Using data-driven segmentation to optimize bank customer engagement and retention strategies.

👤 Role: Researcher & Analysist 🏷️Project Type: Product Analysis 👤 Supervisor: Kseniia Baidina Team Members: 🇲🇲 Angela, 🇷🇺 Daniel, 🇹🇷 Deniz, 🇨🇿 Jin, 🇲🇲 Teddy ****📅 Year: 2023

Overview

Transaction Analysis is a comprehensive data analytics project focused on segmenting bank customers and optimizing engagement strategies through behavioral analysis. This project tackled the challenge of understanding diverse customer segments to develop targeted retention strategies and improve overall banking services.

Key Objectives:

• Identify distinct customer segments using RFM analysis
• Analyze transaction patterns to understand customer behavior
• Determine factors influencing customer churn
• Develop data-backed recommendations for targeted marketing strategies
• Compare purchasing behaviors across different geographical regions
• Test the effectiveness of cashback interventions on customer spending

Challenge

Banks often struggle to effectively segment their diverse customer base, resulting in generic marketing strategies and missed opportunities for targeted engagement. Understanding the factors that influence customer retention, transaction behavior, and churn is crucial for optimizing banking services and improving customer satisfaction.

Solution

We developed a multi-faceted analysis approach that combined RFM segmentation, clustering techniques, statistical hypothesis testing, and geographical comparisons to identify actionable insights. Our solution provided a comprehensive understanding of customer behavior patterns and enabled the development of targeted strategies for different customer segments.

Final Presentation - Angela-Daniel-Deniz-Jin-Teddy.pdf

Process & Methodology

Data Preparation & Cleaning

• Processed transaction data to remove anomalies and negative values
• Filtered data to ensure integrity
• Eliminated extreme outliers
• Established clean datasets for reliable analysis

Customer Segmentation

• Conducted RFM (Recency, Frequency, Monetary) analysis to identify valuable customer segments
• Applied K-Means clustering with Elbow Method to determine optimal cluster count
• Identified distinct customer groups based on transaction behavior
• Analyzed demographic distributions across clusters

Churn Analysis

• Investigated relationship between credit card balance and customer attrition
• Analyzed impact of customer service contacts on churn rates
• Examined correlations between product relationships and customer retention
• Conducted statistical hypothesis testing to validate findings

Transaction Pattern Analysis

• Analyzed distribution patterns of items purchased, cost per item, and total amount
• Examined correlations between transaction variables
• Compared purchasing behaviors across different countries
• Identified high-value geographical segments

A/B Testing

• Evaluated effectiveness of cashback interventions on customer spending
• Conducted Mann-Whitney U test and proportions z-test
• Analyzed statistical significance of the intervention results