Predicting Customer Lifetime Value Using Behavioral Segmentation and Machine Learning - A Data-Driven Approach to Customer Profitability in E-Commerce

Aug 5, 2026

In a competitive e-commerce environment, estimating Customer Lifetime Value (CLV) enables data-driven decisions about acquisition, retention, and investment efficiency. The project analyzes the link between customer behavior and long-term profitability using advanced analytics and machine learning. A large anonymized e-commerce dataset is transformed into behavioral features via the RFM framework, then customers are segmented using K-Means clustering and PCA to improve visualization and interpretability. Findings show strong signals from frequency and monetary value for identifying high-value customer groups and guiding marketing resource allocation.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages33

Reading Time1 h 23 min

AccessFree to download

Acknowledgment

Table of Contents

Abstract

Introduction

Research Scope and Objectives

Literature Search Strategy and Research Questions

Literature Review

Methodology

Modeling Approach

Results and Discussion

Conclusion

References

Predicting Customer Lifetime Value Using Behavioral Segmentation and Machine Learning: A Data-Driven Approach to Customer Profitability in E-Commerce Creative Component Project Report By: Varshith Thudum Master of Science in Information Systems Submitted in fulfillment for the requirements for the degree of Master of Science in Information Systems (MSIS) Major Professor: Dr. Anthony Townsend Ivy College of Business Iowa State University Ames, Iowa 2025 Acknowledgment I would like to begin by expressing my deepest gratitude to my major professor, Dr. Anthony Townsend, whose mentorship and encouragement have played a central role in the completion of this project. Their expert insights, clear guidance, and thoughtful feedback consistently challenged me to think critically and refine my approach at every stage. I ’m also grateful for the support of my department and the learning environment that enabled me to explore this topic thoroughly. The access to tools, resources, and research guidance was essential in bringing this project to life. A special thank you goes to my friends, who were always ready to brainstorm with me, review my work, or just listen when I needed to talk through an idea. Whether it was helping debug a script or offering a second opinion on a model evaluation, your presence made this journey collaborative and motivating. This report is the result of not just individual effort, but a network of support, for which I am truly grateful. Table of Contents ACKNOWLEDGMENT ..............................................................................................................2 TABLE OF CONTENTS ..............................................................................................................3 ABSTRACT ................................................................................................................................4 INTRODUCTION .......................................................................................................................5 LITERATURE REVIEW .............................................................................................................8 METHODOLOGY .................................................................................................................... 11 MODELING APPROACH .........................................................................................................15 RESULTS AND DISCUSSION ...................................................................................................19 CONCLUSION..........................................................................................................................28 REFERENCES..........................................................................................................................31 Abstract In a competitive e-commerce environment, understanding and estimating Customer Lifetime Value (CLV) is essential for data-driven decision-making. This project explores the relationship between customer behavior and long-term value using advanced data analytics and machine learning techniques. Using a large, anonymized e-commerce dataset, we engineered key behavioral features—including Recency, Frequency, and Monetary value (RFM)—to quantify customer engagement and purchasing patterns. To enhance insight and business interpretability, we applied unsupervised learning with K-Means (Huang & Kechadi, 2013) clustering to segment customers based on behavioral traits. Principal Component Analysis (PCA) was used to reduce dimensionality and improve cluster visualization and interpretability. The results confirmed that customer behavior metrics, particularly frequency and monetary value, are strong indicators of customer value potential. The segmentation framework developed in this project can assist businesses in identifying high-value customer groups, personalizing retention strategies, and allocating marketing resources m

FAQs

What is the core objective of the project?

To predict Customer Lifetime Value (CLV) by identifying how customer behavioral patterns relate to long-term profitability in e-commerce.

How are customers represented for modeling?

The project engineers behavioral features using the RFM framework—Recency, Frequency, and Monetary value—to quantify engagement and purchasing patterns.

Why are K-Means and PCA used in the approach?

K-Means clustering segments customers based on behavioral traits, while PCA reduces dimensionality to improve visualization and interpretability of the clustering results.

What behaviors most strongly indicate customer value?

The results highlight frequency and monetary value as strong indicators of a customer’s potential value.

Aug 5, 2026

Upload by Seraphina