Insights - Behaviour Correlations
With the Behaviour Correlation Framework, get to know what your users need to do in their pre period so that they do the goal event you have set for them in the post period.
You can answer product questions like:
What should your users do on Day 0 such that they come back to the app between Day 1 and Day 6?
What should your users do between Day 0 to Day 6, so that they convert by Day 7?
What should users avoid doing on Day 0, so that they do not uninstall the app by Day 1?.
Generating a Correlation Report
- As shown in the above image choose the template based on the analysis you would like to do.
- Retention: Check behaviours correlating with respect to users opening the app.
- Uninstall: Understand what behaviours are correlating with Uninstalls.
- Conversions: Understand what behaviours lead to users doing your goal event.
Choose the date range you would like to analyse.
Set your Pre Period. Example: Set pre period to 0 if you want to understand Day 0 behaviours leading to goal event in post period.
Set your Post Period. Example: Set post period to 1 to 6 if you want your goal event to happen sometime between Day 1 to Day 6.
Choose the events you would like to analyse or allow Apxor to auto pick the top 20 events done by users.
Select event attributes to see behaviour correlation with respect to the event and its attributes.
Example: As shown in the image, you can choose multiple attributes (User, Session or Event) for a single event as well.
For the Retention and Uninstall templates, Post Event is selected by default. If you have chosen the conversion template, Select Post Event for your analysis.
Click Submit to get the behaviour correlation report. A popup with your report link will appear.
Once you click on the report link, you are redirected to the generated report.
You will get a list of events sorted in the descending order based on their impact on the KPI.
You can choose between two types of Behaviour Correlations : Correlations for users who do an event at least 'n' number of times or Correlations for users who do an event exactly 'n' number of times. This can be chosen in the Data Type option in the report after generation.
For example : In this report, you can see what the users have done on their first day (Pre period) at least 'n' times that has caused them to open the app (KPI - Retention) between Day 1 - Day 6 (Post period)
Consider the event Feedback_Generation. Out of 222 Users who have done this event at least 1 time, 49.10% have opened the app in the Post Period. Out of 131 users who have done the event at least 3 times, 51.15% of them have opened the app again the Post period. With this increase in trend, we can see that there’s a strong correlation between the KPI and the number of times the event is done.
There is also a graph, which shows the trend of the event done and its impact on the KPI. From the example graph above, we can infer that Open_Course_Category is 69.57% correlated with Week 1 Retention when the user does it at least 5 times on Day 0. This insight can be made actionable using Apxor’s Action Framework by passively nudging users to do this activity.