|
Optimisation: Does Pareto Still Rule OK? A case study from the UK of how optimisation can lead the limit management in the credit cards business. |
|||||
Trying to manage credit decisions is like walking on a tightrope - it’s easy to slip off either side. Set your limits too high, then you are over exposed on bad debts; too low and customer income is constrained. But the development of low rate balance transfer offers means that trying to find the best limit strategy is increasingly more problematic. Low risk customers are likely to use the interest free period without any intention of retaining a balance beyond this interest free period. After 12 months on the books, it may be that only 40% of customers are active and, of those, only 50% may be retaining a revolving interest generating balance. Even with a much tighter bad rate than the overall book, this means that after you have removed those accounts that go bad, all the profit is being generated by less than 20% of the population - an example of the Pareto Principle at work However, advances in customer decisioning mean that initial credit limits can be set not only on the basis of credit risk but also based on expected customer profitability. If we can predict how different customers behave, we can set limits to maximise profitability. The results from a recent project with a UK credit card organisation have been outstanding: Profit increases by about 19% with little consequential impact on the business.
Experian-Scorex developed models to predict the propensity for the organisation’s customers to behave in different ways. Models predicted the propensity for customers to activate, free ride (i.e., use the interest free period and incur no interest charges), make minimum payments, periodically revolve, transact and go bad. Models also predicted the likely limit utilisation and the implied customer profitability.
In addition, models were also produced which predict longer-term customer behaviour in order to understand the real value of customers over a two-year outcome horizon. The models used a combination of application and bureau data. Credit bureau data identified trends in the use of credit with other lenders - this was critical in predicting how different customers subsequently behaved. Experian used data from its CVM models, which predict different customer behaviours.
Once the models were built, Strategy Optimization determined the best limit for each customer - within the organisation’s business and operational constraints - such as overall capital and loss rates. The strategy was then deployed through Strategy Management. Richard Turner - Optimisation Consultant, Experian-Scorex - from the presentation given at the Experian-Scorex SM User Forum 2006 |
|||||
|
Top |
|||||
 |
|||||
|
The Pareto Principle states that 80% of the consequences stem from 20% of the causes. In customer management, this generally is considered to imply that 80% of profit is produced from 20% of customers. This principle is alive and well in the management of credit card portfolios - and particularly in managing new credit card customers. Most profit is generated from a relatively small proportion of the customer population.
