Understanding Website Statistics
Checking website statistics is part of the testing process. There are two types of data that can be collected: web logs and page tags. Web logs are the trace left behind by every call to a website from a browser, whereas page tags are a special marker code placed on a page which monitors the website use from the browser end. There are differences between these two types of report and in practise both are useful to have. However interpretation of the results requires some understanding of the terms.
A widely used page tag type of system is Google Analytics which gives many statistics in depth. The interpretation of these figures can cause problems as they do not explicitly offer access to the underlying numbers, which are necessary to understand how a site or a page is really doing.
The underlying numbers in which we are interested are of four kinds:
- EE - A link from an External site through this page to another External site. This is a "Bounce" page as only one page is visited on the site.
- EI - A link from an External site through this page to another Internal page on the site. This means at least two pages on the site have been visited.
- IE - A link from an Internal page through this page to an External site. This also means at least two pages on the site have been visited.
- II - A link from an Internal page through this page to another Internal page. This means at least three pages on the site have been visited.
Knowing these figures would give us a clear idea of how the site is doing. In particular a useful figure would be the "II" number to show how many people have stayed on the site. However from the set of default statistics given this is not possible.
Pageviews, Bounce Rate and Exit Rate
Among the default statistics for All Visits to a website are: Pageviews, Bounce Rate and Exit Rate. These are made up from the underlying numbers like this:
- All Visit Pageviews = EE + EI + IE + II
- All Visit Bounce Rate = EE / (EE + EI)
- All Visit Exit Rate = (EE + IE) / (EE + EI + IE + II)
Manipulating these figures will not allow us to find out what each of the underlying numbers is. To do this we have to use an option on Google Analytics which will show "Non-bounce Visits".
Finding Non-bounce Visits
At the top of the Google Analytics page is a heading called "Advanced Segments". It is on the right hand side for the old version of Google Analytics and on the left hand side for the new version. If you click on the down arrow it offers a section called "Default Segments" with a tick in the "All Visits" box in the old version, but blank in the new version. If you are using the new version then tick all "All Visits".
If you scroll down to the bottom of "Default Segments" you will see an option for "Non-bounce Visits". Tick this box, then press the "Apply" button on the screen. The dashboard will then show two lines on the graph and two sets of numbers for each of the types of visit.
What Non-bounce Visits Means
When you compare the numbers for Non-bounce Visits with All Visits you will see the following:
- The Non-bounce Visits number of Pageviews is less than the All Visits number. This is what we would expect as bounces are excluded.
- The Bounce Rate for Non-bounce Visits is zero. Again this is what we expect as we have no bounces.
- The Exit Rate for Non-bounce visits is different from All Visits. This is a result of the bounces not being included.
This enables us to work out that for Non-bounce Visits the statistics are constructed like this:
- Non-bounce Visit Pageviews = EI + IE + II
- Non-bounce Visit Bounce Rate = 0
- Non-bounce Visit Exit Rate = IE / (EI + IE + II)
This breakdown allows us to calculate the underlying numbers and make more detailed judgements about the website. The method of calculation is shown on this page.