Behavior, Funnel and Cohort analysis can be compared or split on the selected event property. Using the split functionality, the analysis can be compared to the values present in the selected attributes.
For Example - From which city users are opening the app, or comparing different campaigns for the Funnel analysis, or Comparing retention driven by the different user programs.
Analysis can be performed on event attributes. In MoEngage analytics, events have event attributes and user attributes attached to one single event, for all events which are generated from SDK. The events which are S2S events (starting from the server) or events generated by MoEngage, do not have user context hence these events may not have the user attributes with them.
Perform Split Analysis
To perform the split analysis, add Split by attribute to the event in the Behavior Page and click Apply button.
Add Split Funnel attribute in the Funnel Page and click Apply button to get split analysis.
Add Split Cohort attribute in the Cohort Page to get split analysis.
User Attribute Freshness
All the user attributes which are present with events have the user attribute value of the time when the event was generated. Events do not have the latest user attribute value.
When an event is generated, at that point the user attributes are not updated or present, for those events the value of the user attribute will not be present.
Splitting based on the event attributes is always on historical data. As split analysis analyses, all the values of the selected attribute, a user who has done an event 'App/Site Open' three times in three different cities will be counted one time for each city when splitting on attribute City.
Split based on latest User Attributes
Split analysis on the latest user attribute is not available with a single split as given above. Split on the current values of the user attribute can be also be performed using Custom Segment Analysis.
Select 'Users that satisfy the filter' in the custom segment analysis section. Select filter basis as User Properties, attribute name, condition, and values. You can select attributes multiple times and get the analysis.
For example, if the split analysis for the latest user attributes are based on City attribute, select city in User Properties and then enter the first city that is NewYork get the analysis, enter second city that is London, and repeat the same with other cities.