Case Study by Navneet Singh

Check out the project at: https://github.com/Navneet022/Product-churn-rate-analysis

🎯 Business Problem

How can we improve customer retention and reduce churn by analyzing user purchasing behavior?

Most businesses face a silent churn problem, where users stop engaging or buying without saying anything. Often, 80% of revenue comes from 20% of customers, so retaining loyal users and reactivating slipping ones is critical.

🔍 Project Overview

A simulated end-to-end product analytics system using SQL, Excel, and Python to replicate real-world e-commerce analysis. The project includes data cleaning, churn modeling, cohort analysis, RFM segmentation, pricing strategies, and business recommendations, all backed by real-world-style messy data and stakeholder-ready insights.

📊 Tools & Stack

• SQL – for cleaning, joining, filtering data
• Excel – for cohort modeling, churn segmentation, pivoting
• Python (Pandas, NumPy, Seaborn, Matplotlib) – for RFM segmentation, churn prediction, seasonal analysis
• Tableau - Customer Churn & Revenue Recovery Dashboard
• Google Colab – environment used
• Dataset – 12K+ orders, 4K+ users, 1K+ products

🛠 Approach

1. Problem Definition

• Began by framing the central question: “When are customers most likely to churn, and how can we intervene to retain them?”
• Expanded the scope to include customer value tiers, purchasing patterns, and seasonal demand shifts for a comprehensive view of retention and revenue.

1. Analysis Plan
   • SQL: Clean, join, and prepare multi-table data to create a complete view of the customer journey.
   • Excel: Build cohorts and visualize churn timing to identify early retention drop-off points.
   • Python: Perform RFM segmentation, advanced analytics, and churn prediction validation.
   • Tableau/Power BI: Present retention curves, segmentation visuals, and seasonal trends in a stakeholder-ready format.
2. Working Hypotheses
   • Significant churn occurs within the first 30 days after signup or first purchase.
   • The “neglected majority” segment has untapped upsell potential.
   • Certain days and months consistently underperform in demand.