Prescriptive Debit Order Analytics: Collections Optimisation And Reduction Of Regulatory Risk

Since it’s establishment in 2006 the Early Debit Order (EDO) system has become integral to South African credit providers’ collections strategies. The system currently offers Non-Authenticated (NAEDO) and Authenticated (AEDO) collections mechanisms. The two methods differ in the way in which the client gives the mandate to debit his or her account. NAEDO uses verbal or written agreement while AEDO requires that the client use his or her debit card and PIN to confirm the agreement. Importantly both of these methods offer the option of credit tracking over periods from 1 day to 32 days. The balance of the relevant account is tracked for a specified period and the debit occurs only if the balance exceeds the debit amount. This allows credit providers to collect more effectively.

Due to perceived issues with non-authenticated debit orders, there is a move by regulators towards a new authenticated collections mechanism. There are also moves to limit the period of tracking available to credit providers.

Prescriptive Debit Order Analytics can be used to optimally time debit order start dates and the associated tracking strategy to maximize the likelihood of collecting the payment and minimize the cost of doing so. Moreover by accurately predicting the success of debit orders on different days of the month the reliance on the tracking feature of the EDO system is minimized and consequently the exposure of the credit provider to regulatory changes in the collections space is also reduced.

Credit providers have a variety of data sources at their disposal to inform the optimal collections strategy, these include:

  • Internal collections history
  • Customer demographic information
  • Loan information
  • Large Employer information (including payroll dates)
  • External collections data from credit bureaus

These data sources can be combined and a wealth of features can be engineered to give a full picture of a clients internal and external payment behavior. Thereafter machine learning algorithms can be used to draw out potentially complex relationships between all features and the timing of clients historical payments. The trained algorithm can then be used to predict the likelihood of being able to debit a particular client on any particular day of the month.

This prediction can be used to calculate the expected collection amounts for all strategies across all clients in the credit provider’s book. A collection strategy can then be chosen that maximizes the expected collected amount. In addition, business rules can be integrated into these strategies, for example a credit provider may not wish to change the debit order strike date date for clients who have not yet defaulted.

Prescriptive analytics therefore offers two important opportunities to credit providers with respect to the use of EDOs. Firstly it allows for the use of optimal debit order strategies. This optimisation will lead to increased collections and reduced cost of collections. Secondly the use of prescriptive analytics can reduce credit provider’s reliance on the current NAEDO and tracking infrastructure and ensure collections strategies will continue to work if faced with a more limited set of options on EDO systems.

Are you interested in learning more about the application of prescriptive analytics to debit order collections? Schedule a demo today!