. Look at the Image below:(Click to see the larger image)



The average bounce rate of my site was changed from 62.67% to 8.33%, what an amazing data. Good performance bounce rate aha. Although my original bounce rate is higher and i really want my bounce rate to be lower, i can not accept this data, because it's wrong. We can get nothing good for our business via some wrong data. 

By the way, meanwhile, pages per visit of my site increased from 3.38 to 4.52, so if there are two codes on the same page, not only the bounce rate was influenced, but also pages per visit, may be there are some other data was also changed, i will continue to analysis. 

Back to the point of bounce rate, If you guy find your bounce rate is lower (super lower), you'd better to check if there are two part of GA data to avoid the wrong data.
But what is the benchmark of bounce rate, what kinds of bounce rate is good and what is not good? In last part of Sidey's post, he gives some advise to us:
In order to give the benchmark, first of all i will ensure that an important factor has no interference, that is new visit rate of site is certain, and about 30%. On this premise, we come to have a look at what does different website bounce rate mean:
  • Less than 10%: less than 10% of the bounce rate may imply a monitoring  technology problem, exception of some super websites, such as Google, Yahoo, or Facebook.
  • 10%-20%:  the probability of monitoring technology problem still exists, however, another possibility is, there may be a strong temptation content on your web site to lead various visitors to click on.
  • 20%-40%: brother, your site is doing a good job!
  • 40%-60%: not the best, but also not bad!
  • 60%-70%: Well, you need to have a look at the main landing page which is poor performance. Because the defining of a new user ratio is 30%, therefore more than 60% bounce rate can describe that the content of your website is existing problems.
  • 70%-90%: there is a serious problem, requiring to correct your website landing page immediately.
  • 90%: OK, need to update the site immediately, if not ok also, to see is there some problem of your web site server? 70% returning users and more than 90% bounce rate, until your web site has no management, or it is really the server problems, may be unable to load the page.
If you have some other opinions about bounce rate, welcome to comment at this post.

2012-08-06

Using Google Analytics to track multiple top-level domains

We separate our site into .com and .co.uk (maybe other domains) to improve our visitors' experience and also to help us better understand our visitors in different locations. Sometimes, we may also want to collect all the sites' data together to make the data more exactly. For example when visitors search a term from google to our .com site, then go to .co.uk site, we will hope google is our source. But if we do not collect all the data together, Google analytics will make .com site to be the source of .co.uk.

How should we do to track multiple top-level domains, in fact Google Analytics have customized the code for us, in your Ga center Admin->Tracking Code


Choose Multiple top-level domains, Iif you have used one top domain code, you will easy to understand this code. Copy the code and paste the code to all your domains' page, then test for about one week to make sure all the code is ok. You are successful to collect all your data together now, so easy.