This project presents a structured, end-to-end marketing analytics workflow that connects experimental analysis with financial performance and strategic decision-making. It evaluates whether advertising improves user behavior, assesses whether that improvement translates into profitability, and determines how to allocate budget under constrained conditions.
The analysis moves beyond descriptive metrics to answer a core business question:
What is the true impact of advertising on conversions and profitability, and how should limited budget be allocated to improve efficiency?
- Ad campaign conversion rate: ~2.55%
- Control group conversion rate: ~1.79%
- Relative conversion lift: ~43%
- Statistical significance: p < 0.001
- Effect size: small but reliable
- ROI across all ad formats: negative
- Optimal budget allocation: prioritize static ads, reduce carousel exposure
Marketing teams often optimize campaigns based on engagement or conversion metrics without evaluating financial outcomes. This creates a disconnect between performance indicators and business value.
This project addresses three fundamental questions:
- Do ads significantly improve conversion rates?
- Are those improvements financially sustainable?
- How should budget be allocated when all channels underperform?
- Cleaned and standardized user-level dataset
- Engineered features such as exposure buckets and time segments
- Structured data for both statistical analysis and visualization
- Two-proportion z-test to compare conversion rates
- Confidence interval estimation for uplift
- Effect size (Cohen’s h) to measure practical impact
-
Simulated revenue and cost structure per ad type
-
Computed:
- Total cost
- Revenue
- Profit
- Return on investment (ROI)
- Developed ROI-weighted allocation model
- Compared equal vs optimized allocation
- Designed strategy to minimize losses under fixed budget constraint
Focus:
- Ad vs control group performance
Metrics:
- Conversion rate
- Absolute and relative lift
- Statistical significance
Insight: Advertising significantly improves conversion rates, but the magnitude of improvement is modest in practical terms.
Focus:
- Financial performance by ad type
Metrics:
- Cost
- Revenue
- Profit
- ROI
Insight: All ad formats generate negative ROI, indicating that increased conversions are insufficient to offset acquisition costs.
Focus:
- Allocation of a fixed ₹50,000 budget
Approach:
- ROI-based weighting
- Comparison of equal vs optimized distribution
Insight: Budget allocation should prioritize minimizing losses. Static ads receive the highest allocation due to relatively better performance, while carousel ads are significantly reduced.
- Conversion improvements do not guarantee profitability
- Cost structure and revenue per conversion are critical drivers of ROI
- Negative ROI across channels signals inefficiency in unit economics
- Budget decisions should be based on relative performance, not intuition
- Scaling campaigns without improving ROI leads to financial loss
- Python (Pandas, NumPy) for data processing
- Statistical methods for experiment validation
- Tableau for interactive dashboards
- Level of Detail (LOD) expressions for stable calculations
- CSV datasets for structured analysis
-
Open the
.twbxfile in Tableau -
Navigate across dashboards:
- A/B Testing
- ROI Analysis
- Budget Optimization
-
Review insights and recommended strategy
- Revenue is simulated due to lack of real financial data
- No customer lifetime value (LTV) included
- No segmentation by user cohorts or geography
- Static assumptions for cost across ad types
- Incorporate customer lifetime value into ROI analysis
- Add segmentation by audience and behavior
- Develop predictive models for conversion and revenue
- Introduce dynamic budget allocation based on real-time performance
This project demonstrates how to move from experimental analysis to business decision-making. It highlights the importance of evaluating both statistical significance and financial impact, and provides a practical framework for optimizing marketing spend under real-world constraints.
The result is a cohesive analytical workflow that aligns data insights with actionable strategy.