Product Retention & Cohort Analysis

Project Overview

Product Retention & Cohort Analysis is an end-to-end analytics project that transforms raw user event data into actionable product insights. This interactive dashboard demonstrates the complete analytical workflow—from Python data processing to executive-ready visualizations.

What This Project Does

This dashboard answers critical product questions that hiring managers care about:

• Which user cohorts retain best? — Visual retention heatmap with color-coded percentages
• What's our Day 7 / Day 30 retention? — Benchmarked curves showing performance over time
• Which acquisition channel delivers highest LTV? — Ranked comparison with retention + revenue metrics
• Does onboarding completion matter? — A/B-style analysis showing +81% retention lift

Key Findings

🎯 Onboarding is the #1 Retention Driver

• Users who complete onboarding: 38% Day 30 retention
• Users who skip onboarding: 21% Day 30 retention
• +81% retention lift from onboarding completion
• Business Impact: Improving completion by 25% → +4.3pp to 30-day retention

📊 Channel Quality Varies 2.1x in LTV

📈 Retention is Improving

Later cohorts (Jan 2024+) show 8-12% better Month-1 retention, suggesting recent product changes are working.

Recommended A/B Tests

1. Onboarding Progress Bar

   • Hypothesis: Visible progress indicator increases completion 15-20%
   • Expected Impact: +4.3pp to 30-day retention

2. Referral Incentive Program

   • Hypothesis: $25 credit per referral shifts mix to higher-LTV users
   • Expected Impact: +$4.70 LTV per 10% mix shift

3. Day-7 Re-engagement Email

   • Hypothesis: Personalized feature highlights prevent early churn
   • Expected Impact: +3-4pp to 30-day retention

Technical Implementation

Data Pipeline

Python/Pandas → JSON API → React/Next.js Dashboard
     ↓              ↓              ↓
  65K events    Processed     Interactive viz
  Cohort logic  metrics       Dark theme UI

Technologies

• Python/Pandas — Data processing, cohort assignment, retention calculations
• React/TypeScript — Component-based dashboard with animations
• Tailwind CSS — Dark theme matching portfolio aesthetic
• GitHub — Source control and notebook hosting

Dataset

• 3,000 users across 21 months (Jan 2023 - Sep 2024)
• 65,493 events (signup, onboarding, purchases, visits, churn)
• Realistic SaaS patterns: ~60% Day 1, ~35% Day 7, ~20% Day 30 retention
• Multi-channel: organic, paid_social, referral with different profiles

Project Structure

cohort-retention-analysis/
├── dashboard/
│   ├── index.html          # Standalone HTML dashboard
│   └── data.json           # Processed data API
├── data/
│   ├── user_events.csv     # 65K+ raw events
│   └── generate_user_events.py
├── notebooks/
│   └── cohort_retention.ipynb  # Full Python analysis
├── app/components/
│   └── CohortDashboard.tsx # React component (this page)
├── README.md
└── requirements.txt

Links

• 🌐 Live Dashboard — Interactive visualization
• 💻 GitHub Repository — Source code & data
• 📓 Jupyter Notebook — Full Python analysis

Skills Demonstrated

This project showcases the exact skills hiring managers look for in Product Analyst and Growth Analyst roles.