Lead Time Variability

The most common measure of lead time variability, calculated as the square root of the variance in historical delivery durations. A higher standard deviation indicates greater unpredictability in supplier performance.

• Formula: σ = √[Σ(xi - μ)² / (n-1)]
• Application: Directly used in safety stock calculations (σLT × Z-score × average demand)
• Interpretation: A supplier with μ=10 days and σ=3 days will deliver within 4-16 days ~95% of the time
• Limitation: Assumes a normal distribution, which lead time data often violates

A normalized measure of dispersion that expresses standard deviation as a percentage of the mean, enabling apples-to-apples comparison of variability across suppliers with vastly different average lead times.

• Formula: CV = (σ / μ) × 100%
• Benchmarking: A CV below 20% indicates a reliable supplier; above 50% signals severe unpredictability
• Example: Supplier A (μ=5 days, σ=2) has CV=40%; Supplier B (μ=30 days, σ=6) has CV=20%—B is actually more consistent despite higher absolute deviation

Lead time distributions are rarely symmetric. They typically exhibit positive skewness—most deliveries cluster around a shorter duration, but a long tail of extreme delays pulls the mean above the median.

• Right-skewed distribution: Mean > Median > Mode
• Fat-tail risk: Extreme delays occur more frequently than a normal distribution would predict
• Modeling implication: Using only mean and standard deviation underestimates the probability of severe stockouts
• Mitigation: Use quantile regression or survival analysis to model the tail behavior directly

Lead time variability originates from multiple, often compounding sources across the supply chain. Identifying the dominant source is essential for targeted improvement initiatives.

• Supplier processing variability: Inconsistent order acknowledgment and production scheduling
• Manufacturing variability: Machine downtime, quality rejections, batch size changes
• Transit time variability: Port congestion, weather disruptions, carrier reliability fluctuations
• Customs and documentation variability: Border clearance delays, incorrect paperwork
• Demand-side amplification: Bullwhip effect where order batching magnifies upstream variability

A critical distinction in supply chain analytics: variability is the measurable, historical dispersion of outcomes, while uncertainty represents the unknown future deviation that cannot be quantified from past data alone.

• Aleatoric variability: Inherent randomness that can be characterized statistically but not eliminated
• Epistemic uncertainty: Knowledge gaps due to limited data, new suppliers, or unprecedented events
• Practical impact: Variability informs safety stock; uncertainty requires scenario planning and buffer capacity
• Black swan events: The COVID-19 pandemic exposed the limits of historical variability as a predictor of future risk

Lead time variability has a multiplier effect on required safety stock. The classical safety stock formula demonstrates that variability in lead time is often more damaging than variability in demand.

• Formula: SS = Z × √(μLT × σD² + μD² × σLT²)
• Key insight: Lead time variance (σLT²) is multiplied by the square of average demand (μD²), making it disproportionately impactful for high-volume items
• Example: Reducing σLT from 5 to 3 days for a product with μD=100 units/day reduces safety stock by approximately 40%
• Strategic implication: Supplier development programs targeting consistency often yield higher ROI than demand forecasting improvements