Begin with the basic data, determine the dimensions of customers, products, and transaction standards, set up business-level tags and profiles, use Model Magic to build precise machine learning models for various business areas, deploy the AI results on the marketing platform, and then deliver the results to customers through various channels. This is the precision marketing process (as shown in the figure above).

As projects usually involve a wide range of areas, limited human resources cannot cover numerous data mining modeling demands and heavy modeling tasks. By setting up an automated modeling platform with Model Magic, the efficiency of modeling work is improved and models are launched more quickly. As a result, customers can meet diverse business needs and simultaneously capitalize on marketing opportunities.

Fund-product marketing

A new fund purchase forecast model is built to provide a marketing list for the initial fund launched monthly and to achieve accurate hit.

Forecast on capital demand

Customers with high capital demand sought by establishing a capital forecast model

Mining of potential customers of credit

The success rate of marketing response is improved through the use of a marketing-response model that helps the screening of credit white-listed customers.

Customer churn alert model

The customer churn probability upon the maturity of a given loan product is predicted through the use of a customer churn alert model. One can also formulate corresponding strategies for prevention.

Customer group operation

Analyze life-cycle operations, including customer life-cycle management, marketing management, and product service management for health insurance cardholders to improve the overall AUM.

Credit-card collection model

The overdue probability of customers migrating from M1 to M2 is scored by setting up the credit card collection risk model (Card C)