From Fake News To Fake Metrics

Written by Joanne Kearney

Joanne has over 10 years’ experience working in digital analytics, executing and managing many large scale projects across the UK and Ireland. Joanne is also an experienced trainer, having developed many customised corporate training schemes and regularly speaks at digital and analytics events.
March 3, 2017

In keeping with the trend in fake news, we’ve decided to look at some fake metrics this month.

These are metrics that we would always warn people to carefully consider, given the manipulated and sometimes distorted story which they can tell. Ronald Coase summed it up well when he said “If you torture the data long enough, it will confess to anything”.

We’ve discussed our top 3 ‘fake’ metrics below:

1) Bounce Rate
How it is defined: A bounce is a single-page session on your site. In Google Analytics, a bounce is calculated specifically as a session that triggers only a single request to the Analytics server, such as when a user opens a single page on your site and then exits without triggering any other requests to the Analytics server during that session.

What to consider: The name suggests that this includes people coming to your site and ‘bouncing’ away again. The term ‘bouncing’ suggests that they don’t stay anymore than a few seconds and in some analytics packages this IS how this metrics is tracked, however, in Google Analytics, the user could have been on that ‘bounced’ page 20 minutes- reading the content and then left and it is still counted as a bounce! Similarly, the user could have visited your site and clicked off to another site, or another page on your site that doesn’t have the same Google Analytics code, in these cases that visit is also counted as bounce (unless for example, you are tracking it as an event and using the non-interaction parameter).

Solution: Use the non-interaction parameter for events- this parameter allows you to determine how you would like bounce rate defined for pages on your site that also include event tracking. We always suggest never focusing on site-wide bounce rates unless you are looking at the trend in comparison to other days/weeks/months. We suggest instead looking at it on an individual page-by-page basis under the All Page report- Bounce Rate is listed in this table but broken out by page:

Screen Shot 2017-03-03 at 10.07.06

2) Average Session Duration
How it is defined: Average session duration is: total duration of all sessions (in seconds) / number of sessions.

What to consider: Based on the name this can be another misleading metric. Google calculates a session duration by using the time difference between the first hit and the last hit recorded so:

For example:

  • Page 1: first hit: 10:00 AM
  • Page 2: first hit: 10:05 AM
  • Page 3: first hit: 10:10 AM
  • 10:10 minus 10:00 = a session duration of 10 minutes (600 seconds)

However, if you visited page 1 of the website at 10am, then page 2 at 10:05am, then page 3 at 10:10am and spent 25 minutes reading page 3 before you clicked exit the website, then your visit duration is STILL only recorded as 10 minutes when it really lasted 35 minutes- you arrived at 10am and left at 10:35am as there was no other hit outside the first hit on page 3. This also means that anyone who comes to your website and only views 1 page without interacting with the content (e.g. a bounce), will be counted as having a 0 session duration.

Solution: Ensure you are tracking all interactions with the page- especially those that don’t automatically send an event to GA e.g. PDF downloads, clicks off to social media clicks, clicks off to any third-party site, clicks on player interactions. 
A more severe solution: send an automatic hit to GA every couple of minutes spent on your website. Note though that this will increase the chance of sampling in your account.

3) % New Visits
How it is defined: A visitor who does not have a previous cookie stored on their device and browser (they are using to access the site) to that say that they have been to the website before.

What to consider: Google Analytics uses cookies to track our journeys across websites. When we visit a website a piece of Javacript is executed and cookies are dropped on our device, on the browser that we use to visit the site from.

Screen Shot 2017-03-03 at 10.47.14

However, we know that people will use different devices for their 2nd visit (e.g. desktop- 1st visit, mobile- 2nd visit) and in some cases, different browsers too (e.g. chrome- 1st visit, safari- 2nd visit) and also delete their cookies before returning to the site and in all these cases, a new set of cookies are dropped and another new visitor is recorded even though the visitor might actually be a returning one!

Solution: You can’t stop users from deleting their cookies or using different devices and browsers but the Google Analytics User ID allows you to connect data from visits across multiple devices: 

In short, these are just 3 metrics that we always warn people to be careful with when using. No analytics package is 100% accurate but some tweaks to your tracking can result in higher data accuracy.


Submit a Comment

Your email address will not be published.

Enjoyed reading this article? Find more like it below…