Business Challenge
The credit card product line of bank has been underperforming for about a year. To boost sales, the bank has been trying to cross-sell this product to new customers in deposit & lending product lines. However, often the customers are declining the offer saying that the suggested credit card types do not meet their needs.​​
Merkle Solution
- Built a Customer Data Platform (CDP) to create unified profiles of existing customers.​
- Profiles include credit card data and relationships with deposit and/or lending products from
bank.​ - Trained a machine learning regression model using these profiles, with credit card type as the
target output.​ - Used the model to recommend the right credit card type for new customers starting a deposit,
lending, or both relationships.
Technology
- XGBoost
- Einstein Studio
- Salesforce Data Cloud
- AWS RedShift
