This project analyzes the marketing funnel performance of a retail brand by understanding how customers move from awareness → engagement → campaign interaction → final conversion.
The goal of the analysis is to identify drop-offs, segment customer behavior, and generate actionable insights that help improve marketing effectiveness and overall ROI.
This project was executed end-to-end with a complete data analytics pipeline including data cleaning, EDA, KPI calculation, segmentation, funnel mapping, and insight generation.
- Understand customer behavior across the marketing funnel
- Identify bottlenecks where customers drop off
- Analyze spending patterns and engagement behavior
- Study campaign performance using past campaign acceptance
- Provide business recommendations to improve conversions
- Present insights that support data-driven marketing decisions
The dataset contains detailed customer information with 29 columns including:
| Feature Category | Description |
|---|---|
| Demographics | Year of birth, education, marital status, income |
| Household Info | Number of kids/teens in home |
| Customer Activity | Web visits, store purchases, catalog purchases |
| Spending Behavior | Wines, meat, fish, sweets, gold products |
| Campaign Data | AcceptedCmp1–5, Response (latest campaign) |
| Churn Indicators | Complaint history, recency |
✔ Total Rows: 2240
✔ Funnel-related columns: WebVisits, Purchases, Campaign Acceptance, Response
NumWebVisitsMonth→ Measures visibility and website interest
NumWebPurchasesNumCatalogPurchasesNumStorePurchasesNumDealsPurchases- Spending columns (
MntWines,MntMeatProducts, etc.)
AcceptedCmp1toAcceptedCmp5→ Past campaign acceptance
Response→ Whether the customer accepted the latest campaign
This funnel structure helps analyze the complete customer journey from awareness to conversion.
- High website visits did not always lead to purchases → awareness does not guarantee engagement.
- High-spending customers (especially in Wines, Meat, Gold products) showed higher campaign conversion rates.
- Deal-driven purchases increased engagement but did not guarantee campaign conversion.
- Past campaign acceptance was a strong predictor of the latest campaign response.
- Target high-value customers with personalized premium campaigns.
- Improve website and catalog purchasing experience to reduce early drop-offs.
- Use historical campaign acceptance to personalize offers.
- Retarget high-visit, low-purchase users with optimized messaging.
- Python → Pandas, NumPy, Seaborn, Matplotlib
- MySQL → Data extraction and SQL-based insights
- Jupyter Notebook → Data cleaning & EDA
- PowerPoint → Insights presentation for stakeholders
| File | Description |
|---|---|
Marketing_jupyterr.ipynb |
Full EDA, funnel mapping, KPI analysis |
presentation.pptx |
Stakeholder presentation |
README.md |
Documentation |
- Identified the main drop-off stage in the marketing funnel
- Improved understanding of customer segments & behavior
- Delivered actionable insights to improve marketing ROI
- Enabled better campaign targeting using behavioral patterns