Back

Bank Customer Churn Prediction

Introduction

One of the greatest problems across many industries is the loss of its customers.

A business that is set out for profit must actively analyze its customer churn data, to understand the dynamics of customer churn, identify underlying causes of churn, and optimize resources by focusing efforts on retaining high-value customers.

In this project, the customer attrition data for a Multinational Bank was collected from Kaggle and analyzed to estimate the churn rate and primary drivers of churn.

Explore Dashboard

Project Objectives

  • The first objective is to use Python to explore and analyze the bank customer churn data and extract insight into the underlying causes of churn
  • To build an interactive dashboard in Power BI that shows top level KPI’s and customers profile based on churn-rate analysis
  • To use Python and Scikit machine learning algorithm and build a model for predicting bank customer churn

Project Links

Github

Power BI

.

Tools used

  • Microsoft Power BI
  • Python
  • Scikit Libraries

Skills Utilized

  • Data Cleaning
  • Feature Engineering
  • Data Modeling
  • Data Analysis
  • Data Visualization
  • Dashboard & Storytelling
  • Model building

Insights

Dashboard 1

KPI Analysis

  • The bank had a total of 10,000 customers 2,038 of whom were churned at a rate of 20.4%
  • There were a total of 4,543 female customers out of which 25% were churned, and the male were 5,457 out of which 16% were churned
  • The average satisfaction score is 3.01 out of 5
  • The average age of the customers is 39 years

Dashboard 1

Churn Analysis

  • Customers between the ages of 41 and 70 had a significant attrition rate, peaking at 56% in the age category of 51 – 60
  • Based on tenure, Customers who had been with the banks for less than a year saw substantial atrrition
  • Customers with more than $200K in their accounts experienced a high attrition rate at roughly 56%
  • There was a 100% attrition for cusomers whose credit scores are below 400

Dashboard 2

Churn Analysis

  • Over 99% of disgruntled customers left the bank. This indicate a lack of effort to resolve their concerns
  • There was a 100% attrition rate for customers who purchased all the four products and 83% for those who purchase three products.
  • 22% of churned customers were unsatisfied with the bank services

Dashboard 3

Churn Analysis

  • Churn is high among dissatisfied customers who have with the bank for 7 to 10 years. This suggest that the bank is losing its loyal clientele as a result of dissatisfaction
  • The churn rate for high net-worth customers is significant across all tenures

Recommendation

  • A loyalty reward program should be implemented to retain long term customers and high-valued customers
  • To significantly decrease customer churn, the underlying causes of complaint should be identified and addressed.
  • Addressing customers complaint and reducing attrition can be achieved through training employees on customer management programs
  • Customers should be informed about the products, this may minimize the quantity of purchases thereby reducing churn