You can access the interactive simulation dashboard here: Profit Optimization Engine - Live Dashboard
🚨 Despite generating $1.98B in revenue, this company is losing $617M.
This project builds a decision engine to identify why and simulate how to fix it.
📌 Executive Summary This project delivers a high-performance decision-support system for a global e-commerce entity operating in Turkey, Germany, and the UAE across D2C and B2B channels. Despite generating approximately $1.98 Billion in gross revenue, the company faced a critical $617.8 Million net loss.
This platform transforms static historical data into an interactive simulation engine, allowing executives to test "What-If" scenarios regarding marketing spend, return rates, and shipping shocks in seconds.
📉 The Business Challenge (Data-Driven Insights) Comprehensive analysis identified several critical "Profit Leaks":
Erosive Return Rates: An overall 17.81% return rate significantly impacts the bottom line, resulting in a $178.6 Million loss in net revenue.
Segment-Specific Crises:
Efficiency Crisis (UAE D2C Entry): This segment exhibits the lowest efficiency with a profit margin plummeting to -34.89%.
Unit Loss Crisis (Germany B2B Premium): Analysis shows the highest loss per order occurs in this segment, reaching up to -$10,000 per transaction.
FX & Operational Pressure: With COGS and shipping costs heavily influenced by USD exposure, international logistics costs ($197.2M) create unpredictable margin pressure.
Marketing Inefficiency: High spending in segments with negative ROI prevents top-line growth from converting into sustainable profit.
🎯 Project Objectives & Strategic Goals The core mission is to provide an actionable framework for profit stabilization:
Granular Profitability Mapping: Visualizing net margins across Country x Channel x Segment using multidimensional heatmaps.
Real-time Scenario Simulation: Utilizing a Waterfall architecture to monitor how variables like marketing shifts or shipping cost shocks affect the final net profit.
Strategic Optimization: Identifying "Quick Win" segments (e.g., UAE D2C Entry) where targeted interventions in return rates can maximize investment efficiency.
🛠️ Technology Stack & Architecture Built on a modern data science stack designed for speed and scalability:
FastAPI: Serves the simulation logic as a high-performance, asynchronous backend API.
Streamlit: Provides a user-centric frontend with interactive sliders for real-time strategic testing.
Pandas & NumPy: Handles complex calculations across 300K+ rows of simulated transaction data.
Plotly: Powers dynamic Waterfall and Heatmap visualizations.
PostgreSQL / SQLAlchemy: Ensures production-ready data persistence and structured querying.
📊 Success Metrics Decision Agility: Reduced complex "What-If" analysis time from hours (manual Excel) to under 1 second.
Impact Potential: Simulations demonstrate that a 20% improvement in return rates can reduce total losses by over $140 Million.
Model Consistency:The generated data maintains internal consistency across financial metrics and business logic.
👤 User Stories Marketing Manager: "If I increase the D2C Premium budget by 20% while reducing returns by 10%, how does the net margin shift?"
Operations Director: "How will a 15% shock in shipping costs (FX-driven) impact our German B2B segment viability?"
📂 Data Architecture
The schema is designed to enable full traceability from order-level transactions to final profitability metrics.
orders: Core transaction data (Country, Channel, Segment, Revenue).
cost_structure: Detailed breakdown of COGS, labor, shipping, and marketing.
behavior: Tracking return status and payment cycles.
fx_rates: Historical and simulated currency conversion rates.
🚀 Recommended Strategy
Based on simulation results, the most effective path to reduce losses is:
-
Prioritize return rate reduction in D2C channels → Even a 20% improvement reduces losses significantly
-
Reduce exposure to unprofitable B2B Premium segments → High loss per order makes this segment structurally unsustainable
-
Optimize marketing allocation → Shift budget toward segments with lower return rates and higher efficiency
-
Closely monitor shipping cost volatility (FX-driven) → Implement dynamic pricing or cost controls
👉 Key Insight: Profitability improvement is not driven by revenue growth, but by cost and return optimization. The primary driver of losses is not insufficient revenue, but structural inefficiencies:
- High return rates eliminate realized revenue
- COGS nearly equals gross revenue, leaving no margin buffer
- Fixed costs (marketing, logistics, labor) push the business into negative profitability
👉 This indicates a fundamentally broken unit economics model.
💡 Final Takeaway: This is not a revenue problem it is a unit economics problem.
Without structural fixes, scaling revenue will only scale losses.
Melek İkiz
Building data-driven decision systems
📍 Türkiye
🔗 LinkedIn: [(https://www.linkedin.com/in/melek-ikiz-520065373/)]