Customer Churn Analysis | Data Analytics Portfolio

Project Overview

Customer churn, also known as customer attrition, refers to when customers stop doing business with a company or service. In this project, I conducted a comprehensive analysis of customer churn data to identify patterns, predict attrition, and develop effective retention strategies.

Understanding why customers leave is crucial for businesses across industries. This analysis focused on identifying key factors that contribute to churn and developing predictive models to identify at-risk customers before they leave, allowing for proactive retention efforts.

Key Objectives

Identify primary drivers of customer churn
Develop predictive models for early churn detection
Segment customers based on churn risk
Provide actionable recommendations for retention
Quantify the financial impact of reducing churn

Business Impact

It costs 5-25x more to acquire a new customer than to retain an existing one. Reducing churn by just 5% can increase profits by 25-95%.

5-25x

Customer acquisition vs. retention cost

Common Causes of Customer Churn

Pricing Issues

Customers may leave due to high prices, unexpected fees, or better pricing from competitors.

Poor Service

Inadequate customer service, long wait times, and unresolved issues drive customers to competitors.

Quality Concerns

Product quality issues, bugs, or service degradation can lead to customer dissatisfaction and churn.

Competitor Actions

Aggressive competitor campaigns, better features, or new market entrants can lure customers away.

Poor Onboarding

Customers who don't experience adequate onboarding may not understand the full value proposition.

Changing Needs

Customer requirements evolve over time, and services may no longer align with their needs.

Methodology

Data Collection & Preprocessing

Gathered customer demographic data, transaction history, service usage patterns, customer support interactions, and churn status. Handled missing values, outliers, and created relevant features for analysis.

Exploratory Data Analysis

Conducted comprehensive EDA to understand the distribution of features, identify correlations, and visualize patterns in customer behavior that may indicate churn risk.

Feature Engineering

Created derived features including customer lifetime value, recency-frequency-monetary (RFM) metrics, and engagement scores to enhance model performance.

Model Development & Evaluation

Implemented and compared multiple classification algorithms (Random Forest, Logistic Regression, Gradient Boosting) to predict churn probability. Used metrics such as AUC-ROC, precision-recall, and F1 score for evaluation.

Customer Segmentation

Used clustering techniques to segment customers based on churn risk and value, enabling targeted retention strategies for different customer groups.

Insights & Recommendations

Translated analytical findings into actionable business recommendations with potential ROI calculations for implementation.

Key Visualizations

Churn Rate by Customer Segment

Bar chart visualization showing churn rates across different customer segments

Analysis revealed that new customers (0-6 months) had a churn rate 3x higher than established customers (>2 years), highlighting the importance of strong onboarding.

Feature Importance in Churn Prediction

Horizontal bar chart showing top predictors of customer churn

The top three predictors of churn were customer service call frequency, contract type, and tenure, accounting for 62% of the model's predictive power.

Explore Other Projects

Airbnb Market Analysis

Comprehensive analysis of Cape Town's Airbnb market, identifying pricing trends and investment opportunities.

View Project

British Airways Service Quality

Analysis of British Airways service quality through customer reviews and operational metrics.

View Project

Kickstarter Growth Analysis

Analysis of successful crowdfunding campaigns, identifying key factors that contribute to project success.

View Project

View All Projects