VarunGowda/pcb-financial-feasibility-analysis

Financial Feasibility Analysis of a Small-Scale PCB Fabrication Center

Extensions: Monte Carlo Simulation & STP-based Segmentation

Based on the IEEE paper: "Financial Feasibility Analysis of a Small-Scale PCB Fabrication Center"

---

Quick Start — Generate All Outputs

git clone https://huggingface.co/VarunGowda/pcb-financial-feasibility-analysis
cd pcb-financial-feasibility-analysis
pip install -r requirements.txt
python run_all.py

This generates 34 PNG figures, 14 JSON data files, and 2 JSON result files. Then open the notebook:

jupyter notebook pcb_financial_analysis.ipynb

---

Paper Summary

Parameter	Value
CapEx	INR 1,80,000
Monthly OpEx	INR 50,000
Total Initial Investment	INR 2,00,000
Revenue Model (Eq. 3)	Fixed N_n = 150, churn α = 10%, avg price P = INR 400
Monthly Revenue (deterministic)	INR 60,000
Break-even	N_BE = C_f / (P - C_v) = 1,200 PCBs
Deterministic Payback	~12 months

---

Extension 1: Monte Carlo Cash Flow Simulation

Model Specification

• Monthly Demand: D ~ Poisson(λ=100)
• Price per PCB: P ~ Normal(μ=400, σ²=40²)
• Monthly OpEx: O ~ Uniform(42,500, 57,500)
• Initial Investment: INR 2,00,000
• Trials: 10,000
• Metric: Months until cumulative net cash flow recovers INR 2,00,000

⚠ Critical Mathematical Insight

With λ=100 (user-specified):

E[Revenue] = λ × μ_P = 100 × 400 = INR 40,000
E[OpEx]    = (42,500 + 57,500) / 2 = INR 50,000
E[Net]     = -INR 10,000/month   ← NEGATIVE!

The expected monthly cash flow is negative, so the project never achieves payback under λ=100. This is the key finding that contrasts with the paper's deterministic model (which uses 150 users).

Minimum viable demand rate: λ ≈ 125 (where E[Revenue] = E[OpEx]).

Results (both scenarios)

Metric	λ=100 (specified)	λ=150 (paper-implied)
E[Revenue]	INR 40,000	INR 60,000
E[Net]	-INR 10,000	+INR 10,000
Recovery Rate	0.00%	100.00%
P(payback ≤ 12mo)	0.00%	0.46%
P(payback ≤ 18mo)	0.00%	30.50%
P(payback ≤ 24mo)	0.00%	82.55%
Mean Payback	N/A (never)	20.81 months
Median Payback	N/A	20.0 months
95% CI	N/A	[14.0, 30.0] months

18 Figures Generated

1. Payback period histogram (λ=150)
2. Cumulative cash flow fan chart vs deterministic
3. Critical comparison: λ=100 vs λ=150
4. CDF of payback period
5. Revenue vs OpEx boxplots (both scenarios)
6. Cash flow distribution at month 12
7. Input distribution validation
8. Sensitivity: payback vs demand rate λ
9. Sensitivity: payback vs mean price
10. Payback heatmap (λ × Price)
11. 50 sample cash flow paths
12. ROI distribution at month 24
13. Probability of profit over time
14. Monthly net cash flow violin plots
15. Monte Carlo convergence
16. Recovery rate vs demand
17. Viability threshold analysis
18. Sensitivity: payback vs OpEx

---

Extension 2: STP-based User Segmentation & Churn Model

Segment Definitions

Segment	Description	N_n	N_e	α (churn)	P (INR)
S1	UG Course Projects	100	30	0.40	300
S2	PG/Research	30	15	0.10	500
S3	Lab/Institutional	20	5	0.05	400

Revenue Equation

R = Σ_k [ N_n,k · P_k + N_e,k · P_k · (1 − α_k) ]

Results

Segment	Revenue (INR)	Share
S1: UG Course Projects	35,400	52.8%
S2: PG/Research	21,750	32.4%
S3: Lab/Institutional	9,900	14.8%
Total	67,050	100%

• Monthly Profit: INR 17,050
• Simple Payback: 11.7 months
• Overall Churn Rate: 27.5%
• CLV: S1=INR 750, S2=INR 5,000, S3=INR 8,000

16 Figures Generated

1. Revenue breakdown by segment (stacked bar)
2. Revenue & user share (pie charts)
3. Sensitivity: total revenue vs α_S1 (primary deliverable)
4. Churn impact waterfall chart
5. User flow (new, retained, churned)
6. Customer lifetime value by segment
7. Revenue heatmap (α_S1 × α_S2)
8. 24-month revenue projection with dynamic churn
9. User base evolution over 24 months
10. Revenue per user efficiency
11. Original aggregate vs STP model comparison
12. Tornado sensitivity chart (±20%)
13. Payback sensitivity per segment churn
14. Segment profile radar chart
15. Break-even analysis per segment
16. Investment recovery timeline

---

Combined Analysis — Key Findings

Metric	Deterministic (Paper)	MC λ=100	MC λ=150	STP Segmented
Monthly Revenue	INR 60,000	E=INR 40,000	E=INR 60,000	INR 67,050
Monthly Profit	INR 10,000	E=-INR 10,000	E=INR 10,000	INR 17,050
Payback	12 months	∞ (never)	~21 months	~11.7 months
P(payback ≤ 12mo)	100% (assumed)	0.0%	~0.5%	N/A (deterministic)

Key Insights

1. Demand Rate is Critical: With λ=100, the project is financially unviable. Min viable λ≈125.
2. Stochastic Reality vs Deterministic Optimism: Even at λ=150, mean payback is ~21 months (not 12), and P(payback ≤ 12mo) is only ~0.5%.
3. STP Segmentation Reveals Higher Revenue: INR 67,050/mo vs INR 60,000 aggregate (+11.75%).
4. S1 Churn is the Dominant Risk Factor: 40% churn but 52.8% of revenue. Tornado chart confirms S1 parameters have the largest impact.
5. Recommendation: Target minimum λ=150, reduce S1 churn below 30%, develop S2/S3 segments.

---

File Structure

├── README.md                        # This file
├── RESULTS.md                       # Pre-computed results summary
├── requirements.txt                 # Python dependencies
├── run_all.py                       # Generate all outputs
├── pcb_financial_analysis.ipynb     # Comprehensive Jupyter notebook
├── src/
│   ├── extension1_monte_carlo.py    # MC simulation (18 figures)
│   └── extension2_stp_churn.py      # STP analysis (16 figures)
├── figures/                         # 34 PNG visualizations (generated by run_all.py)
├── data/                            # JSON data files (generated + pre-computed)
│   ├── ext2_sensitivity_alpha_s1.json
│   └── ext2_tornado.json
└── results/                         # Summary results (pre-computed)
    ├── ext1_monte_carlo_results.json
    └── ext2_stp_results.json

Author

VarunGowda