The time series data of the customers was aggregated to one single row per customer ID. Some of the business questions raised needed the exclusion of some variables that could possibly explain the client behavior through their life cycle. The variables that were identified for analysis were the following:
- Customer ID: This was used as an ID variable and not for analysis.
- Tenure: The number of years the customer has been with the brokerage firm. Right censoring was done in 31 instances where the customer is still active after being with the firm for 20 years.
- AUM: Assets under £0.5 million are classed as 1, between £0.5 and £1 million as 2, 2 to £5 million as 3, and above £5 million as 4.
- Risk appetite: As ...
