故事 · 高毛利产品凭什么排到后面
Story · How a high-margin product loses its slot
一家工厂有两款产品都要过同一台关键机床。产品 P 单位毛利 60 元,产品 Q 单位毛利 45 元 ——
销售经理理所当然要求「多做 P,毕竟它更赚钱」。可财务一算账却发现,多做 Q 公司反而赚得更多。
为什么?因为 P 在那台机床上要占 30 分钟,Q 只占 15 分钟。
换算到每瓶颈分钟:P 是 60/30 = 2 元/分钟,Q 是 45/15 = 3 元/分钟。
机床的每一分钟用来做 Q,比做 P 多赚 50%。瓶颈是全厂产出的总闸门 ——
它每分钟产出多少,决定整个工厂赚多少。所以高德拉特的排产铁律是:
按 T / 单位占用约束时间 从高到低排,先喂饱瓶颈里「单位时间最赚钱」的产品,
而不是被「单位利润」这个表面数字误导。单位利润,是没有把瓶颈时间这个真实代价算进去的假象。
A factory makes two products that both pass through the same critical machine. Product P earns a 60-yuan unit margin, Product Q earns 45. The sales manager naturally pushes for P — "it makes more money per unit". Finance runs the math and discovers the opposite: making more Q actually leaves more money in the bank.
The reason: P ties up that machine for 30 minutes; Q clears it in 15.
Convert to throughput per bottleneck minute: P earns 60 / 30 = 2 yuan/min, Q earns 45 / 15 = 3 yuan/min.
Every minute the machine runs Q, it earns 50% more than running P. The bottleneck is the floodgate of the entire plant —
what it earns per minute determines what the whole factory earns. So Goldratt's scheduling rule is non-negotiable:
sort by T / constraint time per unit, high to low, and feed the bottleneck the products that pay the most per minute — never let "unit margin" trick you.
Unit margin is a mirage: it doesn't price in the real cost of the bottleneck time the product consumes.
1 两种排序,谁榨出更多总产出 Two rankings — which one squeezes out more total throughput
瓶颈序多赚 —Bottleneck wins by —2 每产品的 T/约束分钟 排名 Per-product T / constraint-minute ranking
每根柱是该产品每瓶颈分钟的有效产出(T ÷ 占用约束时间)。 越高越优先排产。注意:单位毛利最高的不一定排第一 —— 占用瓶颈久的产品会被「拉低单价」。 Each bar is that product's throughput per bottleneck minute (T ÷ constraint time). Higher means earlier in the schedule. Note: the highest unit margin doesn't always come first — products that hog the bottleneck get pulled down by their long run time.
3 现实里的产品组合决策 Product-mix decisions in practice
产品组合:订单挑选、产能分配时按 T/约束分钟排序,而非按毛利率或销售额。
Product mix: pick orders and allocate capacity by T per constraint-minute — not by margin rate or revenue.
瓶颈产能分配:瓶颈是全厂闸门,先把它喂饱单位时间最赚钱的产品。
Bottleneck allocation: the bottleneck is the plant's floodgate — feed it the products that earn most per minute first.
T/约束分钟:定价、接单的核心指标,把「时间代价」显式算进每个决策。
T per constraint-minute: the master metric for pricing and order acceptance — it makes the time cost explicit in every decision.
报价决策:占用瓶颈久的订单要么提价、要么外协,否则挤掉更赚钱的活。
Quoting decisions: orders that hog the bottleneck must either pay more or be outsourced — otherwise they crowd out more profitable work.
一句话
In one line
「单位利润」这个数字隐藏了一个致命缺陷:它没把占用瓶颈的时间代价算进去。
当所有产品抢同一个瓶颈时,真正稀缺的不是产品本身,而是瓶颈的每一分钟 ——
这才是全厂产出的总闸门。于是衡量标准必须从「这件产品赚多少」改成
「瓶颈每分钟用来做它能赚多少」(T / 约束分钟)。
按这个指标排序,瓶颈的产出被榨到最大,公司总有效产出随之最大。
一个反直觉但真实的结论是:薄利快产的产品常常打败高毛利慢产的产品,
因为前者在同样的瓶颈时间里能多做好几件。
记住:瓶颈之外的资源有多高效都不重要,只有瓶颈每分钟的产出,才决定整个系统赚多少钱。
"Unit margin" hides a fatal flaw: it never prices in the time cost of occupying the bottleneck.
When every product fights for the same bottleneck, what's scarce isn't the products — it's each minute of the bottleneck itself, the floodgate of the whole plant.
So the metric has to shift from "what does one unit earn" to
"what does one bottleneck minute earn when spent on this product" (T / constraint-minute).
Sort by that and the bottleneck delivers its maximum output — total throughput is maximized along with it.
A counter-intuitive but real result: thin-margin, fast-running products often beat high-margin, slow-running ones, because in the same bottleneck time they finish several more units.
Remember: how efficient the non-bottleneck resources are doesn't matter — only what the bottleneck earns per minute decides how much the entire system makes.
常见误用
Common mistakes
按单位利润 / 毛利率排产。瓶颈受限时按 T/约束分钟排,把时间代价算进去。
Scheduling by unit margin or margin rate. When the bottleneck binds, sort by T / constraint-minute — price the time in.
用单件平摊成本来核算。用有效产出(售价−料成本),固定开销不分摊进单件决策。
Using fully allocated unit cost. Use throughput (price − totally variable cost) — keep fixed overhead out of per-unit decisions.
瓶颈不饱和时也死守这套排序。若瓶颈有空、能全做,则不存在取舍,尽量多接单即可。
Forcing this ranking when the bottleneck isn't saturated. If the bottleneck has slack and can run everything, there's no trade-off — just take all the orders.