Lenddo expert risk models are designed and hosted for customers who would like to improve their current scorecard performance. Data that we have experience incorporating into Scorecards includes demographic and application information, credit bureau reports, banking data including balances and transactions, telecommunications data including call logs, purchasing behavior, as well as many other types and sources of data. The Lenddo Data Science team can extract predictive features from those data sets as well as identify new patterns emerging from the interaction of features with sub-elements of the LenddoScore. Risk officers can use our Custom scorecards to create a fully automated underwriting process. The scorecard is hosted by Lenddo.
Leveraging the appropriate scorecard in effective decisioning strategies to support increased automation with little data is a key challenge. Our team are able to support customers’ policy rules, setting policy and crafting strategies to maximize the benefit of the newly developed models. A Consultant would recommend an appropriate initial strategy to capitalize on the first model and also seek to address credit capacity of applicants through appropriate calculations. Additionally, a strategy tracking methodology identifying key measures that can be used to track the effectiveness of the strategy and compare the performance of the strategy to any challenger strategies that may be developed in the future.
Following the implementation of the first model, it is necessary to periodically review the performance since minimal data is available for initial validation. A semi-annual review is recommended with this service initially involving consulting and agreeing on a comprehensive set of reports that will track the scorecard performance, followed by semi-annual review of the tracking reports produced by Client. This service is important for two reasons: first, to ensure that scorecard performance is closely monitored, problems are identified and resolved early, and scorecard is used in the best way; second, it offers a very good opportunity to receive knowledge transfer and learn the best practice in scorecard usage and tracking. We have included the cost of reviewing an agreed set of such reports to be produced semi-annually in an agreed format. In line with best practice, these standard tracking reports would be Front-end (Characteristic Analysis, Population Stability, Final Decision and Override) and Back-end (Delinquency Distribution, Score to Log Odds Chart, Trade-Off Curves and Scorecard performance Statistics)