Strategizing Against High CAC User Retention & Cohort Analysis

👋
Project Highlights: This page provides an executive summary of the analysis. For the complete code, step-by-step data cleaning, and detailed statistical modeling, please refer to the Full Project Links at the bottom of this page.

In this project, transaction data was meticulously cleaned by handling missing values, duplicates, and outliers to ensure high-quality analysis. A suite of new features was then engineered, including date extraction and monthly activity identification.

Subsequently, a Cohort Analysis was constructed using the first purchase month as the grouping baseline. Retention rates were calculated and presented in a retention matrix visualized via heatmap, providing clear visibility into customer behavior.

📊 Visualization: Retention Matrix Heatmap
Cohort Analysis Heatmap

💡 Key Findings from Heatmap

  • Steep First-Month Drop-off: Retention drops significantly after the first month across all cohorts. Specifically, the Jan 2010 cohort saw a drop to 36% retention in Month 2, meaning 64% of customers churned immediately.
  • Long-term Stabilization: Despite the initial drop, retention stabilizes over time. The Jan 2010 cohort maintained a consistent 35-47% retention rate up to the 12th month, indicating a strong base of loyal customers.
  • Varying Cohort Performance: Cohorts from Feb & Mar 2010 showed slightly lower long-term retention compared to Jan 2010, suggesting potential issues with acquisition quality or campaigns during those months.
  • "Smile/U-Pattern" Effect: A slight retention increase was observed in later months (e.g., Jan 2010 increased from 38% to 43% in Month 10), likely driven by seasonal factors or successful reactivation efforts.

🚀 Actionable Recommendations

1. Revamp the Onboarding Journey

Since the biggest churn occurs in Month 1, focus on the first-purchase experience. Implement immediate incentives like a "Discount for 2nd Purchase" or personalized product guides to secure early retention.

2. Launch "We Miss You" Campaigns

To lift the flattening retention curve for older cohorts, implement automated re-engagement emails with special offers targeting users who have been dormant for 3+ months.

🐍 Python Code: Cohort Logic