From Insights to Action: Using Mobile Benchmark Data to Enhance User Retention

Marketers: keeping tabs on how your industry’s marketing efforts are performing is critical, but not easy. That’s why we share biannual mobile benchmark data, so that you can keep the competition in mind.

It’s crucial, however, that you evaluate your app’s performance based on the right metrics. Some apps, for instance, value “time-in-app” data because of the addictive nature of their product (ie. many Social Media and Gaming apps). That being said, for apps that plan on sticking around for the long-run -- that is, to avoid “flash-in-the-pan,” fad status -- we’ve found that time-in-app isn’t necessarily a strong metric to measure success by.

For businesses that want to build a strong brand, we advise a different approach: Craft great messaging and follow stellar engagement practices, holding retention as your golden metric.

Operation: Keep Users Engaged

Since the cost of serving return customers is a great deal lower than serving new ones -- and it’s common business knowledge that 80% of your profits come from 20% of your customers -- it’s a no-brainer that retention should be considered a primary end goal.

We examined which benchmark metrics contributed to app abandonment in order to understand what makes a solid retention strategy. In the past, we found visible correlation between opt-in and retention, but this time we used the Pearson product-moment correlation coefficient to determine just how related the two variables are.

To start, our results show that push opt-in and the percent of users who abandon an app after one use are significantly negatively correlated. This means that the more users opt-in to receive pushes, the less likely they are to stop using it after only one session. Keep in mind that this data is for apps that have more than 10,000 monthly active users.


App abandonment and opt-in have a correlation coefficient of -0.29, so approximately 29% of the increases in app abandonment are related to a decline in the opt-in rate.

Marketers should keep this visual in mind when evaluating how important these metrics are and focus on improving what they can. For instance, if you want to increase your number of app users in the long-run, a proper push opt-in workflow will do the trick.

Check out some strategies to achieve this here. Notice that this post was written all the way back in 2015, yet there are still many app marketers out there who maintain a broken user experience, preferring to treat people as sheep to be herded rather than individuals to be heard. Nurturing your app users upfront is crucial to retaining them.

The relationship between time in app and retention

Examining the changes in benchmarks from 2017 to 2018, we found that retention is increasing over time, but app usage is decreasing. Among other factors, this change begs the question: on average, is app usage negatively correlated with retention?

Average session lengths and app launches measure usage, so we took a look at the relationship between these metrics and app abandonment. Since users who only complete one app session definitely launch apps less, there is a clear negative correlation between average app launches and app abandonment. We found this to be about -0.44. Using session lengths as the final determining factor, we found that there is in fact a positive correlation between the length of the average app session and the percent of users who abandon an app after one session.


This correlation coefficient came out to 0.27, so there is a weak-medium strength correlation between longer session lengths and users who abandon apps. This implies that the longer a user spends in an app to find what he/she is looking for, the more likely this user is to get fed up and abandon the app.

App marketers are getting better at leading their users to value quickly by using more effective messaging strategies. If you are seeing decreases in your usage metrics, don’t fret. People may simply be obtaining what they’re looking for faster. On the flipside, if your average session lengths are quite high and there isn’t a clear reason why (i.e., users are watching, reading, or listening), you may have bored users on your hands.

Users who opt-in to push launch apps more, but spend less time in them

It seems obvious that users who are willing to receive pushes are going to use apps at a greater rate. This concept partially plays out in the data, where we found that opt-in rates and average app launches are slightly correlated: 13% of movement in either metric is associated with the other, a coefficient of 0.13.

On the other hand, session lengths once again indicate a general disinterest in receiving contact from app marketers. The relationship between opt-in rates and session lengths is approximately -14%, a coefficient of -0.14. If marketers do not lead their users to value quickly, not only will they churn but they will also be less likely to opt-in.

Without this benchmark data, meaningful conclusions that inform effective strategies would be hard to find. A retailer looking to enlarge customer lifetime value or to avoid churn, for instance, would be unable to determine whether they are focusing too much on acquisition at the cost of opt-in, or to see where they are losing users (i.e., is the competition poaching them or is the market soft?). At the end of the day, marketers should always be aware of their oppositional advantage and be on the lookout for growth opportunities. It’s just smart business.


Localytics is the leading analytics and marketing platform for mobile and web apps across more than 1 billion devices and 28,000 apps. Localytics processes 50 billion data points monthly. For this study, we looked at data for apps with more than 10,000 monthly active users across the globe. App abandonment is measured as the percent of users who only use an app once, while opt-in is the percent of total users who chose to receive push notifications. To measure correlation, we used the Pearson coefficient method. The timeframe for this study was June 1, 2018 to August 31, 2018.