We have written a lot on ways to optimise your collection, but the simplest and most effective way is still to segment your debtors based on propensity to pay. Debtor segmentation is a best practice that every serious collections operation needs to utilise.
In this blog, we discuss what debtor segmentation is, what the effect of implementation will be and how to implement it as a solution in your collections environment.
What is debtor segmentation?
The best debt collection and recoveries operations deploy risk-based segmentation to identify a hybrid of high, medium and low propensity to roll and pay distressed borrowers to ensure that appropriate, tilted treatments are applied across the delinquent life cycle.
By analysing your debtors’ previous payment behaviour combined with credit bureau data, you will be able to identify which are high risk, medium risk or low risk. This segmentation into three basic categories gives you the opportunity to treat each debtor appropriately based on risk.
It is critically important to create and implement defined workout protocols that support rehabilitation and establishes an escalation trigger in early-stage collections. Ensure your collections and recovery strategy covers the ‘rehabilitation, repayment and recovery mantra’. Your process should be situational – based on the delinquency cycle and asset/product class. Rehabilitation endeavours should be explored and exhausted before the commencement of litigation. A fundamental requirement for consistent, accurate strategy execution is to deploy a tilted treatment framework across call scripts, short message services and any other engagement channel.
And this will lead to…?
Targeting your debtors with the right message and approach can pay off in various ways, such as better customer experience, increased payments collected, enhanced brand affinity, amplified brand loyalty and improved profits.
Sounds good! But how do I segment my debtors?
Several techniques can be used for debtor segmentation, from the simple k-means algorithm to the advanced model-based clustering to the continuously improving machine learning.
By using a method that includes machine learning, your models will continually update with new information provided and will teach itself if a debtor was identified as low-risk, but in fact, rolled on payment and how to determine such cases in future.
Whichever approach you take; the most important thing would be to ensure you have enough data to segment these debtors. Combining your historical payment and operational data, with credit bureau data, will enable you to get a more accurate picture of a debtor’s propensity to pay or roll.
Read more on our Genius Machine Learning as a Service offering if you are interested in implementing machine learning without the infrastructure cost to segment your debtors, or download our guide to optimising your collections and recoveries operation yield.