Bank Customer Churn Prediction & Retention Strategy

Table of Contents

• Project Review
• Objective
• Key Business Questions
• Dataset Description
• Tools Used
• Data Cleaned
• Dashboard Themes
• Key Insights
• Recommendation
• How to use this Project
• Conclusion
• Author

Project Review

This project aims to analyze customer churn patterns in a bank using historical customer data and develop strategies for improving customer retention. It uses a mix of Excel, SQL, and Power BI to understand customer behavior and predict churn risk.

Objective

To identify key factors contributing to customer churn and recommend data-driven strategies using Power BI visualizations.

Key Business Questions

• What is the overall churn rate?
• What retention strategies can be proposed?
• Which segment of customers are most likely to leave?
• Are inactive or low-product customers more likely to churn?
• Does geography, credit score, or tenure impact churn?
• What customer segments are high value and high risk?

Dataset Descriptions

• Bank Churn Raw Dataset
• Bank Churn Cleaned Dataset
• Source : (Kaggle Dataset)

Tools

• Microsoft Excel (Data cleaning, pivot tables, visualizations)
• SQL (customer segmentation and patterns)
• Power BI (Dynamic, interactive dashboards and storytelling)
• Power Query (data modeling and transformation)

Data Cleaning & Preparation

Added Columns:

• Churn Flag ; based on Exited column
• Tenure Group ; New, Mid, Loyal
• Credit Score ; Low, Medium, High
• Customer Value ; based on number of products
• Risk Level ; based on credit score Missing Values:
• Checked for nulls – dataset was complete
• Ensured columns types were appropriate (e.g., integers for Credit Score)

Dashboard Themes

• Customer Demographic & Churn Overview
• Churn Behavior & Customer Segmentation
• Churn Risk Level + Customer Value
• Retention KPIs & Churn Drivers
• Recommendations

Key Insights

• Majority of churned customers had only 1 product.
• Active members had significantly higher churn.
• Customers with low credit score had increased churn risk.
• Churn was high in Germany.
• Customers with 3 and more products were more likely to churn

Recommendation

• Product Upsell Campaign : Target 1-product customers with loyalty incentives.
• Retention Program : Personalized offers for high-risk geographies.
• Re-Engagement : Reconnect with inactive customers through exclusive offers.
• Credit Score Support : Provide financial education for low-score customers.

How to use this project

SQL

• Open your preferred SQL tool (e.g., MySQL Workbench, DBeaver).
• Go to the ‘Bank_Churn’ folder.
• Use the ‘bank_churn_cleaned_data.cvs’.
• Run the queries from ‘churn_queries.sql’. Power BI
• Open the Power BI file, ‘Churn_dashboard.pbix’
• Use filters to explore churn patterns by segment
• Review the Excel files for cleaning steps and SQL scripts for analysis logic

Conclusions

This Project helped uncover patterns behind customer churn using real-world business analysis techniques. The dashboard can help managers make informed retention decisions.

Author

Elujulo Margaret Kehinde [elujulomargaret@gmail.com](mailto: elujulomargaret@gmail.com)