🚀 📊 Marketing Mix Analysis Sales Optimization using Regression (MMM-Inspired Approach)

---

📌 Overview

Marketing teams often struggle to understand which channels truly drive sales and how to optimize their budget for maximum ROI.

This project applies a regression-based approach inspired by Marketing Mix Modeling (MMM) to quantify the impact of advertising channels on sales and generate actionable business insights.

---

🎯 Business Problem

Companies invest heavily in:

• 📺 TV Advertising
• 📻 Radio Advertising
• 📰 Newspaper Advertising

But key questions remain:

• Which channel drives the most sales?
• Where should we increase or reduce budget?
• How can we maximize return on investment (ROI)?

👉 This project answers these questions using data + statistical modeling

---

📊 Dataset

• 📦 200 observations

• 📈 Features:

  • TV Advertising Spend
  • Radio Advertising Spend
  • Newspaper Advertising Spend

• 🎯 Target:

  • Sales

---

🛠️ Tech Stack

• Python → Pandas, NumPy
• Visualization → Matplotlib, Seaborn
• Statistical Modeling → Statsmodels (OLS)
• Evaluation → Scikit-learn

---

🔍 Project Workflow

1️⃣ Data Cleaning

• Checked missing values & duplicates
• Ensured clean and consistent dataset

2️⃣ Exploratory Data Analysis (EDA)

• Identified strong relationships between variables
• Observed high correlation between TV, Radio → Sales

3️⃣ Model Building (OLS Regression)

• Built regression model:

  Sales ~ TV + Radio + Newspaper
  

• Evaluated statistical significance using p-values

4️⃣ Model Refinement

• Identified Newspaper as statistically insignificant
• Improved model interpretability by focusing on key drivers

---

📈 Model Performance

Metric	Value
R² Score	0.897
Adjusted R²	0.896
F-statistic	570.3
Model Significance	p < 0.001

👉 The model explains ~90% of variance in sales, indicating strong predictive power.

---

📊 Key Insights

• 📻 Radio is the strongest driver of sales
• 📺 TV significantly impacts sales
• 📰 Newspaper has no meaningful impact
• 📌 Marketing channels contribute unequally

---

💰 ROI Insights

• 🥇 Radio → Highest return per unit spend
• 🥈 TV → Moderate ROI
• ❌ Newspaper → Low / negligible ROI

---

🔮 Scenario Simulation (What-if Analysis)

new_budget = pd.DataFrame({'TV': [200],'Radio': [50],})
predicted_sales = final_model.predict(new_budget)

print(f"Predicted Sales:{predicted_sales[0]: .2f} units")

👉 Enables:

• Budget simulation
• Sales forecasting
• Decision support

---

📢 Business Recommendations

• 🔼 Increase investment in Radio advertising
• 🔼 Optimize TV campaigns
• 🔽 Reduce or eliminate Newspaper spend
• 🎯 Adopt data-driven budget allocation

---

⚠️ Limitations

• Assumes linear relationships

• Residuals are not perfectly normally distributed

• No external factors included:

  • Seasonality
  • Competition
  • Pricing

---

🚀 Future Scope

• 📈 Implement advanced Marketing Mix Modeling (MMM)

• ⏳ Add time-series analysis (trend & seasonality)

• 🔁 Apply Adstock (lag effect of advertising)

• 🤖 Explore advanced ML models:

  • Random Forest
  • XGBoost
  • Ridge / Lasso

---

📂 Project Structure

Marketing-Mix-Analysis/
│
├── data/
│   └── Advertising.csv
│
├── notebook/
│   └── Marketing Mix Modeling (MMM)_Sales Driver & ROI Analysis.ipynb
│
└── README.md

---

🧠 Key Learnings

• Built and interpreted an OLS regression model
• Translated statistical output into business insights
• Identified high-impact marketing channels
• Applied data-driven decision-making concepts

---

🧾 Conclusion

This project demonstrates how a regression-based MMM approach can uncover hidden insights in marketing data.

👉 By focusing on high-performing channels like Radio and TV, businesses can significantly improve:

• ROI
• Budget efficiency
• Overall sales performance

---

⭐ If you found this project useful, consider giving it a star!