What Is Material Requirements Planning (MRP)?

MRP meaning: what material requirements planning actually does

In manufacturing, MRP means Material Requirements Planning—a production-planning and inventory-control method that calculates which materials, components, and subassemblies you need, in what quantities, and by which dates, so finished goods can ship without chronic stockouts or excess stock. It is built for dependent demand: the need for a bolt, PCB, or subassembly is derived from the end item it goes into, not from an independent reorder point on every part.

Outside manufacturing—especially in India and some other retail markets—MRP also means Maximum Retail Price printed on consumer packaging. That is a labeling rule, not a production system. This guide uses MRP only in the operations sense (Material Requirements Planning).

The demand split is the conceptual core. Independent demand is external: customer sales orders and forecasts for finished goods (and sometimes service parts sold as-is). You plan it with MPS, forecasts, and often safety stock. Dependent demand is calculated: if you will ship 100 four-legged tables, you need 400 legs—no separate forecast for legs. MRP monitors that parent–child relationship so component plans track finished-goods plans instead of floating free. Reorder-point systems treat every SKU as independent; they work for simple items and fail when multi-level BOMs couple demand across levels.

MRP is a push system: it works backward from a master production schedule (firm orders plus forecast for finished goods), explodes that demand through the bill of materials (BOM), nets on-hand stock and open supply, then offsets lead times so purchase and manufacturing orders release early enough to hit need dates. The three classic objectives run together: materials available for production and delivery; lowest practical inventory; and coordinated manufacturing, delivery, and purchasing. It fits discrete manufacturing with multi-level BOMs—assembly, fabrication, electronics, machinery—where one missing fastener can stop a line. Process and hybrid plants still use MRP-like netting for purchased ingredients and packaging when recipes and BOMs drive component need.

The three inputs and the outputs MRP produces

MRP is only as good as the data it consumes. Operations management literature treats three primary inputs as non-negotiable: without them the calculation cannot produce a trustworthy plan.

• →Master Production Schedule (MPS): quantities and timing of end items to produce, driven by firm orders and forecast.
• →Bill of Materials (BOM): hierarchical listing of every component and subassembly per end-item unit, with quantity-per at each level.
• →Inventory status records: on-hand, scheduled receipts already in flight, allocations/reservations, and lead times (buy and make).

How the MRP method works: four steps, level by level

Most modern explainers frame the MRP method as four process steps that map cleanly to what planners actually do each week: (1) identify demand and explode requirements through the BOM, (2) check inventory and open supply (netting), (3) schedule planned purchase and manufacturing orders with lead-time offsets, and (4) surface exceptions—shortages, late receipts, quantity changes—so people can replan before the line stops. That four-step story is the same math as classical time-phased netting; it just names the exception loop that makes MRP a living process rather than a one-shot report.

Under the hood, the engine processes the BOM level by level using low-level codes so every parent is fully netted before children are planned. That matters when a common part appears under multiple assemblies: demand is accumulated first, then netting runs once, so you do not under-order shared components.

At each level the engine still repeats explosion, netting, and time-phasing. BOM explosion multiplies parent demand by quantity-per (and, when configured, scrap or yield factors) to derive gross requirements. Netting subtracts available supply: net requirement is driven by gross demand minus projected available (on-hand plus scheduled receipts, less allocations), with safety stock treated as inventory that should not be consumed under normal planning. If projected available falls short of demand (and the safety-stock floor), the shortage becomes a net requirement and triggers a planned order. Time-phasing offsets each net requirement backward by lead time to set the planned order release date—the moment a PO or manufacturing order must be released so material arrives when needed.

Scrap and yield change the math in step one, not as a later “buffer guess.” If a process consistently scraps about 5% of a component, ordering for 100% yield means you will be short every run. BOM lines or routing parameters that carry scrap factors inflate gross requirements so planned orders cover expected loss. Wrong or missing scrap factors are a classic reason netting looks correct on paper and fails on the floor.

