Machine Learning and Marketing Optimization for Commercial Banking

How to optimize marketing campaigns so that at any time, each customer is offered the services and products with the highest probability for conversion and highest profitability to the bank.

Banks increasingly operate multi-channel marketing campaigns on a daily-weekly-monthly basis. With analysis garnered from the statistical models, the campaign management system applies business rules and propensity models' scores to define the target population for each campaign.

This new Marketing Optimization approach turns targeted outbound campaigns from Product Centric to Profit Centric. This integrated machine learning methodology matches not only highest score product to each individual customer but also the impact of this specific sale on the bottom line. It takes into account numerous dynamic customer characteristics, product parameters, bank profitability and marketing strategy requirements. The outcome is a short list of 15-30 highest probable and profitable leads for each branch each day and a few hundred leads for the call center each day [rather than thousands] for the variety of product combinations. In this way, the call center can focus on scooping the crème only. The high score is achieved by automatically matching the conversion probability of each customer/product combination with the expected profitability of each sale.

Many institutions face a major obstacle to putting this modern marketing optimization philosophy into practice. To produce such highly accurate sales leads the bank needs to constantly design and run hundreds of sophisticated machine learning models. With predictive analytics or machine learning development tools available in the market, the bank has to invest a huge force of statisticians and analysts to manually develop the models and keep them up-to-date.

On the other hand, banks that have implemented Marketing Optimization through machine learning measure substantial increase in channel marketing ROI. Call center operations become much more effective as they work according to accurate models that easily keep track of the latest marketing dynamics.

With Sales BRAINs™, G-STAT’s automated machine learning application, the pain of manually designing, developing and running multiple models is gone. Hundreds of models are automatically developed by marketing analysts [rather than statisticians and data scientists] in hours [rather than weeks and months]. The models constantly follow changing market-product-customer conditions as well as profit impact of each sale, to produce the optimal ROI from each customer interaction. Campaign management becomes much more efficient with these highly targeted and hot leads. Sales BRAINs easily integrates with any existing DWH and CRM platform to start producing high score results in days.