Safety stock
Safety stock is extra inventory maintained as a buffer against variability in demand and supply. This buffer protects against stockouts when actual demand exceeds forecasts, when suppliers deliver late, or when quality issues reduce usable supply. Determining appropriate safety stock levels balances the cost of extra inventory against the cost and risk of stockouts.
Examples
Demand variability buffer: Historical demand for a component averages 1,000 units per week with standard deviation of 200 units. Safety stock of 400 units (2 standard deviations) provides 97.7% confidence of meeting demand during lead time.
Supply variability buffer: A supplier with 4-week lead time delivers anywhere from 3.5 to 5 weeks. Safety stock covers the additional half-week of demand during extended lead times, preventing stockouts when deliveries are late.
Combined buffer: Safety stock calculations consider both demand variability and supply variability together, as both contribute to stockout risk during the replenishment lead time.
Definition
Safety stock calculation methods range from simple rules of thumb (two weeks' supply) to statistical formulas based on demand variability, lead time variability, and desired service level. More sophisticated approaches provide better results but require better data and analytical capability.
Key inputs to safety stock calculation include demand variability (how much demand fluctuates), lead time and lead time variability (replenishment timing and reliability), and target service level (acceptable stockout probability). Each factor significantly influences required buffer levels.
Safety stock represents tied-up capital and storage space, so minimizing it improves efficiency. Reducing demand forecast error, shortening and stabilizing lead times, and improving supplier reliability all enable safety stock reduction without sacrificing service levels.
Different items warrant different safety stock approaches. Critical items justify higher service levels and more safety stock. Commodity items with many alternatives need less buffer. Obsolescence-prone items should minimize stock despite variability.