Want to save this article for later?
On this blog, we regularly talk about what churn is. Basically: it’s an important metric to track if you intend to build a sustainable app. You need to see how many users uninstall and where they drop off in order to fix the leak in the bucket.
So if churn is so important to your growth, the big questions are: how do you go about conducting churn analysis? And is there a way to successfully predict user churn?
Churn is a metric that quantifies how many users have uninstalled your app over a specific time range. And churn analysis is the act of gleaning actionable insights from those numbers so you can work toward better user retention.
How do you get actionable insights, exactly?
Start with the most important one: churn rate. This is the percentage of users who stop using an app within a given time period. For an app to be healthy, the number of new users must be larger than the number of users who uninstall.
As of the first half of 2018, the average churn rate for app users across all industries is 80% within 90 days.01
But there are other numbers to track as well, including your monthly and annual churn rate or your probable monthly churn.
If you want to figure out where in your app users begin becoming inactive, or eventually uninstalling, you’ll need to use funnel reports.
Funnels allow you to identify how users navigate your app and where they drop off before reaching a conversion step. You’ll get answers to questions like:
The data from funnel reports help you plot the actions performed by the user so you can figure out how to move them away from churn and back on a path toward conversion.
Once the actual dropoff points are found, you can begin to pinpoint why your users uninstall. Or rather, come up with hypotheses that can be verified via A/B tests or other methods.
We conducted a survey of 2,000 app users and found that the top 3 reasons for uninstalling are: no longer using the app, a lack of storage space, and too much advertising in the app. Your churn analysis may show you similar results.
Just keep in mind that app uninstalls aren’t forever. Often, depending on your industry or app category, your users may only be taking a temporary break from your app.
The point of this all being: once you know why your users churn, you can go about optimizing the user experience.
So the thing about churn analysis is you can only get so far with calculating probable monthly churn.
There are existing predictive models that use statistics to guess at outcomes. With a lot of them however, there is a reliance on quantifying risk using static data about the customer as they are at this moment. Or even as they were in the past. Many traditional approaches based on older statistical and data-mining methods are rooted in older data and show which users performed a particular set of behaviors previously.
Predictability of future actions has only recently become available to marketers as big data and AI tools have improved.
The problem with churn is that it’s based on human behavior. And we all know that humans behave in random, unpredictable ways. One day they’re spending hours on your app, the next, they’ve uninstalled it and moved on.
While these older approaches still offer value and can successfully identify a low percentage of at-risk customers, they aren’t always accurate. Which means, you marketing dollars are being wasted chasing after the wrong segments.
But what if there was a way to let artificial intelligence and machine learning predict which users were going to churn before it even occurred? That would be a powerful tool to help determine where to allocate marketing resources so you can stem the churning tide.
CleverTap’s Intent Based Segmentation tells mobile marketers which users have a higher propensity to achieve any “goal” you predetermine. One of them could be churn, but it could just as well be predicting if a user on a streaming app will watch an action or science fiction movie.
Our data science engine creates an intent model for any goal by considering millions of different data points as inputs:
This allows mobile marketers to create more engaging long-term strategies that stop users from churning and improve overall business performance.
For more on intent based segmentation, read:
But more than simply identifying potential user behavior such as churn, our tools also provide mobile marketers with the capability to take immediate action on those insights.
For example, our RFM analysis tool automatically breaks down your entire user base into specific segments based on their scores for recency, frequency, and monetary value – i.e., how recently they last used the app, how often they use your app, and how much they’ve bought from you already.
On the left side of the grid you see how the CleverTap engine has identified your hibernating, at-risk, and can’t lose them user segments. And best of all, you can initiate engagement campaigns directly from this view.
For more on RFM analysis, read:
See how today’s top brands use CleverTap to drive long-term growth and retention