DDMRP: Stock the Part, Not the Product

You build twelve pump models. Demand for each one jumps around: 30 units one month, 5 the next. But all twelve use the same cast iron housing, and that casting takes ten days to arrive. If you could keep stock in only one place, where would you put it?

Not in finished pumps: you'd need twelve piles, and one of them would always be the wrong one. Put it in the casting. One pile protects all twelve models. That's the core idea of DDMRP.

DDMRP (demand driven MRP) is a planning method that places inventory buffers at a few chosen points in the product structure, sizes them from actual average usage, and reorders based on a "net flow" position instead of a forecast. Between the buffers, ordinary MRP logic still does the work. Carol Ptak and Chad Smith first described it in the 2011 edition of Orlicky's Material Requirements Planning, the book that made MRP a standard in the first place.

Why DDMRP decouples demand

Classic MRP passes every change straight down the bill of materials. A forecast change for one pump model ripples into castings, seals, bolts and paint, and purchase suggestions flip from week to week. A buffer at the casting stops that wave. Above it, assembly responds to real orders. Below it, the foundry gets a steady signal.

It also shortens the lead time the customer sees. With castings on the shelf, the customer waits for machining and assembly, not for the foundry. That's the decoupling point from make to order vs make to stock, turned into a planning rule.

A DDMRP buffer, calculated

The casting has an average daily usage (ADU) of 20 and a decoupled lead time (DLT) of 10 days. We'll use a lead time factor of 0.5 and a variability factor of 0.5, both typical "medium" settings.

• Yellow zone (covers demand during the lead time): 20 × 10 \= 200
• Red zone (the safety part): red base 20 × 10 × 0.5 \= 100, plus red safety 100 × 0.5 \= 50, so 150
• Green zone (sets order size and frequency): 20 × 10 × 0.5 \= 100 (or the minimum order quantity, if that's larger)

Stacked up: top of red \= 150, top of yellow \= 350, top of green \= 450.

Now today's position. You have 180 castings on hand and 120 on order, and 40 are needed for pump orders due today.

Net flow \= 180 + 120 − 40 \= 260

260 is below the top of yellow (350), so you reorder. The order fills the buffer to the top of green: 450 − 260 \= 190 castings.

Notice what the signal ignored: next quarter's forecast. It used stock you have, supply that's already coming and demand that is real today.

The buffer also moves with usage. If ADU climbs to 30, the same formulas give red 225, yellow 300, green 150 and a top of green of 675. Nobody has to remember to edit a number.

The fuel gauge analogy

Net flow is how a sensible driver reads the fuel gauge before a long trip: what's in the tank, plus the full can in the trunk, minus the 200 miles you must drive before the next station. Look at the tank alone and you stop too early, or too late.

DDMRP vs a reorder point

A classic reorder point for the same casting might be ADU × lead time + safety stock \= 200 + 50 \= 250, with a fixed order of 200. Three differences show up right away:

1. What gets compared. Many reorder-point setups look at on-hand stock. On hand is 180, below 250, so they order 200 more, even though 120 are already on the way. Net flow counts the 120 and the 40 due today.
2. Order size. The reorder point always orders 200. DDMRP orders the gap to the top of green: 190 today, a different number tomorrow.
3. Upkeep. A reorder point set when ADU was 20 stays at 250 when usage doubles, until someone notices. DDMRP buffers recalculate from ADU.

Neither method survives bad data. A wrong lead time or an unbooked receipt breaks both.

Myth: DDMRP replaces MRP

It doesn't. DDMRP keeps the bill of materials explosion; it just stops the explosion at buffered parts. Between buffers you still calculate dependent demand exactly as MRP does. What changes is that a few strategic parts are protected by buffers instead of being planned purely from forecasts.

The second myth is that DDMRP means more inventory. Often the total goes down, because you swap finished stock of twelve models for one stock of castings. But placement matters. A buffer on every part is just a bigger warehouse.

How to start

Pick one shared, long-lead part with bumpy demand, like the casting. Check its real lead time and usage over the last three months, not the values in the item master. Size the buffer, run it next to your current method for a few weeks, and compare stockouts and stock value. Accurate stock by location matters more than the formulas; see inventory management for manufacturers.

Where factory.online stands

factory.online gives you the building blocks for buffer-style planning, such as stock by location, open supply, demand linked to orders and MRP shortages, but it is not a certified, full DDMRP implementation. For the casting, the planner sees the 180 on hand, the 120 on open purchase orders and the 40 required by today's pump orders in one place, which are exactly the inputs to the net flow calculation.

If shared parts keep causing your shortages, book a demo and bring the one you'd buffer first.