Engagement and Event Tracking

Engagement with end users is a critical measure of an application's success and can generally be defined as an end user's focused interaction. In web analytics, proxies for measurement of engagement include average session length, average page views and return visits. In this post, we'll show how to use mobile app analytics to record new and more useful measures of engagement.

Most mobile app analytics services report average or median session length and pageviews. More sophisticated services include loyalty reports with a frequency distribution of session lengths, which is more informative than a simple average or median.

But apps provide a richer experience, and app publishers want more focused, customized measurements of engagement. For a news app, the number of articles read might be a better measure of engagement than session lengths. For an advertisement, it might be the amount of time spent on the page. For other apps, it might be the number of tabs or sections of the app that were accessed during a session. With Localytics, event tracking can be used to capture the ideal engagement metric for an app.


Simple engagement metrics: Totals and averages

When first instrumenting an app, many will tag each article read or tab opened as an event. This is how you might tag a website, and the result is a simple count of articles. Then, it's easy enough to compute a total number of articles read. The total can also be divided by the number of users for a simple average of articles read per user.

Engagement: Total and average articles read

In the above chart, a total of 407,000 articles were read over seven days by a total of 20,350 unique users. The overall average is 20 articles per user. This would suggest that our average user is pretty engaged. But it also smooths over large variations in engagement.


Improved engagement metrics: Articles by unique session

Fortunately, app analytics provide more options. One recommendation is to count the number of articles read during an app session and report the result in an event when the app closes. Now for each unique user and each unique session we have a count of the total articles read.

Engagement - articles read Localytics

In this chart we have the same 407,000 articles, same 20,350 users and same average of 20 articles per user. But by looking at our "article read" event frequency distribution, we can see real levels of engagement! The highest category is actually 21 to 25 articles per session, and 26 to 30 articles is third highest. Then, there are many users who were not engaged at all, with over 3,500 sessions with 5 articles or fewer read.

This is far more useful. If the app owner wants to up-sell users with a subscription model, he knows exactly which users are the most engaged and most likely to convert. Comparing engagement levels to past conversions will probably identify exactly the right threshold to target.


Other uses of event tracking

The same technique can be used to track other functionality and other forms of engagement with apps, subsections of apps and especially advertising. As an example, instead of measuring occurrences (ie, articles read), the exact amount of time spent reading an article or interacting with an ad can be easily tracked.

We'll continue to provide suggestions for tracking events, measuring engagement, creating conversion funnels and optimizing advertising in future posts.

For more information, see Localytics app analytics services and complete implementation and event tracking documentation.