There are a number of Google Analytics metrics that continually cause confusion in the world of digital marketers. The five most prevalent ones are average time on site/page, users, direct traffic, the next and previous page path dimension and bounce rates. This blog post sheds some light on these and explains exactly how each one is calculated and how it should be interpreted.
1. Average time on page / Average time on site
Google Analytics calculates the time spent on a page by looking at the difference in time between one page-load and the next. The problem that arises as a result of this is that Google Analytics is not able to calculate the amount of time spent on the last page that was viewed in a session.
To counteract this, the average time on page metric is calculated by excluding all page views where the given page was the last page in a session. As a result, pages with high exit rates will have a time on page that is taken from a small sample of the total page views.
Average time on site on the other hand does not try to account for the inability to track the amount of time spent on the last page of a session. Average time on site is simply calculated as the total amount of time up until the last page of the session is loaded. This means that the time on site metric is nearly always lower than the true amount of time spent on a site.
Below is an example for two different visit scenarios:
For this visit the time spent on the various pages would look like this:
- Time on Page A = 5 min
- Time on Page B = 5 min
- Time on Page C = The Time on page C cannot be calculated, this is therefore excluded from the average time on page calculation. (If GA is unable to calculate an average time on Page C from the remaining page views, this is displayed as 0).
- Time on Site = 10 min
For this visit the time spent on the various pages would look like this:
- Time on Page A = 2 min
- Time on Page C = 2 min
- Time on Page B = The Time on page B cannot be calculated, this is therefore excluded from the average time on page calculation. (If GA is unable to calculate an average time on Page B from the remaining page views, this is displayed as 0).
- Time on Site = 4 min
While here the time spent would look like this:
- Time on Page A = The Time on page A cannot be calculated, this in therefore excluded from the average time on page calculation. (If GA is unable to calculate an average time on Page A from the remaining page views, this is displayed as 0).
- Time on Site = 0 min
The average time on the various pages and the average time on site for all the visits would thus be:
- Average Time on Page A = (5min + 2min)/2 = 00:03:30
- Average Time on Page B = 00:05:00
- Average Time on Page C = 00:02:00
- Average Time on Site = (10min + 4min + 0min)/3 = 00:04:40
It is especially important to keep this in mind when dealing with websites and pages that have high bounce and exit rates. This is when Google Analytics has to rely on small sample sizes, which may result in misleading or skewed information.
Google Analytics calculates users based on cookies that are stored within a web browser. As a result, Google Analytics sees each new device and each new browser on each individual device as a different user. It also means that every time someone clears the cookies from their browser, that browser will appear as a new user.
This means that user numbers are nearly always inflated and will seldom be equal to the actual number of people who have visited your site.
The below image shows how easily two individual people can be recorded as five separate users in Google Analytics.
Note: This assumes you have not implanted a user-ID solution. If you have implemented a user-ID solution, we suspect you already understand the inner workings of how GA tracks users.
3. Direct Traffic
When first learning about the various traffic sources, many people are taught that direct traffic is a result of someone entering the URL directly into the browser or clicking on a bookmark. While both of these do result in direct traffic, it is important to note that the direct bucket also catches a lot more traffic. The real definition of direct traffic is any traffic where the first page of the session does not contain a referrer within the headers of the HTTP request.
This can occur for a number of reasons. A referral from https to http will not include referral information. Certain apps do not pass on referral information. Untagged emails and links from an untagged promotional pdf, excel or word document will not contain referral information. You also get browser extensions that can modify or remove referrer information. And finally some websites mask or remove referral information when sending visitors to another site (for security purposes).
So while clicking a bookmark or entering the URL directly into the address bar will be counted as direct traffic, it is important to remember that there are a number of other traffic sources that fall under direct traffic. Most of the time, this is simply because GA doesn’t know where that traffic is coming from.
4. Next Page Path and Previous Page Path
The “next page path” and “previous page path” dimensions are used to see which pages are visited either before or after a specific page. These dimension are seldom used, as they need to be chosen as a secondary dimension or selected within a custom report. Unfortunately, when they are used, they are often used incorrectly.
This arises because of the terminology used to describe these terms, specifically the terms “next” and “previous”. The assumption is often made that this refers to the page before and after the “current” page. Something like this:
Unfortunately, this is not how these dimensions were designed to work. This is because the “next page path” and “previous page path” are designed only ever to be used directly alongside one another. This means that the “next” page path is the page directly after the “previous” page path, as depicted below (note the lack of a “current” page):
The best way of using the “next page path” and “previous page path” dimensions together is in a custom report. Here is an example, that shows the pages most commonly visited after the English version of the home page of amazeemetrics.com:
Setting up the Custom Report:
The custom report should include the “next page path” and “previous page path” dimensions, paired with the page view metric. To make this report more readable, it is also suggested to filter either “next” or “previous” pages to focus on only one page.
The custom report will then show the pages viewed either before or after a specific page. In the example below, we are looking at the pages viewed directly after the English version of the home page.
As you can see, this Custom Report does not make use of the “page” dimension, but only the “next page path” and “previous page path” as specified.
5. Bounce Rate
Google Analytics defines a bounce as a single-page session on your site. The website’s bounce rate is thus defined as the number of bounces divided by the total number of sessions.
The problem is, bounce rate is also calculated for each page. Here it is calculated as the number of bounces on that page divided by the number of times that page was the first page of a session (Landing Page).
This page level bounce rate is often confused with exit rate. The exit rate is an indicator of how many people leave the site from a page (similar but not the same as bounce rate). It is calculated as the number of times that a session ended on a page divided by the total number of times that page was viewed. It is important to note the difference between bounce rates and exit rates. To highlight these differences, we have included an example below:
Consider only these four visits to a website.
As a whole, there are only two sessions that count as bounces, Session 2 and Session 4. Thus the website bounce rate is 2/4 = 50%
- Page A has three page views (Visit 1, 2 and 3) and two exits (visit 2 and 3), this equals an exit rate of 2 / 3 = 66.66%.
- Page A was also the landing page twice (Visit 1 and 2) and has one bounce (Visit 2), this equals a bounce rate of 1 / 2 = 50%
- Page B also has two page views (Visit 2 and 3), it however has no exits, so an exit rate of 0%. It was also the landing page once (Session 3), but as this was not a bounce, it has a bounce rate of 0%.
- Page C has three page views (Visit 1, 3 and 4) and one exit (Visit 4), so an exit rate of 1 / 3 = 33.33%
- Page C was only the landing page once though (Visit 4), as this was also a bounce, Page C has a bounce rate of 100%
- Finally, Page D had only one page-view (Visit 1), where it was also the exit page, resulting in an exit rate of 100%. As Page D was never the landing page, it had no bounces and has a bounce rate of 0%.
As you can see, the website’s bounce rate, the bounce rate per page and the exit rate per page, can all tell very different stories.
If you have any questions regarding any of these metrics or their explanation, please send me an email.