Our Solution
We developed state-of-the-art data mining models to improve efficiency of debt collection activities by deeply scrutinizing client behaviours, risk profiles, debt case characteristics and impacted payments. The data mining models disclose evidence for risk management, early prediction, intervention and treatment, towards delivering discriminative business rules, optimising current business lines, and discovering inside factors triggering client/debt risk. In order to reduce the cost of collection and the number of collectable debt cases, an effective and practical method is to optimise resources, for example, as follows.
- Finding characteristics of the clients who are likely to self-finalise cases even without any actions. It will be helpful to reduce cost without impacting the collection.
- Finding characteristics of debts/clients which are likely to respond to a given action. It will enable each debt case to be matched with the most effective and low-cost action. Assigning easily-collectable cases to auto action, instead of manual processing, will allow FTE to be allocated to most productive actions.
- Looking for the next best action for a case according to its current status and situation.
- Predicting potential high-risk debt cases, to make early-prevention on those cases.
- A huge number of debt cases tend to remain unprocessed. If each case was assigned with an accurate priority score, which is based on business requirement, it will guide manual actions to deal with high priority cases first, that would improve the efficiency of manual processing.