Does anyone have a working example of Customer Retained where the aggregation level changes based on the dimension viewed.
I have a case where I would like to compute Retained Customers based on Base Customers (n) months ago in the same time period.
Customer could be Retained on Global, but not on a Country or County level.
So the aggregation is different based on the X-Axis in the dimension.
I have a working application where the expression is super slow since I have to first compute an AGGR() of the base list of Customers (n) months, then get a COUNT(DISTINCT Customers)
This is too expensive on CPU for just 6M unique records.
Is there a better way to compute Retained information where we can push this into the load script or a more efficient way to get A * B where A is base list of customers, B is current list of customers.