Business Intelligence Report  /  Sania Sultana
Business Analysis Case Study · E-Commerce Domain

Customer Churn &
Revenue Leakage Analysis

A data-driven investigation into why ZylaCart — a mid-scale D2C e-commerce brand — was losing ₹1.2Cr/month in preventable revenue. From raw behavioral data to boardroom recommendations.

Analyst Sania Sultana
Domain E-Commerce / D2C Retail
Dataset 120K transactions · 18 months
Tools Used Excel · SQL · Power BI · Python
Projected ROI +₹3.6Cr/year
Period Jan 2024 – Jun 2025
01 Business Context & Problem Statement

🎯 Business Problem

ZylaCart's management observed a 34% decline in repeat purchase rate over 6 months despite increasing new customer acquisition spend. CAC rose 58% while revenue grew only 12% — a clear sign of a retention & monetisation breakdown.

📋 Analytical Objectives

Identify the root cause of revenue stagnation by analysing customer lifecycle, purchase patterns, cart abandonment, and product-level profitability. Quantify each leakage point and recommend prioritised interventions.

📊 Scope of Analysis

18-month transactional data covering 120K orders, 48K unique customers, 6 product categories, and 3 regional markets (South, West, North India). Focus on FY2024–25 post-pandemic normalisation phase.

⚠️ Stakeholder Impact

Revenue leak of estimated ₹1.2Cr/month identified across cart abandonment (₹42L), inactive customers (₹38L), and poor cross-sell conversion (₹40L). Business needs actionable fixes within 90-day sprint cycle.

Hypothesis

High first-purchase discount dependency is creating transactional-only customers with no loyalty trigger, resulting in <90-day churn post-acquisition.

Methodology Used

RFM Segmentation · Cohort Retention · Funnel Analysis · ABC Product Classification · What-if Revenue Modelling

Expected Output

Prioritised recommendation matrix with estimated revenue impact, effort scoring, and 30-60-90 day implementation roadmap for the product & marketing teams.

02 Executive KPI Summary — FY2024–25
💰
Total Revenue
₹8.4Cr
▲ +12.3% YoY
vs ₹7.5Cr FY2023–24
📉
Repeat Purchase Rate
28.4%
▼ −34% vs prev year
Benchmark: 45–55%
🎯
Customer Acq. Cost
₹840
▲ +58% — critical
LTV:CAC = 2.1x (low)
👥
Customer LTV
₹1,764
→ Flat (no growth)
Target: ₹3,500+
🛒
Cart Abandon Rate
68.2%
▲ +9pts — worsening
Industry avg: 55–60%
Monthly Revenue vs New Customer Trend
₹ Revenue (bars) & New Customers Acquired (line) — Jan 2024 to Jun 2025
Revenue by Customer Segment
% contribution to total revenue
Champions 31%
Loyal 24%
At-Risk 22%
Lost 14%
New 9%
03 RFM Customer Segmentation Analysis
RFM Score Distribution — 48,000 Customers
Recency × Frequency × Monetary analysis across 5 scored segments
Segment Breakdown
% of customer base with avg LTV
🏆
Champions

18% of base · Avg LTV ₹4,200

💛
Loyal Customers

22% of base · Avg LTV ₹2,800

⚠️
At-Risk

29% of base · Used to buy but inactive 90d+

🌱
New Customers

31% of base · 1 purchase only, high churn risk

04 Conversion Funnel & Cohort Retention
Purchase Conversion Funnel
From website visit to successful order placement — Q4 FY2025
Website Visit
1,00,000
100%
Product View
62,000
62%
Add to Cart
38,000
38%
Checkout Start
22,000
22%
Payment Done
14,000
14%
Order Placed ✓
11,800
11.8%
🔴 Critical drop: Cart → Checkout — 42% abandonment at checkout initiation, primarily due to unexpected shipping costs and mandatory account creation.
Cohort Retention Heatmap
% of customers returning each month after first purchase
Cohort M0M1M2M3M4M5
Low
High Key insight: All cohorts drop to <20% by M2
05 Product & Category Performance
Revenue by Category — ABC Classification
A-class (top 20% products generating 80% revenue) vs B & C-class
Avg Order Value Trend
Monthly AOV movement — signals pricing sensitivity
06 Key Findings & Actionable Recommendations
F1

Core Finding Discount Dependency Cycle Breaking LTV

83% of first-time purchases are made using a discount code (avg 20–30% off). These customers show a 91% churn rate within 60 days if not given another discount. ZylaCart is essentially paying ₹840 CAC to acquire customers who buy once at a loss and never return — creating a unit economics death spiral.

💡 Estimated Revenue Recovery: ₹1.8Cr/year if repeat rate lifted to 45%
R1

Recommendation Launch Value-Based Loyalty Programme

  • Replace flat discounts with a points-based rewards system (earn on every ₹100 spent)
  • Introduce tier upgrades at ₹2K, ₹5K, ₹10K lifetime spend with exclusive perks
  • Target "At-Risk" segment (29%) with personalised win-back emails at Day 60 of inactivity
  • A/B test "loyalty reveal" at checkout to increase perceived value of repeat purchase
R2

Risk Mitigation Fix Cart Abandonment (₹42L/month loss)

  • Remove mandatory account creation — switch to guest checkout with optional save
  • Show shipping cost upfront on product page, not at checkout — reduces checkout shock
  • Implement 3-step cart recovery email sequence (1hr → 24hr → 72hr with social proof)
  • Add UPI one-click payment; 48% of drop-offs happen at payment step
R3

Opportunity Cross-Sell to Champions Segment

  • Top 18% Champions generate 31% revenue — zero cross-sell campaigns currently active
  • Build product affinity matrix using co-purchase data (SQL analysis completed)
  • Create "Curated for You" email series based on category purchase history
  • Projected uplift: +15% revenue from existing customers, zero new CAC spend
Recommendation Priority Matrix
Effort vs. Revenue Impact — 90-day implementation roadmap
🟢 Do Now (0–30 days)
• Guest checkout toggle
• Upfront shipping cost display
• Cart recovery email #1 (1hr)
• UPI payment integration
🟡 Plan (30–60 days)
• Loyalty programme MVP launch
• At-Risk win-back campaign
• 3-step recovery email sequence
• Champions cross-sell series
🔵 Scale (60–90 days)
• Personalised product affinity model
• Tier loyalty upgrade mechanics
• CAC reduction via referral system
• Monthly cohort tracking dashboard
07 Methodology & Tools
01
Data Collection
SQL queries on transactional DB, GA4 export, CRM data
02
Data Cleaning
Python (pandas), removed nulls, normalised customer IDs
03
RFM Scoring
Excel pivot, quintile scoring, 5-segment classification
04
Cohort Analysis
SQL cohort query, heatmap via Python matplotlib
05
Funnel Mapping
GA4 funnel report + custom SQL drop-off queries
06
Dashboarding
Power BI (KPI views), Excel dynamic charts
07
BRD / FRD
Requirements doc, user stories, stakeholder sign-off
Microsoft Excel (Advanced)
Power BI Desktop
Python (pandas, matplotlib)
SQL (PostgreSQL)
Google Analytics 4
Google Sheets
Confluence (BRD docs)
Jira (Agile tracking)