Lot-sizing rules then shape order quantities. Lot-for-lot orders exactly the net requirement each period; fixed order quantity rounds to a standard pack or batch; period-order-quantity and cost-balancing methods (for example EOQ-style or Silver-Meal) trade setup cost against holding cost. Classical MRP assumes infinite capacity: it plans materials first and validates capacity separately—why MRP II later closed that loop with capacity requirements planning and shop-floor feedback.

Worked netting example: from parent demand to planned orders

A single-level numeric walk-through makes the method concrete. Suppose Product A uses 3 units of Component X per finished unit. Beginning on-hand for X is 500; safety stock is 100; purchase lead time is 2 weeks; lot sizing is lot-for-lot; an open purchase order for 300 of X is due in week 2.

MPS for Product A by week: 50, 80, 60, 100, 70, 90. Gross requirements for Component X (MPS × 3) are therefore 150, 240, 180, 300, 210, 270. Week by week, projected on-hand starts at 500. After week 1 demand (150) it falls to 350—still above safety stock, so no planned order. Week 2 consumes 240 but receives the open PO of 300, ending at 410. Week 3 ends at 230. Week 4 demand of 300 would drive projected on-hand negative; with a 100-unit safety floor the net requirement becomes 170. MRP places a planned order receipt of 170 in week 4 and a planned order release of 170 in week 2 (week 4 minus 2-week lead time). Weeks 5 and 6 similarly generate planned order releases of 210 and 270 in weeks 3 and 4 so projected on-hand never drops below the safety stock floor.

In a multi-level BOM the same logic cascades: planned manufacturing orders for Product A become gross requirements for subassemblies; those planned orders explode into purchased parts. Purchasing works from planned purchase orders (or requisitions) for buy parts; the shop works from planned manufacturing orders for make parts. If Component X were itself made from two purchased parts, each child would get its own grid, lead-time offset, and exception list.

Purchasing MRP: planned orders, PO conversion, and exceptions

For buy items, MRP does not place purchase orders by itself. It creates planning elements—planned purchase orders or purchase requisitions—that purchasing converts (or rejects) into firm supply. Internally made items get planned production or process orders that manufacturing converts into firm work orders. That handoff is where “purchasing MRP” lives in practice: planners and buyers review the planned supply queue, firm what should run, and cancel or reschedule what demand no longer supports.

Exception messages are the planner’s daily worklist. Typical actions include create a new order, change quantity, reschedule earlier (expedite) or later, reschedule and change quantity, or cancel. Good teams triage exceptions by risk: missing parts that stop current jobs first, then near-term firm orders, then longer-horizon forecast noise. Ignoring exceptions is how MRP plans go stale even when the math was correct yesterday.

Run cadence matters as much as the algorithm. Many SME plants run net-change MRP daily (or after major order/receipt events) so the plan keeps pace with demand, and a full regenerative plan weekly as a clean refresh. Running only weekly when demand and receipts move daily leaves the midweek plan stale; running full regeneratives many times a day without dampening floods planners with trivial action messages. A practical pattern: planner owns the exception list every morning; buyer firms near-term purchase suggestions inside the frozen window; shop supervision confirms make-order releases against capacity and material readiness.

Data discipline decides whether purchasing trusts the plan. Open POs must carry correct due dates (scheduled receipts); BOMs must match how you actually build; inventory must match the floor; lead times must match supplier reality. Regenerative MRP recalculates the whole item set; net-change MRP recalculates only items touched by new demand, inventory, or BOM edits—faster for frequent re-runs. Either way, firming too early freezes a bad plan; firming too late leaves purchasing without supplier lead time. Time fences (frozen near-term, slushy mid-horizon, free far out) reduce “MRP nervousness”—constant reschedules that make the shop stop trusting the system.

