Par levels, reorder points, safety stock, and stock rotation

A handful of core concepts run most inventory systems, coffee bar or otherwise. Par level is the quantity that should be on hand or prepared for a given period — set too low and the bar runs out mid-service, set too high and product ages out or ties up cash unnecessarily. Reorder point is the stock level that triggers placing a new order, calculated so the order arrives before stock runs out — in its simplest form, average daily usage multiplied by lead time (how long a supplier takes to deliver), plus a safety-stock buffer.

Safety stock exists specifically to absorb the unpredictable parts of the system: demand spikes above average, or a supplier's lead time running longer than usual — without it, even a mathematically correct reorder point can still run out during genuine variability. Lead time itself deserves its own attention because it varies by supplier and item: a locally available syrup might arrive in a day, while a specialty single-origin bean shipped internationally might take weeks, and reorder points need to reflect each item's actual lead time rather than one blanket assumption. FIFO (First In, First Out) and FEFO (First Expired, First Out) are both stock-rotation rules, and though closely related they aren't identical: FIFO uses arrival order, which works well for shelf-stable goods where age and expiry track closely together, while FEFO uses actual expiry date, which matters more for perishables like dairy or fresh produce where two batches received on different days can still expire in a different order than they arrived — a fresher-looking delivery can sometimes have a shorter shelf life than older stock, which is exactly the case FEFO exists to catch.

Counting and variance review closes the loop: comparing what should be on hand (based on purchases and recorded usage) against what's physically counted surfaces waste, theft, or portioning drift that a purely theoretical par-and-reorder system can't detect on its own — which is why regular physical counts remain necessary even with good forecasting.