Setting interest rates is a complicated problem. If you set rates too low, you are cutting your profit. If you set them too high, your clients won’t accept them or will soon move on. The competition is very stiff, especially online. Most banks use risk-based pricing, assigning each class of clients interest rates according to the probability of default. This method can be improved by data-science algorithms. By giving clients who are more price sensitive a little bonus at the expense of clients who are less price sensitive while maintaining the average interest rate for all risk classes, we can improve acceptance, reduce turnover, and boost profit while still complying with strict banking regulations. Find more in the video or read about the use case at bigdataforbanking.com.