From MRP to MRP II to ERP: how the discipline evolved

Joseph Orlicky, working at IBM, is widely credited with formalizing modern MRP—the time-phased, dependent-demand, netting-and-offsetting logic that replaced pure reorder-point thinking for complex product structures. Black & Decker (1964) is the standard reference for the first widely recognized production implementation; academic histories also note early collaborative work such as J.I. Case with IBM in the early 1960s. APICS (now ASCM) ran its MRP Crusade in the early 1970s; adoption scaled from hundreds of companies mid-decade toward thousands by the early 1980s.

MRP II (Manufacturing Resource Planning), formalized by Oliver Wight in the early 1980s, extended materials planning into a closed loop: capacity planning (rough-cut and CRP), shop-floor feedback, financial plans in dollars as well as units, demand management, and what-if simulation. ERP (Enterprise Resource Planning)—a term Gartner coined around 1990—extended that manufacturing core enterprise-wide across finance, HR, sales, supply chain, and CRM on a shared data model. MRP is not obsolete; it is the embedded materials-planning engine inside modern manufacturing ERP.

In 2026 the labels are blurry on purpose. Cloud ERP suites ship MRP as a module; lightweight cloud MRP tools add light capacity and shop-floor tracking; APS (advanced planning and scheduling) pairs material logic with finite capacity; Demand-Driven MRP (DDMRP) buffers strategic decoupling points and pulls more from actual demand signals instead of pure forecast push. Practitioners still often run ERP for orders, inventory, and finance while using a separate scheduler for complex job shops—because classical MRP plans materials, not always a realistic machine-hour schedule.

MRP, APS, and MES: how the planning layers fit in 2026

Teams often ask whether they need “an MRP system” when they already have ERP, or whether APS or MES replaces materials planning. They sit on different horizons and answer different questions. MRP (usually inside ERP) explodes BOMs and nets inventory over days to months: what materials to buy or make and when planned orders should release. It classically assumes infinite capacity and validates capacity later. APS (advanced planning and scheduling) is finite-capacity: it sequences work on real machines, crews, tooling, and alternate routings so promised dates respect bottlenecks. MES (manufacturing execution) is the short-horizon execution layer—dispatch, tracking, quality, downtime, and live re-sequence when the floor diverges from the plan.

A practical stack for many discrete plants is: ERP/MRP owns orders, inventory, BOMs, purchasing, and finance; APS or a finite scheduler owns capacity-realistic sequencing for complex job shops; MES owns shop-floor truth and closes the loop. Job-shop manufacturers still commonly report that ERP holds work orders and allocations while a custom or third-party scheduler reads ERP data live and drives the floor—because pure MRP material plans alone do not tell machine 12 what starts at 10:00. Modern cloud ERP often embeds light scheduling; heavy custom or highly constrained plants still pair ERP with dedicated APS/MES.

Choose layers by failure mode. If stockouts and excess come from wrong BOMs, lead times, or open PO dates, fix MRP data and exception discipline first. If materials look fine but every promise slips on the same bottleneck work center, add capacity planning or APS—not another material-only replan. If plans are good but the floor never reports completions, scrap, or substitutions accurately, inventory and MES feedback—not a fancier explosion—are the bottleneck.

How Microsoft Dynamics 365 handles MRP

Microsoft Dynamics 365 splits MRP across two products sized for different manufacturers, and Flectic implements both for SMEs across Canada, the UK, and the USA.

Dynamics 365 Business Central (mid-market) runs MRP and MPS from a single Planning Worksheet. MPS calculates a master production schedule for end items driven by sales orders and forecast; MRP then calculates material requirements for dependent, non-MPS components at lower BOM levels. Planners choose Calculate Regenerative Plan for a full replan or Calculate Net Change Plan for incremental updates. The engine emits Action Messages—New, Change Quantity, Reschedule, Reschedule and Change Quantity, and Cancel—which Carry Out Action Message converts into planned or firm planned production orders, purchase orders, or transfer orders.

