07-29-2015 03:52 PM
I am trying to predict attrition / customer churn in a non-contractual setting (direct marketing) using SAS Enterprise Miner - unfortunately most of the case studies online are for contractual settings (telcos, subscriptions etc.).
What should the data structure look like if the churn date does not exist?
The definition of retention that I am using is -
From a base of customers who bought during the 13-24 mo period, those who bought during the 0-12 mo is classified as retained and those who haven't are classified as 'churned'.
How should my data to look like if I want to do survival analysis in Enterprise Miner v 12.1?
Thank you so much in advance.