Useful for: Either showing how sticky a website is, or conversely how many pages a user must visit to reach their goal.
Be careful of: The fact that Google Analytics also counts repeated views of a single page by the same user (e.g. they hit refresh or trigger a reload)
Our preferred tool: Google Analytics
Pages per session is one of the most straightforward metrics in Google Analytics, and unlike it’s ‘average session duration’ brother doesn’t have many nuances to consider. In fact, its only minor nuance is that the metric will count any repeated views of a single page, so if a user clicks the refresh button or goes back to a page which they have previously visited, this will be included within the stats.
One thing you should not do is to use this metric as a one-size-fits-all litmus test for how well a site is performing. Let’s take two examples to demonstrate this point…
Website A is a well respected online magazine, with engaging content in the form of easy to digest articles. Website A decides to implement a new sidebar with links to their other articles, and use pages per session as a metric to gauge whether or not people are consuming more of their content. In this instance, pages per session could be a good metric in determining whether or not the new sidebar has increased engagement on the website.
Website B is an e-commerce site who decides to add two new pages to their checkout flow. In this instance their pages per session will likely increase, but they also find that the new pages adversely affect their conversion rates as users find the additional pages an irritation, and thus have a poorer experience on the website.
In summary, be careful when using this metric as an indicator for website performance, and if you do try to use it alongside other metrics (e.g. conversion rates, average session duration, etc) to get a more accurate picture on overall UX.
Useful for: Measuring the impact of meta description optimisations or schema implementation
Useful for: Providing an indication of how ‘sticky’ a website is, and whether any CRO work has yielded improvements. […]
Useful for: Not much by themselves!
Be careful of: The fact that there are both good bounces and bad bounces