Reordering policies on the Item Card (Lot-for-Lot, Fixed Reorder Qty., Maximum Qty., and Order) combine with manufacturing policy (Make-to-Stock vs Make-to-Order), safety stock, reorder point, time bucket, order modifiers, and low-level codes to govern multi-level BOM planning. As in classical MRP, Business Central planning assumes infinite capacity and validates it later in scheduling—pairing the plan with manufacturing routings and work centers when you need executable capacity.

Dynamics 365 Supply Chain Management (large enterprise) uses the Master Planning module with Planning Optimization. Coverage settings cascade from Coverage groups to Item coverage to Master plan overrides, with coverage codes including Manual, Per requirement, Per period, Min/Max, Priority, and Decoupling Point for DDMRP. Output is planned orders (purchase, production, transfer, kanban) that are firmed manually or automatically within a firming time fence, with positive and negative days shaping the action messages planners see.

How Odoo handles MRP

Odoo's Manufacturing and Inventory apps take a route-and-rule-driven approach rather than one monolithic time-phased MRP pass—a model that suits smaller manufacturers who want lighter setup while still covering BOM explosion, replenishment, and shop execution.

The building blocks are Bills of Materials (BoMs) that define components and quantities (and optional Operations when Work Orders is enabled), and Manufacturing Orders (MOs) as the execution unit. Multi-level and sub-assembly BoMs, by-products, flexible consumption, kit BoMs, and BoM variants are all supported. When Work Orders is enabled, each operation in the BoM links to a Work Center that tracks capacity, cost, scheduling, and efficiency; operators run orders through the Shop Floor tablet interface. That operations sequence is your manufacturing routing in Odoo terms.

Replenishment is driven by Reordering Rules and routes. A reordering rule keeps forecasted stock between a minimum (trigger) and a maximum (target) quantity; a Buy route creates an RFQ, while a Manufacture route creates a manufacturing order. The Make-to-Order (MTO) route triggers procurement or production only after a sales order is confirmed and reserves the output to that order—suited to low-volume, high-variety, or expensive items—while Make-to-Stock relies on reordering rules against forecasted quantity. A daily scheduler evaluates each product's forecasted quantity (on-hand plus incoming minus outgoing) against its minimum and fires any Auto rules, with lead times used to time replenishment.

When spreadsheets stop working: SME path from Excel to ERP MRP

Most small manufacturers do not jump from paper to enterprise APS. They climb a short ladder. Spreadsheets and reorder lists work while SKU count is low, BOMs are shallow, and one planner holds the whole plan in their head. They break when multi-level BOMs, shared components, long lead times, and frequent demand changes make netting and time-phasing too error-prone to maintain by hand—stockouts and expedites become the default planning system.

A dedicated cloud MRP or manufacturing module answers what/how much/when without forcing a full finance+HR cutover on day one. Embedded ERP MRP (Business Central Planning Worksheet, Odoo Manufacturing + Inventory, or similar) is usually the right target once purchasing, inventory, production orders, and accounting need one shared truth. Full multi-site ERP with Master Planning, APS, or MES is warranted when capacity constraints, multi-warehouse transfers, or regulated traceability outgrow a single planner’s exception list.

Stay on spreadsheets only if product structures stay simple and you can still hit on-time delivery without heroic expediting. Move to structured MRP when BOM explosion and lead-time offsets are already happening in someone’s head or three linked workbooks. Prefer an ERP MRP module over a pure-point tool when you need planned orders to become real POs, work orders, and inventory transactions without CSV shuttle. Add APS/MES only after materials data is trustworthy—finite scheduling on fiction still ships late.

