6. Prioritize Collections Agencies
When collections strategies fail, predictive collections analytics can still be used to maximize recovery rates. One effective use case for predictive analytics is predicting which collections agency is most likely to recover from a specific account. Models can rapidly analyze historical collections data to determine which agency is has had the most success with similar accounts in the past. Models can also take into account the bandwidth of an agency, number of cases recently sent to each agency, and the likely percentage of a debt each will recover.
Using this information, a collections team can prioritize the allocation of debts to specific agencies based on value-at-risk and other key KPIs in an effort to maximize the amount recovered.
7. Optimize Collections Models
Perhaps the most valuable tool predictive analytics can bring to your team is the ability to optimize collections models. From machine learning models to champion-challenger tests, analytics can help you ask and answer ‘what if’ questions to iterate on and improve your collections models. You can analyze model performance in real-time to get early warning notifications of dips in model performance. You can test new data sources to find ways to improve model accuracy. You can identify and eliminate the strategies that lead to increased recidivism rates. Model optimization helps you create incremental improvements and test new strategies across the entire collections ecosystem, which in turn improves efficiency, increases repayments, and lowers the level of loan loss reserves your organization needs.
5
Made with FlippingBook Digital Publishing Software