
Chapter 5. Can you predict the customers who are likely to leave? 85
All the variables defined in 5.2, “The data to be used” on page 77, are supposed
to be merged into one table or view to be used as input of churn prediction
modeling. The following describes where and how you can get the right data set.
5.3.1 Churn indicator
The value of the churn indicator variable is decided, based on the definition of
the churn, and this variable is built, based on the defined data window,
forecasting window, and time lag. The churn indicator is on if customers who are
active in the data window have left the company in the forecasting window;
otherwise it ...