Blu Attrition Model

Identifying at-risk customers before they leave a company is the fastest way to help stem possible attrition. Once you have the ability to identify these at-risk customers, you can help to counteract their behavior by reaching out to them to help save the relationship. Reflex Blu has developed an attrition model that employs a prediction methodology and scores your entire customer database based on a number of predictive factors that ultimately lead to attrition. Our model is based on a prediction methodology rather than a trigger based methodology, because balance decline triggers can, when used only by themselves, can be misleading as balances will often rise and fall in active deposit accounts creating false signals of attrition risk that are actually normal, self correcting fluctuations.

A “profile” of the database is developed using standard analytical techniques. This profile can now be applied to the current set of households to flag those most likely to close accounts or significantly drain balances in the coming months.

Key At-Risk Factors
The key to the model is that we are modeling the behaviors of your “best” customers. We have defined these “best” customer relationships as having the following key characteristics:

• Multiple Accounts
• Higher Balances and product usage
• No Fee Generation

Therefore, we typically build our models to predict any attrition loss to the bank using a definition inclusive of all the below:

• Termination of the entire relationship
• Account closure or decline in buying behavior

Our model takes into account multiple relationship behaviors and ranks and scores these behaviors to ultimately give your company a clear view of the true at-risk activity your customers maybe exhibiting. Blu identifies the critical at-risk households and also wraps a communication strategy around how to speak to the most at-risk customers and then evaluate how these customers continue to score for the duration of the model.

< Return to nsite