A tiny wobble in shop-floor demand becomes a wild swing at the factory. Each link in a supply chain, to be safe, orders a bit more than it sold and adds buffer — and those margins compound upstream, so the maker sees demand lurching between famine and flood when customers barely changed. It’s why supply chains overshoot and crash. Slide the lead time and watch the whip crack.
The bullwhip effect: order variance grows at each tier moving UPSTREAM from the customer, even when end demand is nearly steady. With replenishment lead time L and a moving-average forecast over p periods, the variance amplification per tier is at least 1 + 2L/p + 2L²/p² — caused by demand forecasting, order batching, price promotions and rationing/gaming. So a retailer’s modest demand swing becomes a distributor’s larger one and a factory’s violent one, driving alternating gluts and shortages. A fail-loud self-check throws unless the upstream variance exceeds the downstream (amplification > 1). ◆ real supply-chain theory, node-verified.
The Lee–Padmanabhan–Whang forecasting-driven amplification formula (one exact cause); real bullwhip compounds several causes and information-sharing can damp it — the variance-grows-upstream result is exact.