Predicting Large Deposits of Wealth Management Clients

www.boirefillergroup.com

In developing response models or defection models, creating the objective function can be as simple as determining whether or not a given person purchased an item (response) or whether or not a person has become inactive (defection). Once the objective function is determined, the modeling process attempts to determine the characteristics that best predict the behaviour in question. Straightforward with respect to response or defection, this process can become quite complicated if our objective function is not well defined. What follows provides description of a solution to a modeling challenge arising out of a Client need to identify a difficult-to-define behaviour.

A Boire Filler Group Client manages the investment portfolios of high-income wealth management clients. Initially, the Client desired to produce a solution that identified clients within its portfolio that were at risk of moving their entire investment portfolio to another manager. In this case, the objective function could be determined by identifying those clients that redeemed their entire portfolio. From there, it was simply a case of identifying the key triggers that predict complete redemption behaviour.

Here is the paper

Complex Objective functions: Predicting Large Deposits of Wealth Management Clients (PDF).