March 8th, 2010 by Lars Johansson
Are you one of the many people who like to see metrics presented in a layer on top of the links on your website? That functionality is often referred to as site overlay or visual overlay.
Do you know how to interpret a site overlay? If you think that it shows how many times visitors have clicked the links on the page you’re looking at, guess again. That is actually only the case if all the links on a page are leading to different pages. If you have two links leading to the same page, your site overlay may be fooling you.
If we take the Site Overlay function in Google Analytics as an example, it shows you how many times a linked page has been visited, not how many times someone has clicked the specific link leading to that page. If you have three different links leading to the same page, all three links will have the same value in Site Overlay. It looks only at which page you visited after you clicked, not the click itself.
See this example where I created a very basic page containing eight links leading to the same page. I clicked one of the links seven times. If you look at the example, you may think that Site Overlay is suggesting that I clicked each individual link seven times. That’s not the case.

Check what the site overlay function in your web analytics tool is actually telling you!
Tweet This Post
Posted in Google Analytics | No Comments »
March 8th, 2010 by Lars Johansson
My colleague Christoffer has played around with negative event values in Google Analytics and proven that it is possible to combine negative and positive values and report on them. It’s not recommended by Google, but it is possible to do.
Say you let visitors rate a page and that −10 equals “very unhappy,” 10 equals “very happy,” and 0 equals “no opinion”. If you have three votes and one is for “very happy,” one for “very unhappy,” and one for “no opinion,” the “very happy” and the “very unhappy” votes would cancel each other out. Now, if you’d have two votes for “very unhappy” and one for “very happy,” you’d get ((−10×2) + 10) ÷ 3 = −3.3. You would instantly be able to tell that there are more “unhappy” than “happy” votes.
Now, you should be looking at “Avg. Value” and not “Site Avg.” in the Event Tracking-Actions report.
In this example code, you should replace “My page” with the name of the page the visitor is rating:
pageTracker._trackEvent(‘Rating’, ‘My page’, ‘Very unhappy’, -10);
pageTracker._trackEvent(‘Rating’, ‘My page’, ‘No opinion’, 0);
pageTracker._trackEvent(‘Rating’, ‘My page’, ‘Very happy’, 10);
Tweet This Post
Posted in Google Analytics | No Comments »
February 18th, 2010 by Lars Johansson
I made four podcasts about testing back in 2007. Names of tools have changed, but the idea and principle behind testing is still relevant.
If you would like to give my podcasts a listen, try:
Podcast with Bonnie Thomas, Marketing Director of Memetrics
Interview with Matt Roche, CEO of Offermatica
Wine or Gasoline? A Podcast with Mario Fantoni of TaguchiNow
Podcast with Mark Wachen, CEO of Optimost
I am currently planning a new round table podcast about testing featuring thought leaders.
What would you like me to ask them? Comment!
Tweet This Post
Posted in Test and Target | No Comments »