The Data Analytics Blog

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

All Posts

How Machine Learning Can Improve Your Debt Collection Process

March 22, 2018 at 12:05 PM

Machine learning, for all its cool applications, is at its core the generation of predictive models using advanced algorithms that learn from data. If we have enough reliable and stable data to feed it, we can build models and make predictions on just about anything. If you are new to machine learning, read more on What is Machine Learning?

In a collections operation, having enough data shouldn’t be a problem. But knowing how to use the data at your disposal to increase your collection yields is the challenge. In this blog, we discuss how machine learning can help improve your debt collection process.

Top machine learning applications for collections & recoveries

Credit application scoring

Using machine learning during originations can reduce credit risk, and therefore future bad debt inflow, by identifying which customers are low risk. Internal data can also be used to improve the accuracy of the predictions further. It can also make predictions and recommendations on what credit limit and at which rate to offer which customers.

Identify propensity to pay or roll

Using machine learning, you can identify the accounts most likely to roll forward to a worse delinquency state and those most likely to pay at any given time, driving internal pre-delinquency and delinquency strategies. In addition, this will allow you to determine the best treatment for each account, achieve higher collection yields and decrease your operational costs by increasing your call centre agent efficiency and effectiveness.

Identify the best time to call customers

You can use ML to predict the best time and the best number on which to contact debtors. Then you can prioritise calls by time of day when a debtor is most likely to answer. This will not only reduce your telephone costs but will increase connection and collection rates, by improving your call centre agent effectiveness. The only requirement is that you have historical dialler information at a number level with previous outcomes that can be fed into the model to make predictions.

Yield forecasting

Standard forecast comparisons of month-on-month; month-on-year etc. data will not give you accurate forecast predictions to create intervention strategies to mitigate potential risk. For collection operations, a critical success factor is the ability to predict month-end outcomes accurately within the early stage of the billing cycle.

Using machine learning methods, it is possible to forecast yields accurately, and predict which remedial actions will have a short or long term impact if you are not tracking to your targets. Use these insights to inform your business and drive collections strategy.

What are some other benefits of machine learning?

Other than increasing your collections through the decisions it makes, other benefits of using machine learning in your business include:

  • Yielding insights humans don’t see in large sets of data
  • Possible higher accuracy of predictions over and above your current models
  • Removes human biases from decisions
  • Higher productivity and less time spent on laborious tasks
  • Higher conversion rates and yields
  • Lower employee turnover

What does it take to get started with machine learning?

If you are outsourcing the ML to experts that know what they are doing and your use case is known to the data scientists, not much! Stable and reliable data, of course, and enough of it for algorithms to pick up patterns and learn. Infrastructure and talent is a requirement if you’re setting your machine learning up from scratch, but there are many applications and software available with machine learning capabilities. This means that you can purchase tools and systems with built-in machine learning capabilities. You don’t have to worry about building a machine learning application from scratch, hiring data scientists in-house or keeping the system maintained and up-to-date.

Find out more about utilising machine learning in your business by downloading our Guide to Machine Learning for Business.

Using machine learning in business - download guide

Perry de Jager
Perry de Jager
Perry has been involved in Collections and Recoveries for the past 22 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

Solving the Credit Unaware Challenge with Psychometrics

At Principa, we engage with clients and organisations across the entire credit lifecycle and track the focus of the South African credit industry. For nearly ten years the focus has consistently been in the collection space, but recently (since early 2021) this has changed and a large number of our clients are focused on acquisitions and originations.

Predicting Customer Behaviour (PART 2)

In Part One of this two-part blog, we started providing a short overview of just some of the propensity models that Principa has developed. In this Part Two, we continue to look at different types of propensity models available across the customer engagement lifecycle that are used to predict behaviour and solve business problems. 

PART 2: How to Cure the Post Pandemic “Collections” Symptoms

In PART 1 of this two-part series, we explored how the current socio-economic climate resulting from the lingering financial hangover caused by the pandemic is negatively impacting the consumer's ability to settle a debt.