Mastering Customer Segmentation: The Critical Business Questions RFM Analysis Can Solve - Implementation

Kishore Babu Valluri

7/5/20253 min read

To implement RFM Analysis in Python, you’ll need transaction data (with columns such as Customer ID, Purchase Date, and Amount), which can then be used to calculate Recency, Frequency, and Monetary values for each customer.

Let's assume that the transaction data is in a pandas DataFrame. Below is the Python code to implement the RFM analysis and answer the various business questions:

1. Required Libraries:

import pandas as pd

import numpy as np

from datetime import datetime

import matplotlib.pyplot as plt

import seaborn as sns

2. Calculate Recency, Frequency, and Monetary:

We calculate Recency, Frequency, and Monetary values for each customer.

# Get today's date (most recent transaction date in the data)

today_date = df['TransactionDate'].max() + pd.Timedelta(days=1)

# Recency: Days since last purchase for each customer

recency_df = df.groupby('CustomerID').agg({

'TransactionDate': lambda x: (today_date - x.max()).days

}).rename(columns={'TransactionDate': 'Recency'})

# Frequency: Count of transactions for each customer

frequency_df = df.groupby('CustomerID').agg({

'TransactionDate': 'count'

}).rename(columns={'TransactionDate': 'Frequency'})

# Monetary: Total amount spent by each customer

monetary_df = df.groupby('CustomerID').agg({

'Amount': 'sum'

}).rename(columns={'Amount': 'Monetary'})

# Combine Recency, Frequency, and Monetary into a single DataFrame

rfm_df = recency_df.join(frequency_df).join(monetary_df)

# Display the RFM table

print(rfm_df)

3. Business Questions Implementation:

(1) Which customers are most likely to buy again?

Customers who are recent and frequent buyers are more likely to buy again. We can identify them by looking at the Recency and Frequency values.

# High Frequency & Low Recency -> Most likely to buy again

likely_buy_again = rfm_df[(rfm_df['Frequency'] > 2) & (rfm_df['Recency'] < 30)]

print("Customers likely to buy again:")

print(likely_buy_again)

(2) Who needs to be retargeted or re-engaged?

Customers who have low frequency and high recency are at risk. These are customers who haven't purchased recently.

# High Recency & Low Frequency -> Customers to retarget

at_risk_customers = rfm_df[(rfm_df['Frequency'] <= 1) & (rfm_df['Recency'] > 30)]

print("At risk customers (need retargeting):")

print(at_risk_customers)

(3) Which customers are worth investing more in?

Customers who spend a lot of money and have a high frequency of purchases are worth investing in.

# High Frequency & High Monetary -> Worth investing more in

high_value_customers = rfm_df[(rfm_df['Frequency'] > 2) & (rfm_df['Monetary'] > 400)]

print("Customers worth investing more in:")

print(high_value_customers)

(4) Who are your most loyal and engaged customers?

Customers who frequently make purchases and spend a significant amount of money are the most loyal.

# High Frequency & High Monetary -> Loyal & engaged customers

loyal_customers = rfm_df[(rfm_df['Frequency'] > 3) & (rfm_df['Monetary'] > 300)]

print("Loyal and engaged customers:")

print(loyal_customers)

(5) How do I predict customer churn and retain valuable clients?

Customers who have high recency, low frequency, and low monetary are at risk of churning. These customers need attention.

# High Recency & Low Frequency -> Potential churn

churn_risk = rfm_df[(rfm_df['Frequency'] <= 1) & (rfm_df['Recency'] < 60) & (rfm_df['Monetary'] < 200)]

print("Churn risk customers:")

print(churn_risk)

(6) How can I increase customer lifetime value?

Focus on high-frequency and high-monetary customers and personalize offers to increase their lifetime value.

# High Frequency & High Monetary -> High potential for increased LTV

increase_ltv_customers = rfm_df[(rfm_df['Frequency'] > 3) & (rfm_df['Monetary'] > 300)]

print("Customers with potential for increased LTV:")

print(increase_ltv_customers)

(7) How do I optimize my marketing spend?

Focus your marketing efforts on high-value customers to maximize ROI.

# High Frequency & High Monetary -> Optimize marketing spend

optimize_spend = rfm_df[(rfm_df['Frequency'] > 3) & (rfm_df['Monetary'] > 300)]

print("Customers to focus marketing efforts on:")

print(optimize_spend)

(8) How do I identify seasonal trends or patterns?

You can analyze Recency and Frequency by month or season to identify buying patterns.

# Adding a 'Month' column to check seasonal trends

df['Month'] = df['TransactionDate'].dt.month

# Calculate average spending by month

monthly_spending = df.groupby('Month').agg({

'Amount': 'sum',

'CustomerID': 'nunique'

}).rename(columns={'Amount': 'TotalAmount', 'CustomerID': 'UniqueCustomers'})

print("Monthly spending and customer activity:")

print(monthly_spending)

Visualizations:

To visualize the results, we can create some plots.

# Plot Recency vs Frequency for customers

plt.figure(figsize=(8,6))

sns.scatterplot(data=rfm_df, x='Recency', y='Frequency', hue='Monetary', palette='viridis', size='Monetary', sizes=(20, 200))

plt.title('Recency vs Frequency with Monetary Size')

plt.show()

# Plot customer segments based on Frequency and Monetary