Our Solution for Inaccurate Time on Page Metrics in Google Analytics

Most publishers track Average Time on Page as a way to understand how readers are engaging with online content. But this basic metric doesn’t help you understand user engagement in a meaningful way; simply measuring how long someone remains on a page doesn’t tell you whether they moved through the content, clicked links, or responded to a call to action. 

To overcome this lack of insight, publishers are adding a scroll-tracking function to gain more information into the types of content that most strongly capture readers’ attention. The challenge, though, is that scroll tracking can impact the Google Analytics (GA) Average Time on Page calculation, creating the impression that users are spending an inaccurately long time on a page. The solution? We have developed a custom calculated metric to help publishers get a clearer read on page engagement.

Average Time on Page in Google Analytics

Content marketers and publishers may be suspicious when they notice an unusually long Average Time on Page in their reports. Pages that have a high percentage of single page visits tend to be more frequently impacted by seemingly inaccurate Average Time on Page statistics.

While Google declines to make its analytics formulas public, we’ve surmised that GA’s Average Time on Page metric is calculated as follows:

Total Time on Page / (Pageviews – Exits) 

This makes sense for a basic implementation tracking mostly pageviews. The reason exits are excluded from the equation is that in a pageview-only implementation, there is no timestamp after the initial pageview, so therefore there is no Time on Page for that user. 

However, once exit links and scroll tracking are implemented, there are other events to provide data for GA to compute Time on Page. For example, a user might land on an article, then scroll through and continue reading for 3 minutes. The scroll events are tracked and GA factors that activity into the total time on page for the content. But if the user exits from the page, the basic metric still excludes the view from the equation. Therefore the time is included but not the view. 

How to Implement a More Accurate Engagement Metric

For pages with a high volume of exits — where users land directly on the page and then exit — the inflation to the metric can be dramatic. You might see that most pages have an Average Time on Page in the 2-minute range, for example, and certain outliers in the 12-plus-minute range. 

To compensate, we have devised a calculated metric that swaps in bounces for exits. The result is that only users who land on the page without any interaction event (and so no time on page) are excluded from the denominator in the calculation. Users who scroll or later click an exit link (thus generating an event and factoring into the total time) will be included in the total views to the page. The result is a much more accurate engagement metric for content reports.

Organizations that use GA can create up to five calculated metrics per view; GA360 users have up to 50. This is easy for GA administrators to do, no code required. Here is how to add our calculated metric to your GA reporting: 

Click on the Admin link, then navigate to the selected View. Under the View column, click Calculated Metrics > New Calculated Metric and set up the following calculation:

Total Time on Page / (Pageviews – Bounces) 

Generating More Reliable Time on Page Data

Without our revised calculation, GA will automatically compute an Average Time on Page that’s ridiculously long and inaccurate for some pages, particularly those with a high bounce rate. If you see wild swings on the average time in GA reporting, the first question to ask is, “Did this start to happen when we implemented scroll tracking?”

The problem isn’t just that you’re getting misleading information about a handful of pages — it’s that inaccuracy undermines confidence in your site data across the board. And distrust of data prevents organizations from using analytics as a smart basis for decision making. 

We’re finding that as publishers seek deeper understanding of how users engage with content, GA’s out-of-the-box Average Time on Page calculation is not especially revealing. Scroll tracking provides a more accurate view of how visitors are moving through a page to read the full text, provided you adjust the formula in GA accordingly. 

With our adjusted Average Time on Page metric, you’ll gain more meaningful, more reliable user engagement data. 

Did this interest you? Reach out to learn
how we can help you.

Contact Us

When you reach out to Empirical Path, you get in touch directly with our fully qualified and experienced consultants.   Whether you need an analytics audit, training, support, or a custom quote on Google Analytics 360 licensing, let’s talk today.