Apr
25
The Search for Better Web Analytics
Filed Under Best Practices, Measurement and Tracking, News, Web Analytics
Understanding your blog or website traffic requires more than merely counting your visitors.
You ought to know too where your visitors are coming from and the suitability of each source of traffic, as well as how effectively your website is performing when visitors arrive.
Average time on site and bounce rate are two of the statistics that help you gauge the relevance of your website to your site’s visitors.
Google Analytics Bias
Google Analytics is a popular free tool for monitoring your website traffic that has many great features including advanced visitor segmentation, a favorite of mine.
Sadly, Google Analytics reports bounce rate and time on site statistics that are biased. The Google bounce rate is too high, while the time on site is too low. This problem tends to be especially acute in the case of blogs.
I defined Google bounce rate and discussed the bias in Google Bounce Rate Misleads Bloggers (Oct. ‘08). Please read that introductory article if this subject matter is unfamiliar to you.
Cause of Google Analytics Bias
Google Analytics assumes that a visitor who views only a single page on a site is dissatisfied and leaves immediately, which is often not the case. A visitor can spend time on a single page and leave contented, especially if that page contains a blog post.
However, Google’s assumption was much simpler and more cost effective for Google to implement at the time they made it than an alternative would’ve been. Yet, a change in that one assumption would permit Google Analytics to provide more useful estimates of both bounce rate and time on site.
Better Web Analytics
Technology has advanced since Google formulated its bounce rate. Web analytics service providers now have access to faster servers and greater bandwidth.
In search of better web analytics, I have tried other service providers.
My favorite is Clicky, an innovative web analytics service that exploits newer technology.
Clicky devised and implemented an approach to estimating bounce rate and time on site that is quite elegant and superior to that which Google implemented in Google Analytics.
The Clicky script installed on each website communicates at pre-assigned intervals with the server. This interaction enables Clicky to estimate time on site, even if the visitor views just a single page. Clicky then makes an arbitrary yet fair assumption that any visitor who was on the website for at least 30 seconds found the site to be relevant.
The difference in methods is striking: My overall Google bounce rate last week was 79% vs. my overall Clicky bounce rate of 26%. Google average time on site was 1:55 vs. Clicky 3:41.
Perfect timing.
No comments:
Post a Comment