Machine learning-driven insights for customer conversion

: optimizing onboarding processes through predictive analytics at AM by froots

Abstract

This study, undertaken in collaboration with the asset management company froots, tackles the challenge of improving client conversion during digital onboarding in the financial services sector. By utilizing predictive analytics and machine learning, the research examines how demographic, behavioral, and engagement data can enhance user conversion rates while ensuring adherence to regulatory standards. Multiple machine learning models, including LightGBM, XGBoost, and Random Forest, were employed to predict user conversion likelihood with precision and reliability. Key factors such as referral sources, financial capacity, and engagement timing emerged as the most significant predictors of successful onboarding. To better interpret the findings, features were categorized into actionable, non-actionable, and regulatory groups, offering practical insights into their role in conversion outcomes. The study’s results informed the development of targeted strategies, including the creation of prioritized user lists and the use of dynamic incentives, both of which effectively reduced drop-offs and improved onboarding efficiency. These strategies demonstrate the value of predictive analytics in optimizing resource allocation and enhancing user experiences. By integrating machine learning into onboarding processes, this research bridges the gap between regulatory demands and user-focused strategies. The findings offer a scalable and efficient framework for improving customer acquisition in the financial services sector, highlighting the potential of data-driven approaches to address real-world challenges.

Date of Award	11 Feb 2025
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Nicolò Bertani (Supervisor)