Analytics Deep Dive Part Three: The four most confusing metrics in google analytics
Home

Analytics Deep Dive Part Three: The four most confusing metrics in google analytics

Author: Aditya Khanna | Categories: Digital Analytics, Digital Marketing, Google Analytics

analytics

I have been living, eating and swallowing analytics nowadays and trying to help our customers make right decisions based on their clickstream data. Because of this, I (hopefully) have an interesting perspective on both how to set up solid analytics processes, but also what marketers seem to get hung up on.

Most of us love Google analytics because it’s free, easy to implement and has a beautiful interface which any non-technical person can use.

Unfortunately, while those clever charts and dashboards make it look very easy to interpret Google’s metrics, the trust is a little less clean cut. Google Analytics can lead to some extremely dangerous miscalculations unless you understand how Google calculates their metrics.

Over and over again, I see the same interesting metrics in Google Analytics create a good deal confusion and have sometimes led to bad business decisions. Today, I will look at four metrics which are important for any analytics professional (or any other marketer) using Google Analytics to understand data before they make business decisions.

1 &2: Bounce Rate vs. Exit Rate

I have a lot of seen managers confuse bounce rate and exit rate. While these terms look very similar, they measure two very different page elements.

Here are their distinct definitions:

Bounce rate: In analytics (especially Google Analytics) we consider bounce as a visitor who visits a page and does not trigger any event or visits a second page.

Exit Rate: In analytics, we tabulate the Exit Rate, when a visitor exits from a page (closing the browser window or open another website in the same browser tab) is counted in calculation of the Exit Rate.

And here are the differences:
  1. For a Bounce to happen the first page of the session should be the last page of the session. And exit happens only on the last page of the session. So this means that if a page has a bounce rate of 100% it may or may not have an exit rate of 100% and vice-versa.
  2. Bounce does not happen in every session, but exit happens in every session.

Example:

Here are a few sample user sessions and on a conversion journey:

Session 1: Page A > Page B > Page C > Exit

Session 2: Page B > Page C > Page A > Exit

Session 3: Page C > Page B> Page A > Exit

Session 4: Page A > Exit

In an example above where the user had 4 sessions. The only session which had a bounce was session 4 where the user visited Page A and did not visit any other page.

So, your analytics program would make the following calculations (bear with me, I know this is math, but it is important to understand).

  • Bounce rate of Page A with 4 sessions, 2 entry and 1 Bounced session = 50%
  • Bounce rate of Page B with 3 sessions, 1 entry (Page B was not involved in 4th session) and 0 bounced = 0%
  • Bounce rate of Page C with 3 sessions, 1 entry (Page C was not involved in 4th session) and 0 bounced = 0%

In the same example, the user leaves the website with page A as the last page in 3 sessions.

  • Exit rate of Page A with 3 exits in 4 sessions =75%
  • Exit rate of Page B with 0 exits in 3 sessions = 0%
  • Exit rate of Page C with 1 exits in 3 sessions = 33.33%

See the difference?

3 & 4: Average Time on Page vs. Average Time on Site

Time on page and time on site are also metrics that can twist your mind at times. Although they look very simple by their names, the actual calculations have some interesting nuances which make them a little trickier than they appear at first glance. Here, I will walk through each metric and provide an illustration of what it might look like.

Average Time on Page: On the surface, the average time on page metric should seem pretty simple: This is the average length of visit for each page. Simple, right? Unfortunately, no.

That calculation includes a major flaw in this report that I also mentioned in my Analytics Deep Dive Part One: 4 Metrics Your Company is Probably Reading Wrong article.

Digital Analytics

Google calculates time on the page based on “jump visits.” This means that it calculates by taking a difference between your page visit and next page visit. So, if you spend 5 second or 20 minutes on a page and do not visit a second page, you the time on site is calculated as 0:00 in Google analytics.

Moreover, Google also discounts these bounced/exit visits when it calculates it’s average time on page. So every visit which had a bounce is not used in the calculation of average time on page.

So the actual Time on page metric that we see is exclusively based on non-bounced sessions for that page.

Interesting, Right? Here’s what the calculation looks like:

Average Time on Page = Total Time on Page A/ (Pageviews – Exits)

Example:

Session 1: Page A (30 seconds) > Page B(60 seconds) > Page C(50 Seconds) > Exit

Session 2: Page A (60 seconds) > Exit

Session 3: Page A (20 seconds) > Page B(10 seconds) >Exit

In this example, the time on page for Page A = (30+20)/2=25 Seconds Why? Because session 2 had an exit from page A so it was not considered in the calculation.

Average Time on Site:

Average time on site is another data point that appears simple. However, this measurement of the duration of a visitor’s overall session on your website is also calculated based on the summation of jump visits from various site pages.

While time on site considers the number of exits in the calculation, the time a visitor spent on the last page is NOT counted as a part of a visitor’s overall time on your site, because as there are no further jumps from there. So, if that page was an eBook they read for 20 minutes, for example, those minutes wouldn’t count.

Example:

Using the session example above, let’s explore this site visit based on the amount of time a visitor might have spent on each page.

Session 1: Page A (30 seconds) > Page B(60 seconds) > Page C(50 Seconds) > Exit

Session 2: Page A (60 seconds) > Exit

Session 3: Page A (20 seconds) > Page B(10 seconds) >Exit

In this example, the time on site for the 3 sessions above = (30+60+0+0+20+0)/3=36.66seconds

This is the basic reason why sometimes you may see average time on a page much more than time on site for the same page.

Check for yourself:

To completely understand this unique calculation dynamic, go to your own Google analytics and see these data dynamics for yourself. I would recommend looking at the following elements:

1: Look for your takeaways from your Exit rate and Bounce Rates:

  • Bounced sessions are the ones with only 1 page view.
  • When you have a page bounce rate of more than 50-60% it is important to drill down and check its entrances, exit rate and total page views Because bounce is only based on the entrances and if the page does not have enough entrances or its exit rate is much lower than the bounce rate then the bounce metric may have little or no significance in your reporting.

2: Review your Average time on page and Average time on site:

  • Remember: Average time on page is calculated by discounting the exit visits while average time on site considers sessions with exit visits (as 0 seconds visits).
  • Don’t be surprised when you see 2 different values for Average time on page and Average time on site for the same page.

What did you find there? Did anything surprise you? Please let us know in the comments. 

Measurement Model