The decoupling point is the place in the flow where the customer order stops being a forecast and becomes a commitment. Upstream of it, everything is decided on anticipation and carried as stock. Downstream, everything is triggered by real demand. Its position determines simultaneously the lead time you can promise and the capital you must tie up, and it is the only parameter in the domain that settles both questions at once.
Why it matters
A supply chain runs on two incompatible logics. Upstream, you anticipate: you launch, build and stock on the basis of a forecast, with all the uncertainty that entails. Downstream, you execute: every movement answers a real order, and uncertainty disappears. The decoupling point is the exact boundary between these regimes, and every chain has one, whether it chose it or not.
Its position governs everything else. Placed far downstream, at the distributor, it lets you serve the customer instantly but forces the whole chain to work on forecast, hence to carry the matching stock, including that of variants that will not sell. Placed far upstream, it cuts stock to almost nothing but makes the customer wait for the full build.
The signature shows the two resulting curves, and their opposition is perfectly legible. Stock falls as the point moves upstream, customer lead time rises, and no position improves both at once. It is a pure trade-off, and that is why it belongs to leadership rather than to planning.
What is striking in practice is that this position is rarely the result of a decision. It is inherited from the company’s industrial history, from product design, or from a commercial promise made long ago. Many organisations discover, on mapping it, that they are paying for considerable stock to deliver a lead time their customers were not asking for.
The intermediate position deserves particular attention. Assemble to order, where you stock common sub-assemblies and combine them once the order arrives, is very often the best compromise, for a reason that goes beyond the simple trade-off. By stocking what is common rather than what is specific, it lets stock benefit from pooling across variants, exactly as postponement does.
The decoupling point then reads as the synthesis of the whole network cluster. Moving it upstream is pooling; keeping it downstream is favouring proximity. The pages on pooling, postponement and multi-echelon each describe one aspect of this same decision, seen from a different angle.
The expert lesson is to make lead time a negotiating variable rather than a given. Moving the point back requires lengthening the customer promise, which seems unacceptable until you check it segment by segment. On an often significant share of the portfolio, customers readily accept a few extra days in exchange for price or guaranteed availability. It is that negotiation, not a calculation, that unlocks the opportunity.
The mechanism
The usual positions form a continuous scale usually summarised into a few regimes. Delivery from stock places the point closest to the customer. Distribution from stock steps back by concentrating finished goods. Assemble to order stocks sub-assemblies. Make to order keeps only components. Engineer to order commits nothing before the customer agrees.
The rule governing the choice is simple to state. The lead time you can promise is that of the stretch downstream of the point, and that alone. Everything upstream must be ready in advance, hence anticipated and financed. Moving the point therefore means choosing what share of total lead time you absorb yourself through stock, and what share you ask the customer to wait.
One important constraint limits the reachable positions: the point can only sit where the product is still common to several possible demands. This links the page directly to postponement. A product whose specificity appears very early in the process offers no interesting intermediate position, and the only way to create one is to redesign the product.
Finally, nothing requires a single position. One company can serve its fast, regular references from stock, assemble its mid-range variants to order, and make its long tail to order. This differentiation by segment, guided by the segmentation matrix, is almost always superior to one regime applied to the whole catalogue.
The traps
Most chains have a decoupling point nobody chose, fixed by industrial history or by an old commercial promise. Mapping it frequently reveals considerable stock financed for a lead time the market no longer demands.
Moving the point back means lengthening the promise, which seems impossible until the question is asked segment by segment. On part of the portfolio, a few extra days are readily traded for price or a guarantee of availability.
Serving the whole catalogue on the same logic means over-financing slow references or under-serving fast ones. The decoupling point is differentiated by segment, exactly like service level and replenishment policy.
The rollout
Identify, family by family, the exact place where the customer order meets the flow. This map alone often reveals positions nobody had decided.
Compare the current promise with what customers genuinely ask for, by segment. The gap between the two is the room available.
Spot the steps where the product is still common to several demands. These are the only candidate positions, and they depend on design as much as on process.
Assign a distinct regime to fast, mid and slow references, building on the existing segmentation rather than applying one rule.
Make lead time part of the commercial offer, tradable against price or a guarantee, rather than a constraint imposed on the supply chain.
Neighboring concepts
From knowledge to action
That boundary sets both your lead-time promise and your tied-up capital, and it is almost always inherited. Our Inventory & distribution file maps it and differentiates it by segment.