The Data Analytics Blog

Our news and views relating to Data Analytics, Big Data, Machine Learning, and the world of Credit.

All Posts

Implement Collection Strategy Best Practice With Debtor Segmentation

February 15, 2018 at 10:08 AM

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.

increase your collection and recovery yields

Perry de Jager
Perry de Jager
Perry has been involved in Collections and Recoveries for the past 12 years, spending time in different market segments ranging from law firms to investment companies. At Principa, Perry has worked on extended projects within both South Africa and the Middle East with some of the largest financial organisation, providing on-site consulting within the collections and recoveries space covering strategy, process, people and technology.

Latest Posts

10 Machine Learning Books To Read For Budding Data Scientists

Machine learning and artificial intelligence are exciting fields, and we've been writing about these topics for a couple of years now. While a lot of what we talk about on our blog is advanced implementations of machine learning and can be overwhelming to beginners, the core concepts of machine learning are actually pretty easy to grasp. There are many resources and cheat sheets available online, but we believe the old fashioned way of learning is sometimes the best: with a good book. Few resources can match the in-depth, comprehensive detail of a good book.

Our Top Artificial Intelligence, Machine Learning And Data Science Podcasts

Our data scientists are keen readers and avid podcast listeners. In this blog, we list a few of the podcasts that cover topics such as data science, machine learning and artificial intelligence, and that we’d recommend if you’re looking to start exploring the world of podcasts.

Where To Learn Data Science Skills In South Africa

In this blog, we’ve created a (non-exhaustive) list of courses you should consider if you want to learn essential data science skills in South Africa. These courses are mostly classroom training from South African institutions, but if you’re more interested in online learning, check out our blog Where To Learn Essential Data Science Skills Online.