Developed in collaboration with Teradyne’s Supply Base Management (SBM) group, this project optimizes the excess inventory approval process. By addressing demand volatility and long supplier lead times, we transitioned Teradyne from a subjective, manual oversight model to a data-driven, algorithmic decision framework.
Goal: Reduce bias, improve approval efficiency, and enhance capital utilization using Python-based quantitative modeling.
Teradyne faced recurring excess inventory and financial leakage due to:
- Demand Volatility: Rapid shifts in market requirements.
- Lead Time Complexity: Long and variable supplier lead times.
- Subjective Workflows: Approval systems relied on qualitative judgment rather than real-time data.
Financial Impact:
- Increased holding costs and reduced profitability.
- Significant capital tied up on the balance sheet.
- Lower operational efficiency in the supply chain.
We synthesized 18 months of historical data across three primary domains:
| Data Source | Scope | Key Attributes |
|---|---|---|
| Excess Inventory | 1,408 requests | Site, Part Number, PO Price, Qty Ordered |
| Demand Data | 17 Files | Planned Demand, Parent Site, Due Date |
| Parts Data | 17 Files | Cumulative Lead Time (weeks), Site |
The pipeline cleanses and merges disparate datasets (Demand, Inventory, and Lead Times) into structured snapshots for specific fiscal periods (2022 H2, 2023 H1, and 2023 H2).
The framework utilizes Lead-Time-Adjusted Demand to determine if a purchase is truly necessary.
-
Adjusted Date Calculation:
$$\text{Adjusted Date} = \text{Requested Date} + (\text{Cumulative Lead Time} \times 2 \times 7)$$ -
Lead Time Segmentation: * Short-term (< 182 days)
- Medium-term (182–365 days)
- Long-term (> 365 days)
-
Suggested Need Formula:
$$\text{Suggested Need} = \text{Aggregated Demand} - (\text{On Hand} + \text{Open Orders})$$ -
Potential Excess Quantification:
$$\text{Total Excess Cost} = (\text{Suggested Need} - \text{Quantity Ordered}) \times \text{PO Price}$$
The implementation of the algorithmic framework led to immediate improvements in financial oversight:
- ⚡ Enhanced Oversight: 8% increase in requests escalated for high-level management review, ensuring better control over high-spend approvals.
- 🔍 Risk Visibility: Identified hidden excess risk across multiple periods:
- 2022 H2: 7% visibility
- 2023 H1: 43% visibility
- 2023 H2: 55% visibility
- 💰 Capital Efficiency: Significant reduction in projected excess spend and optimized working capital.
- Language: Python
- Libraries: Pandas (Data Manipulation), NumPy (Numerical Logic), Matplotlib/Seaborn (Visualization)
- Environment: Jupyter Notebook / Google Colab
- ML Integration: Implementing machine learning for predictive demand forecasting.
- Real-time Monitoring: Dashboard for real-time stockout and excess monitoring.
- MOQ Optimization: Incorporating Minimum Order Quantities into the "Suggested Need" logic.
W.P. Carey School of Business – MSBA Applied Project
- Term: Spring 2024
- Team: 441
This project proves that transitioning from subjective judgment to objective, lead-time-adjusted modeling allows supply chain groups to proactively mitigate financial risk and improve overall resilience.