RFM Customer Segmentation Analysis

Project Overview

This project applies RFM (Recency, Frequency, Monetary) modeling on retail transactional data to segment customers based on their purchasing behavior. The goal was to uncover distinct customer groups and recommend targeted marketing strategies for retention and growth.

---

Key Insights

• Analyzed 5,860 customers from 800,000+ transactions
• Discovered 7 customer segments using RFM scoring
• Top 5 customers contributed $1.9M+ revenue
• Identified 615 high-value customers at risk of churn

---

Customer Segments Discovered

Segment	Customers	Priority	Description
At Risk Customers	1,470	🔴 High	Recent but infrequent buyers needing re-engagement
Potential Loyalists	1,430	🟡 Medium	Moderate recency and frequency with growth potential
Lost Customers	842	⚫ Low	Inactive customers requiring win-back campaigns
Champions	720	🟢 High	Best customers: recent, frequent, high spenders
Can't Lose Them	615	🔴 Critical	High-value customers at risk of churn
Loyal Customers	506	🟡 Medium	Frequent buyers but not necessarily recent
New Customers	123	🟡 Medium	Recently acquired customers

---

Visualizations Created

• Segment Distribution → Horizontal bar chart of customer groups
• Average RFM Scores → Compare Recency, Frequency, and Monetary metrics
• RFM Heatmap → Spending patterns across Recency/Frequency matrix
• Distribution Charts → Histograms for R, F, M values
• Top Customers → Highest revenue contributors

---

Marketing Strategies

• Champions → VIP treatment, exclusive offers, premium loyalty rewards
• At Risk Customers → Win-back campaigns with 15–20% discounts
• Can't Lose Them → Personal outreach + retention packages
• New Customers → Welcome series and onboarding journey

---

Tools & Dependencies

• Python 3.7+
• pandas | numpy
• matplotlib | seaborn

---

Business Impact

This segmentation provides actionable insights for:

• Retaining high-value customers
• Targeting at-risk customers before churn
• Optimizing marketing spend with precision
• Enhancing customer lifecycle management

---

Technical Implementation

RFM Calculation

rfm = df_clean.groupby('Customer ID').agg({ 'InvoiceDate': lambda x: (snapshot_date - x.max()).days, # Recency 'Invoice': 'nunique', # Frequency 'TotalSales': 'sum' # Monetary }).reset_index()

rfm.columns = ['CustomerID', 'Recency', 'Frequency', 'Monetary']

---

Project Structure

├── RFM_Analysis.py                         # Main Python script
├── README.md                               # Project documentation
├── RFM Analysis Visuals (Page 1).png       # Visualizations page 1
├── RFM Analysis Visuals (Page 2).png       # Visualizations page 2