DDMRP reorganises replenishment around two ideas. First, positioning strategic decoupling points in the chain, where you accept holding stock in order to break the propagation of variability. Second, protecting each of those points with a three-zone buffer and steering not on physical stock but on the net flow position, that is, on-hand plus on-order minus qualified demand.
Why it matters
Classic requirements planning rests on a chain of dependencies: a finished-goods forecast is exploded into components, which are exploded in turn, each level inheriting the assumptions of the one above. The mechanism is rigorous but fragile, because it also transmits errors. A minor correction at the top propagates upstream while amplifying, producing what the literature calls system nervousness: orders that change constantly while real demand has barely moved.
Pull flow answers that problem, but it demands a regularity many chains do not have, and it sits poorly with deep bills of material and long lead times. DDMRP falls between the two. It keeps the structure of the bill of material, but refuses to let the calculation run through the entire chain end to end.
Its first idea is decoupling. At chosen points, you decide to hold stock, not for comfort but to break the transmission of variability. Upstream of that point you no longer react to downstream jolts; downstream you no longer suffer upstream inertia. The decoupling point also shortens the lead time the customer perceives, since the promise now depends only on the stretch that follows it.
The second idea is a change of steering indicator. You look neither at physical stock, which ignores what is already on order, nor at the forecast alone, which ignores what is genuinely committed. You steer on the net flow position, which adds on-hand and on-order and subtracts qualified demand, that is, firm orders and identified spikes. It is the only indicator that reflects the real situation at the moment of deciding.
The signature above shows the whole articulation. On the left, the buffer profile and its three zones, each with a distinct role: green sets the frequency and size of orders, yellow covers consumption during the decoupled lead time, red absorbs uncertainty. On the right, the same buffer crossed over time by two lines, the net flow position that triggers orders and the on-hand stock that is actually consumed.
Sliding variability reveals the most instructive point of the method. Only the red zone dilates, but since it serves as a base, the whole profile rises. Watch the day-32 spike: with a thin red zone, on-hand grazes a stockout; with a thick one, the same spike passes unnoticed. Uncertainty is not paid for in degraded service, it is paid for in buffer height, and that height is explicit, readable, arguable in a steering committee rather than buried in a parameter.
The expert lesson concerns the order of priorities. DDMRP performance comes first from the positioning of decoupling points, and only then from buffer sizing. A chain with poorly placed points will not be saved by well-computed zones. Conversely, well-chosen points deliver most of the gain, even with approximate buffers that can be refined later.
The mechanism
The method unfolds in three stages. First you position the decoupling points, looking for the places where holding stock produces the most effect: where lead times lengthen, where variability amplifies, where a component is common to many finished products, and at the boundary of what the customer is willing to wait. This choice is structural and reaches far beyond parameter setting.
Then you size each point’s buffer. The yellow zone represents expected consumption during the decoupled lead time, the core of the coverage. The green zone determines the frequency and size of orders, accounting for minimum quantities and the desired order cycle. The red zone, finally, is protection against uncertainty: a base proportional to the lead time, plus a supplement that grows with observed variability.
Daily steering then runs on the net flow position. While it stays in the green, nothing happens. As soon as it enters the yellow, an order is proposed to bring it back to the top of the green. If it drops into the red, the situation calls for attention, because the buffer is being eaten into. The colour code replaces the reading of dozens of parameters and makes the system legible to teams who are not specialists.
A word on what the method really is, often misunderstood. DDMRP invents no new mathematics: the zones reduce to notions already met here, lead-time coverage, lot size, safety stock. Its contribution lies elsewhere, in the articulation. It forces an explicit choice of where to decouple, it gives one single decision indicator, and it makes the whole thing visual. This is not a theoretical breakthrough, it is a discipline of execution, and that is precisely what many organisations lack.
The traps
Placing a buffer at every level of the bill of material means financing stock across the whole chain while decoupling nothing at all, since nothing flows freely any more. Decoupling draws its value from being rare: a few well-chosen points absorb most of the variability, a multitude of points merely ties up capital.
The demand-driven label suggests forecasting becomes useless. It does not: the average usage that sizes the zones is a forecast, and the spikes identified in qualified demand are another. The method reduces dependence on detailed long-range forecasting, it does not eliminate it.
Buffers are designed to breathe with consumption, lead times and seasonality. Many deployments compute the zones once, then freeze them. The device then loses its main quality, adaptation, and reverts to one more static configuration, with a coat of colour on top.
The rollout
Record, along the flows, where lead times lengthen and where variability amplifies. This map is the raw material for choosing decoupling points, and it has value even if the deployment stops there.
Favour components common to many products, customer lead-time boundaries and rhythm breaks between upstream and downstream. Every point retained must be justified by what it decouples, not by the comfort it brings.
Handle yellow through decoupled lead-time coverage, green through lot and frequency logic, red through measured variability. Three distinct roles call for three distinct calculations, not a uniformly applied multiple.
Give planners a single indicator combining on-hand, on-order and qualified demand. Without it, teams will keep deciding on physical stock and the method will stay cosmetic.
Automate periodic recalculation of the zones and plan adjustments for seasons and launches. A living buffer is what separates a successful deployment from a repainted configuration.
Neighboring concepts
From knowledge to action
Performance comes first from positioning, only then from zone sizing. Our Inventory & distribution file maps your flows and places the points where they genuinely break variability.