Inventory Optimization: Safety Stock and EOQ

Inventory optimization is the practice of holding just enough stock to protect service levels without tying up excess capital or warehouse space. Two classic tools anchor the discipline: the Economic Order Quantity (EOQ), which sizes replenishment orders, and safety stock, which buffers against demand and lead-time uncertainty.

EOQ answers a simple but consequential question: how much should be ordered at a time to minimize total inventory cost? Ordering too frequently in small batches drives up ordering and transportation cost per unit; ordering rarely in large batches drives up carrying cost from holding excess stock. The classic EOQ formula balances these two opposing cost curves to find the order quantity where total cost is lowest.

Safety stock exists to absorb the two sources of uncertainty that any deterministic order-quantity formula ignores: demand can spike above forecast, and supplier lead time can stretch beyond the average. The size of the buffer needed depends on the variability of both, and on the target service level a company is willing to pay for — a 99% in-stock target requires meaningfully more buffer stock than a 95% target, since the cost of protecting against rare extreme events grows quickly.

• Demand variability — how much daily or weekly demand fluctuates around its average
• Lead time variability — how reliably a supplier delivers within the quoted lead time
• Target service level — the acceptable probability of stocking out before replenishment arrives

Not every SKU deserves the same level of inventory scrutiny. ABC analysis ranks items by their contribution to revenue or volume, typically finding that a small share of SKUs (the "A" items) drives the majority of value. This segmentation lets planners apply tighter forecasting, more frequent review, and higher service-level targets to A items, while managing C items with simpler, lower-touch rules.

• A items — high value or velocity, reviewed frequently, tightly forecasted
• B items — moderate value, standard review cadence
• C items — low value or velocity, simple reorder-point rules, larger buffer tolerance

Pure EOQ and safety stock formulas assume stable, independent demand — an assumption that breaks down for seasonal products, promotional spikes, or new product launches. Mature inventory planning layers statistical forecasting, multi-echelon inventory optimization (which considers stock positioned across an entire network rather than one location at a time), and periodic manual override for known future events on top of the classic formulas rather than replacing them entirely.