605 movies. Decades of franchises. One question: when does the sequel actually pay off?
Who: Gyanesh Samanta — solo author, written as a research paper.
What: A data-driven study of franchise economics in cinema — comparing originals vs. sequels across budget, lifetime gross, ROI, and genre moats.
When: Authored February 2026.
Where: Jupyter notebook + accompanying paper, using a curated MovieFranchises.csv dataset of 605 movies across major studios and franchises.
Why: Hollywood has bet hard on franchises for two decades. This paper examines whether the bet is still paying off — and which franchise traits separate the winners from the burnouts.
The dataset spans 605 movies, 13 fields per movie (lifetime gross, budget, year, studio, rating, runtime, vote average, franchise ID, and more), going back to Star Wars: A New Hope (1977, $11M budget, $775M gross).
The analysis answers four questions:
- Does more budget mean more gross? Yes — but with diminishing returns. Past a certain budget threshold, sequels show meaningfully lower marginal ROI than their originals.
- Which franchises set the ceiling? Star Wars and the MCU (Avengers arc) dominate the lifetime-gross leaderboard, defining the benchmark every other franchise is measured against.
- Are franchise ROIs declining? Yes. Rising production and marketing costs have compressed ROI on modern installments even as gross headlines grow.
- What makes a "moat"? Action and Adventure genres carry most successful franchises, with PG/PG-13 ratings dominating the high-gross tier.
The paper contributes a working framework for evaluating sequel marketing decisions: budget tier × genre × installment number is a stronger predictor of ROI than star power alone.
| Budget vs. Gross | Top Franchises |
|---|---|
 |
|
| ROI Over Time | Genre Moats |
|---|---|
 |
 |
Additional plots (analysis_plot_3 through analysis_plot_8) cover studio splits, rating distributions, and runtime trends.
- Python 3.x — language
- pandas / NumPy — data wrangling
- Matplotlib / Seaborn — visualization
- Jupyter Notebook — analysis environment
Sequel-Marketing-in-Cinema-Analysis/
├── Cinematic_Franchise_Analysis.ipynb # Full notebook
├── MovieFranchises.csv # 605 movies, 13 columns
├── images/ # 9 generated plots
└── README.md
git clone https://github.com/GyaneshSamanta/Sequel-Marketing-in-Cinema-Analysis.git
cd Sequel-Marketing-in-Cinema-Analysis
pip install pandas numpy matplotlib seaborn jupyter
jupyter notebook Cinematic_Franchise_Analysis.ipynbRun all cells top-to-bottom; plots render inline and are also saved to images/.
This is a research artifact, but extensions welcome — particularly:
- Sentiment analysis correlating marketing hype with opening-weekend gross
- Predictive modeling (Random Forest / XGBoost) for "success" classification
- A "star power" index from cast metadata
Released for academic and educational use.
Authored by Gyanesh Samanta (@GyaneshSamanta). Dataset compiled from public box-office and franchise databases.