• Eugene Tan

Delinquency

Delinquency refers to falling behind on required monthly payments. It is a business challenge for telcos, financial services and utilities.


Making accurate judgments on the likelihood of customer delinquency is essential for the purpose of responsible consumer practices and the health of a product portfolio.


Challenge


The challenge was to find innovative ways of improving the detection of a customer’s likelihood of delinquency and provide insights of typical customer profiles related to it. This is a space where organisations have significant maturity in data, process and policies, so it was an extra challenge to look at this problem with fresh eyes.


Solution


We spent time deeply understanding the problem and the desired objectives. We set up agile ways of working to create speed and transparency. August processed millions of transactions, and thousands of variables to automatically create and deploy a machine learning model in a matter of hours that predicts customer delinquency.


Outcome


  1. High engagement with business stakeholders was an important outcome. Equally, our delinquency model was 12x the speed compared to the existing processes.

  2. Provided new insights and ability to understand the customer profiles and key drivers for credit delinquency by:

  • Late fees

  • Account balance

  • Account activity

Value

  1. An improvement of 40% in detecting customer delinquency over existing solutions.

  2. The model captured 52% of the delinquent customer accounts in just 10% of the customer population.

  3. Provided the client an ability to offer payment plans to its customers to lower the rate of customer credit delinquency.

5 views0 comments

Recent Posts

See All