An international bank develops financial products, recruiting new customers through marketing campaigns using standard CRM selection techniques. The bank wanted to know whether there was a faster, non-traditional way to build their marketing campaigns to achieve higher ROI on their marketing budget.
Using demographic data and customer purchase patterns provided by the bank, we trained an A.I. to establish a connection between certain customers and certain products, based on customers’ profiles.
We determined the probability that a customer would buy a new product under the campaign. After training several different algorithms – what we call discovery experiments – we defined a model to match customers to products. This allowed us to increase marketing campaign results, applying this AI to a new campaign with 156,000 customers.
This project followed our standard A.I. delivery approach, starting with model experimentation and data analysis during the discovery phase, and moving through POC until it was deployed in production. The results surpassed expectations.
The campaign’s success rose from 3% with the first 291,000 customers to 23% success with the following 156,000. Considering new sales, the bank’s ROI for this campaign grew by 736%.
Before taking working with us, the bank carried out an initial campaign using traditional methods with over 291,000 customers. They obtained 8,905 sales, or a 3% success rate, similar to what they had obtained in past campaigns. This success rate is what prompted their desire to integrate A.I. into their workflow.
When they launched a new campaign using the A.I., they returned 35,117 sales out of 156,000 customers, or 23%.