An accurate demand forecast only tells you the expected value — it says nothing about how wrong you might be. This lab turns a forecast plus its error into a stocking policy: a safety stock buffer that absorbs demand spikes during the supplier's lead time, and a reorder point that triggers a replenishment order before the shelf runs dry.
σ_LT = σ · √(lead time), Safety Stock = z · σ_LT, ROP = (avg. demand × lead time) + Safety Stock.z (90% → 1.28, 99.9% → 3.09), which grows the safety stock and cuts stockout days at the cost of holding more inventory.Doubling forecast error roughly doubles the required safety stock, but reaching from a 95% to a 99.9% service level with the same error can need over twice as much buffer again — service level and forecast accuracy trade off against working capital tied up on the shelf.
A 3D inventory silo is fed by a scrolling demand forecast with a widening uncertainty cone; watch how forecast error, lead time and service-level targets translate into safety stock, reorder points, and stockouts.
Safety stock and the reorder point are computed live from the forecast's error, the supplier's lead time and the target service level, then rendered as bands and rings on the silo — showing why an accurate mean forecast is only half the inventory problem.
Set the average demand, forecast error, lead time and service level. Watch daily demand deplete the silo, trigger a replenishment order when stock crosses the reorder point, and a shipment fly in after the lead time elapses.
Because lead-time demand variance scales with the lead time itself, doubling a supplier's lead time can require over 40% more safety stock even if daily forecast error stays exactly the same.