I have a technical question around best practice with building segments.
I have an offline trait from my data warehouse, for customer’s product holding and 3 other traits for lifecycle.
I want to segment the customers with product ‘A’, into 4 product lifecycles. Early life, in life, recontracting window and out of contract.
For early life, recontracting window and out of contract I have made positive inclusions, however for in life I have removed all the other options i.e. ‘AND NOT’ for early life, recontracting window and out of contract. (Negative inclusion)
My assumption is that Product A has 10m devices, if I remove 3 options, say totalling 4m, then I’ll be left with 6m. This does not seem to be the case as I’m seeing good volume in all other segments apart from in life.
The acid question is – do you always need to make positive inclusions rather than assumed inclusion by removing all other options? (i.e. does it handle remainders?)