起源故事 · 库存越多越缺货的怪圈
Origin Story · The "More Stock, More Stockouts" Paradox
Goldratt 在《目标》之后把约束理论推向供应链,发现一个反常识:很多企业仓库爆满,门店却天天缺货。
根子在于「各级照预测往下推」——每一层都怕断货、加安全库存、再放大上一层的预测误差,于是货全囤在了「错误的地点」。
他的解法干脆利落:别让货往下沉,把大部分库存留在最靠上的中央仓,下游只留刚好够卖的小量,
靠实际销售触发的频繁补充把货拉下去。库存待在中央,等看清哪儿真要货了再发——聚合的需求更平稳,决策推迟到信息更全的时候。
After The Goal, Eliyahu Goldratt turned the Theory of Constraints toward the supply chain and surfaced a counter-intuitive truth: many companies have warehouses bursting at the seams while their stores run dry daily.
The root cause is the tiered forecast-and-push model — every level fears a stockout, layers on safety stock, and amplifies the forecast error of the level above. Inventory ends up sitting in the wrong location.
His fix is brutally simple: stop pushing stock downstream. Hold most inventory at the upstream central warehouse, keep only the bare minimum at each downstream point, and let actual sales trigger frequent small replenishments that pull stock down on demand.
Inventory stays central and only moves once you can see where it's truly needed — aggregated demand is steadier, and decisions are deferred until you have better information.
1 配销网络:货囤在哪、怎么流 Distribution Network: Where Stock Sits, How It Flows
预测推货Forecast Push2 两个本该「此消彼长」的指标,一起降了 Two Metrics That Usually Trade Off — Falling Together
常识里「降库存」和「降缺货」是跷跷板。但 TOC 拉动靠需求聚合 + 频繁补充,把两根柱子一起压下来——这正是它最反直觉、也最值钱的地方。 Conventional wisdom treats "cut inventory" and "cut stockouts" as a seesaw. TOC pull replenishment uses demand aggregation plus frequent replenishment to push both bars down at the same time — the single most counter-intuitive and most valuable property of the model.
3 现实里的 TOC 配销 TOC Distribution in the Wild
连锁零售:门店只留小量、总仓集中持有,每日按 POS 销量补货——爆款不断货,长尾不压仓。
Chain retail: stores carry a thin buffer, the regional warehouse holds the bulk, and daily POS sell-through drives replenishment — hits stay on shelf, long-tail SKUs don't choke the warehouse.
备品备件:分布式囤备件成本高且常缺。中央池 + 按消耗补,关键件可得性反而更高。
Spare parts: distributed spares are expensive and still often missing. A central pool plus consumption-pulled replenishment actually lifts availability of critical parts.
医药/快消分销:保质期敏感品类尤其怕呆滞。拉动补充让货「待在上游」,临期风险大降。
Pharma & FMCG distribution: shelf-life-sensitive SKUs hate dead stock. Pull replenishment keeps inventory upstream and slashes expiry risk at the edge.
电商前置仓:中心仓集中库存、前置仓小批高频补,兼顾时效与周转。
E-commerce forward warehouses: the central DC holds the depth, forward warehouses get small high-frequency replenishments — fast delivery without killing inventory turns.
一句话In One Line
为什么各点放得更少,反而更不缺货?两个机制:一是需求聚合——把货留在中央,
多个门店的随机波动相互抵消,总需求比单点平稳得多,同样服务水平所需的安全库存更小;
二是推迟决策——库存不急着往下沉,等真看清哪个门店在卖什么再发,少押错宝。
再加上按实际消耗、频繁小批补充,库存就从「赌预测」变成「跟着真实销售走」——
既不囤死,也不断货。
Why does holding less at each point produce fewer stockouts? Two mechanisms.
First, demand aggregation: keep stock central and the random swings across many stores cancel out — aggregated demand is far steadier than any single point, so the safety stock needed for the same service level is much smaller.
Second, delayed commitment: inventory doesn't rush downstream; you wait to see who is actually selling what, then ship — far fewer wrong bets.
Combine that with small, frequent, consumption-driven replenishments and inventory shifts from "betting the forecast" to "tracking real sell-through" — neither overstocked nor out of stock.
常见误用Common Mistakes
只把库存搬到中央,却没提高补货频率。拉动靠「小批高频补」,补得慢门店照样断货。
Moving stock to central without raising replenishment frequency. Pull only works with small, high-frequency replenishment — slow cadence and stores still run out.
仍按预测决定补多少。补货量要由「实际消耗」触发,不是猜下周卖多少。
Still sizing replenishments off a forecast. Replenishment quantity is triggered by actual consumption, not a guess at next week's sales.
缓冲设死不动态调整。用缓冲管理(红黄绿)随趋势调高调低各点持有量。
Hard-coding the buffer and never touching it. Use dynamic buffer management (red-yellow-green) to flex each location's holdings with the trend.