Bank-customer-churn-analysis
Customer Churn Analysis Report
- Prepared by: Ifenna Daniel
- Client: Confidential Financial Institution
- Sector: Banking and Finance
- Date: June 2025
Introduction
Customer churn is a critical issue for financial institutions, impacting revenue and long-term business sustainability. My clients are experiencing a significant issue with customer retention.
Objectives
My role is to design and deploy a machine learning model that predicts which customers are likely to churn based on the given variables present in the dataset. Also to answer the following questions
- What are the key factors influencing customer churn?
- How does customer churn vary across countries?
- Which customer segments are most at risk of churning?
- What is the accuracy of the predictive model?
Technical Tools
- Regression Tree
- Decision Tree
- R Programming
- Tableau
Data Preprocessing and Manipulation
Performed exploratory data analysis. There were no missing values or duplicates. The column Customer_Id was dropped to optimize the dataset for analysis.
Also, discovered distribution of customer churn was analyzed, revealing the following:
- Churned Customers: 20.4%
- Non-Churned Customers: 79.6%
The cleaned data was then imported into R for modeling.
Model Approach
Decision Tree Vs. Regression Tree
- The Regression Tree outperformed the Decision Tree with an accuracy of 86.1%, offering more precise probability estimates for churn.
- The Decision Tree provided interpretable rules but had a slightly lower accuracy of 83.6%.
Thus, the Regression Tree model was selected for deeper insights.
Key Findings
1. What are the key factors influencing customer churn?
- Age: Customers over 50 are more likely to churn.
- Active Membership: Inactive customers show a higher churn probability.
- Number of Products: Customers with less than two products have higher churn risk.
- Estimated Salary: Customers with lower estimated salaries tend to churn more.
- Country: Customers from Germany exhibit a higher churn rate.
2. How does customer churn vary across countries?
- Germany: Highest churn rate, especially among older customers with limited product holdings.
- France & Spain: Lower churn rates, especially for customers with active memberships.
3. Which customer segments are most at risk of churning?
- Customers over 50 years old
- German customers with low salaries and inactive memberships
- Individuals with only 1 or 2 products
Conclusion & Recommendations
To reduce customer churn, the bank can implement the following strategies:
- Target Inactive Customers – Launch engagement campaigns for inactive members.
- Increase Product Adoption – Encourage single-product holders to adopt additional products.
- Monitor High-Risk Customers – Develop retention programs for customers over 50 years old.
- Customize Regional Strategies – Focus on Germany with personalized offers.
- Salary-Based Incentives – Provide special incentives for customers with lower estimated salaries.
Regression Tree Visualization