Why MRP fails—and when not to start with full MRP

MRP does not fail because the math is mysterious. It fails when inputs lie and the organization treats the engine as set-and-forget software. The highest-frequency failure modes are well documented across implementation checklists and practitioner war stories: inaccurate BOMs (wrong qty-per, missing parts, stale revisions), phantom inventory (system on-hand that the shelf does not hold because receipts, issues, and scrap were never captured), stale lead times (catalog promises instead of last-quarter supplier performance), missing scrap or yield factors (ordering for perfect process), unstable MPS (sales churning the near-term schedule daily), and planners who ignore action messages and keep a private spreadsheet as the real plan.

Inventory accuracy is often an event-capture problem, not a once-a-year counting problem. If receiving, bin moves, picks, scrap, and returns are not recorded when they happen, every netting run plans against fiction—and buyers correctly stop trusting planned orders. BOM integrity is the other pillar: if the BOM is not the recipe you actually build, explosion multiplies the wrong world. Practitioners who teach BOM thinking from first principles (finished good → subassemblies → purchased parts with vendors, MOQs, and lead times) are describing the same discipline MRP assumes is already true in the database.

There are also times when full multi-level regenerative MRP is the wrong next move. Do not start there if product structures are still shallow, item masters are incomplete, or the floor cannot name a reliable on-hand for A-items—fix scan discipline and BOM audits first. Job shops with extreme high-mix custom work often need ERP for orders, inventory, and finance, but a finite scheduler or dedicated planning layer for capacity-realistic sequencing; classical infinite-capacity material plans alone will not tell machine 12 what starts at 10:00. If your real pain is a single bottleneck machine while materials are already easy, APS or capacity planning beats another material-only replan. If demand is pure make-to-order with almost no shared components, simple MTO routes and reordering rules (as in lighter Odoo-style models) may deliver more trust than a nervous full-grid MRP.

A healthy system shows measurable trust signals: BOM accuracy north of about 98% on in-scope items, inventory record accuracy about 95%+, planner override rates that are not routine (high override rates usually mean data debt, not “smart planning”), manageable meaningful action messages, and improving stockout and on-time delivery trends. Until those signals move, expanding scope multiplies noise.

MRP implementation for SMEs: readiness checklist and what breaks

MRP implementation for small and mid-size manufacturers is usually incremental: spreadsheets or basic inventory tools first, then a lightweight cloud MRP or manufacturing ERP module, with BOMs, inventory accuracy, production planning, and purchasing before broader finance and CRM scope. Empirical work on manufacturing SMEs consistently favors phased rollout over big-bang cutover, starting where pain is acute—stock control, production planning, and BOM integrity. Practitioners and implementation guides converge on the same gate: data preparation is the number-one failure mode, not the brand of software.

Treat readiness as measurable, not aspirational. Implementation checklists used across MRP programs commonly target 98%+ BOM accuracy against actual builds, 95%+ inventory record accuracy, and lead times validated against real supplier and manufacturing performance—not catalog promises—for critical items. Also lock make-versus-buy flags, lot-sizing rules (lot-for-lot, fixed order quantity, order multiples), safety stock or coverage policies, and open purchase and production orders with correct due dates before the first live regenerative plan. Without those inputs, every planned order is fiction and buyers stop trusting exceptions.

Two process risks dominate after data. Training and discipline: planners must run exception reviews on a cadence (daily for near-term, weekly for the horizon), not only press Calculate. Capacity is the third trap: infinite-capacity material plans that ignore machine hours and labor look green in MRP and red on the floor; close that loop with routings, rough-cut or finite scheduling, or a realistic capacity check before you promise customer dates. Phased scope beats heroics—start with the top revenue or highest-pain product families, prove the handoff from planned orders to purchasing and the shop, then expand.

Platform fit matters more than brand labels. Business Central suits manufacturers who want a tightly integrated planning worksheet with deep production-order and routing control. Odoo's modular, route-driven model suits teams that want to start with Manufacturing and Inventory and expand. Because Flectic implements both Dynamics 365 and Odoo, an ERP Readiness Call maps your BOMs, planning triggers, purchasing handoff, and capacity model—not a single-vendor pitch.

Frequently asked questions