3 Site Averages That Can be Misleading

  • 1 + = 3

3 Site Averages That Can be Misleading

Do you use site averages to make important decisions about your website? Web analytics tools commonly report metrics as averages, like Avg. Session Duration, Pages per Session, and Sessions per User. If you take these metrics at face value without analyzing the underlying distribution, you may be misinterpreting the behaviour of your visitors.

Let’s say you open the Audience Overview report in Google Analytics and see an Avg. Session Duration of 2 minutes. This may lead you to think that half your visitors spend more than 2 minutes on your site, while the other half spend less than two minutes. If so, you would be wrong!

The distribution of session duration, pages/session, sessions/user, and other similar metrics are often heavily skewed. Most visitors are usually at the lower end of the scale, while a smaller segment of highly-active visitors tend to pull up the average.

Below we will outline examples explaining why averages alone can be misleading by using data from the Google Merchandise Store Demo Account.

  1. Average Session Duration

    In the screenshot below, taken from the Audience Overview report, the average session duration for the month of December 2017 is 2 minutes and 6 seconds.
    Average Session Duration highlightedIf you switch over to the Engagement report, under the Session Duration tab, you will see that the session duration for the majority of sessions was actually 0-10 seconds. There are still a significant number of users in the higher duration buckets, but it’s probably not what you were expecting, given the average alone. Although mathematically correct, the average here does not show you the full story.Engagement Report - Session Duration

  2. Pages per Session

    In the screenshot below, the Pages / Session metric has a value of 3.75, implying that the average user will see 3.75 pages per session.Pages / Session Metric Highlighted

    However, if you switch to the Page Depth tab in the Engagement report, you will once again see that the distribution is heavily skewed. The Page Depth for the greatest number of users is 1 page. If you would have only looked at the Pages per Session metric (which is also Average Page Depth) you may have thought the average user sees more pages on your website than they really do.Engagement Report - Page Depth

  3. Number of Sessions per User

    In December 2017, the Number of Sessions per User is 1.30 for the Google Merchandise Store.

    Number of Sessions per UserSwitching to the Frequency & Recency report, you will notice that most users have 1 session. In this case, the average number of Sessions per User is more representative, but it still hides the several thousand users who have 5+ sessions. And those may be your most valuable customers! Frequency and Recency Report

We’re not saying that averages are wrong. In fact, taking an average can be a useful summary of your data (assuming you calculate it correctly). However, an average is uninformative and potentially misleading without its underlying distribution. So, next time you see an average on a report or dashboard, make sure you look at the distribution of the metric before making any decisions about your site or visitors.

Have any questions or concerns? Feel free to reach out to us in the comments below!

By |2019-05-17T11:57:13-04:00February 8, 2018|1 Comment

One Comment

  1. garuav Tripati 2019-07-22 at 05:48 - Reply

    this is really helpful thanks for the points you have mentioned

Leave A Comment