Projects

Churn Prediction

Understanding and predicting churn to improve customer retention strategies

Funding: IFAB for PMI

Enabling technology: Big Data Analytics, Machine Learning

Sustainable Development Goals

Customer churn rate refers to the percentage of customers who stop using a product or service within a given period of time. The project aims to understand and predict churn in order to improve customer retention and loyalty strategies.

Objectives

Understanding and predicting customer churn in order to improve understanding of the patterns and factors that influence it.

Initial challenge

Explaining why Emil Banca’s customers stop using a product or service, and therefore understanding what to improve and how.

Solution

The project involves two approaches:

XGBoost with demographic data, using a machine learning algorithm trained on customers’ demographic data to identify patterns and factors that influence churn.
Survival Analysis with Lifelines, a time-based analysis that studies how features evolve over time up to the event in question (in this case, churn). This makes it possible to identify risk factors and quantify the “probability of churn” over time.

Benefits

The expected outcome is the ability to compare and integrate the results of the two approaches, in order to provide effective, well-informed recommendations on how Emil Banca can reduce customer churn.

IFAB’s role

In collaboration with EmilBanca, IFAB developed a predictive system for analyzing and forecasting customer churn, making its data science expertise available in support of the bank’s customer retention and loyalty strategies. The work began with building a solid, reliable data foundation, through data collection, organization, and quality assurance activities — an essential precondition for the robustness of predictive models.

Building on this foundation, IFAB carried out an in-depth analysis of customer behavior to identify recurring patterns and the factors influencing the propensity to churn, translating the data into actionable insight. The work concluded with the design and validation of predictive models for estimating individual-level churn probability, designed to guide targeted, effective retention initiatives.

The collaboration with EmilBanca allowed Fondazione IFAB to apply advanced predictive analytics methodologies to the banking and financial sector, consolidating expertise in understanding customer behavior and developing tools to support commercial and retention decisions.