SUPSupply & productionStop 4 of 12
MRP: from BOM to orders
Exploding the master schedule through the bill of materials into the components and raw materials to buy or make, and when.
Read atChange any time
Once the master schedule says “make 30,000 bottles of shampoo in week 6”, every component follows from arithmetic, not forecasting. The bill of materials (BOM) is the recipe: one bottle of shampoo needs one bottle, one cap, one label and 500 ml of liquid. MRP (material requirements planning) multiplies through:
30,000 shampoos → 30,000 bottles, 30,000 caps, 30,000 labels, 15,000 litres of liquid
Then, for every component, MRP runs the same steps:
- Gross requirement: what the plan needs.
- Net requirement: subtract stock on hand and orders already on the way. With 8,000 caps in stock and 5,000 arriving, it needs 17,000 more.
- Lot sizing: round to the supplier’s pack size or an economic batch.
- Lead-time offset: if caps take 2 weeks to arrive, the order must go out in week 4.
The output is a set of planned orders: purchase requisitions for bought parts and production orders for parts made in-house. In ERP systems such as SAP, this runs automatically, often every night.
The key idea behind MRP is dependent demand. Components don’t need their own forecast, because their demand depends on the finished goods plan. That is why MRP works best for manufacturing with clear BOMs, and why the plan is only as good as its inputs: wrong lead times, inaccurate stock records or outdated BOMs flow straight into wrong orders.
Words to use
- Bill of materials (BOM)
- The list of components and quantities needed to make one unit of a product.
- Dependent demand
- Demand calculated from a higher-level plan rather than forecast on its own.
- Lead-time offset
- Placing an order earlier by the lead time so it arrives when needed.
The MRP record
For each component, MRP lays out the weeks and works through them one at a time: gross requirements from the parents, scheduled receipts, projected balance, net requirements, planned receipts (lot-sized) and planned order releases (offset by lead time).
The MPS makes 20,000 shampoos in week 4 and 30,000 in week 6, one cap each. Caps: 8,000 in stock, 5,000 arriving in week 3, ordered in multiples of 10,000, lead time 2 weeks.
| Week | 1 | 2 | 3 | 4 | 5 | 6 |
|---|---|---|---|---|---|---|
| Gross requirements | 20,000 | 30,000 | ||||
| Scheduled receipts | 5,000 | |||||
| Projected balance (start 8,000) | 8,000 | 8,000 | 13,000 | 3,000 | 3,000 | 3,000 |
| Net requirements | 7,000 | 27,000 | ||||
| Planned receipts | 10,000 | 30,000 | ||||
| Planned order releases | 10,000 | 30,000 |
Week 4 needs 20,000 but only 13,000 will be there, so 7,000 is missing, rounded up to 10,000 and ordered in week 2. Week 6 is short by 27,000, rounded to 30,000 and ordered in week 4.
Multi-level explosion
MRP works down the bill of materials level by level. The planned orders for a parent become gross requirements for its components, which may themselves be made from other components. The shampoo liquid, for example, might be made in-house a week before filling, so its planned orders create requirements for surfactant, fragrance and water.
- Low-level codes: a component used at several levels of different bills is netted only once, at its lowest level, after all its parents’ requirements are known.
- Pegging: each requirement is linked back to the parent orders that caused it, so a planner can see which customer order is affected by a late part.
Messages planners act on
MRP also checks existing orders against the new plan:
- Expedite (reschedule in): an order arrives after it is needed.
- De-expedite (reschedule out): an order arrives earlier than needed.
- Cancel: an order is no longer needed.
- Release: a planned order has reached its release date and should become a real purchase or production order.
Infinite capacity and fixed lead times
Classic MRP assumes every work centre has unlimited capacity and every order takes its planned lead time. Neither is true. MRP II added capacity requirements planning and feedback to close the loop, and advanced planning systems schedule against finite capacity, but the core MRP calculation still offsets by fixed lead times.
The lead-time trap
Planned lead times are mostly queue time, not working time. When parts arrive late, planners lengthen the planned lead times to “be safe”. Orders are then released earlier, more work enters the shop, queues grow, and actual lead times grow too, which leads to even longer planned lead times. Little’s law explains it: with the same throughput, more work in process means longer flow times. The fix is the opposite instinct: control what is released and shorten queues.
Nervousness
Small changes in the master schedule can cause large changes in component plans, especially with lot-sizing rules such as minimum order quantities and period lot sizes, which can flip an order from one week to another. Suppliers see orders appear, disappear and move. Remedies: firm planned orders in the near term, time fences, lot-for-lot sizing on low-cost items, and ignoring small reschedule messages.
Data is everything
MRP is arithmetic, so its output is exactly as good as its inputs:
- Inventory accuracy: if the system thinks 8,000 caps are in stock and only 5,000 are, every plan built on that is wrong. Cycle counting keeps records honest.
- Bills of materials: wrong quantities or outdated components generate wrong orders.
- Lead times, lot sizes and safety stocks: parameters set years ago and never reviewed are a classic source of bad plans.
Demand-driven MRP
DDMRP (Ptak and Smith) keeps MRP’s bill-of-materials logic but places strategic stock buffers at chosen points in the product structure. Buffers are sized from average usage and lead time, and replenished according to how far their stock has fallen, rather than from a forecast exploded through every level. This decouples lead times and dampens nervousness. Classic MRP remains the backbone of most ERP systems; S/4HANA’s MRP Live runs the same logic much faster in memory.
Interview kit
Practise it out loud
Answer each question yourself before opening the model answer.
MRP takes the master production schedule, explodes it through the bill of materials into gross component requirements, nets them against stock and open orders, applies lot sizes, and offsets by lead time to create planned purchase and production orders. It relies on dependent demand: components are calculated, not forecast. Its quality depends on accurate BOMs, lead times and stock records.
Why don't components need their own forecast?
Because their demand depends on the finished-goods plan. If you know you’re building 30,000 bottles of shampoo, you know you need 30,000 caps. Forecasting them separately would add error. Only items with independent demand, like spare parts sold directly, need their own forecast.
MRP keeps suggesting orders for parts you already have. What would you check?
Stock accuracy (are the counts and storage locations right?), open purchase orders that were received but not posted, BOM quantities, lead times and lot-size settings in the material master, and whether safety stock is set too high. Most MRP problems trace back to master data.
Connected stops
Where this idea feeds in, or gets its inputs from.
Sources and further reading
- Silver, E. A., Pyke, D. F. & Thomas, D. J. (2017). Inventory and Production Management in Supply Chains (4th ed.). CRC Press. Chapter on material requirements planning.
- Orlicky, J. (1975). Material Requirements Planning. McGraw-Hill.
- Ptak, C., & Smith, C. (2016). Demand Driven Material Requirements Planning (DDMRP). Industrial Press.