Reorder point

Reorder point = (average daily sales × lead time in days) + safety stock

A product selling 6 units a day from a vendor who takes 21 days, held with 40 units of safety stock, has a reorder point of (6 × 21) + 40 = 166 units. When stock touches 166, the order goes.

The first term is what you will sell while you wait. The second is your allowance for the two things that will not behave: demand running hotter than average, and the delivery arriving late.

The multiplication is arithmetic. Safety stock is a decision about how much you are willing to pay to avoid an empty shelf, and it is where the real thinking happens.

Two things drive it up. Demand that swings week to week, because an average that is right on average is wrong most weeks. And unreliable lead times, which matter more than long ones: a vendor who reliably takes six weeks is easier to plan than one who takes two or five unpredictably.

Two things drive it down. A product you can reorder quickly, because the exposure window is short. And one where a stockout costs little, because the customer will take a neighboring size or color instead.

Where you want a number rather than a judgement, the conventional calculation is (maximum daily sales × maximum lead time) − (average daily sales × average lead time). It prices the gap between a bad week with a slow delivery and an ordinary one, which is exactly the exposure safety stock exists to cover, and it needs your own observed maximums rather than an estimate.

That last consideration is worth asking rather than assuming, and it is why a reorder point set purely per product can mislead. Whether a stockout in one color of a family that carries four costs as much as one with no substitute on the shelf depends entirely on your customers: in some ranges a shopper takes the next color happily, and in others the color was the reason they came in. Only your own sales history answers it.

The formula assumes demand continues at roughly the rate you measured. For a core line that holds all year, that is reasonable. For anything seasonal it is not, and the failure is expensive in both directions.

Late in a season the reorder point still says to order, because stock has fallen and the trailing average is still high, and the delivery lands as the range stops selling. Early in a season it says to wait, because the trailing average is low, and by the time the rate rises the lead time has already cost you the weeks that mattered.

The correction is the same one weeks of supply needs: judge against the same weeks last year rather than the weeks just gone, and check the reorder against how many selling weeks remain rather than against the stock level alone.