MRP vs ERP vs APS

Executive summary

Material requirements planning is a netting calculation. It explodes a bill of materials against a production schedule, subtracts what is already on hand or on order, and offsets the result by fixed lead times to say what should be ordered and when. It assumes capacity is unlimited and that lead times are constant. MRP II added feedback loops and a financial view, and enterprise resource planning absorbed the whole calculation into a shared database alongside finance, procurement, and human resources, but neither changed the underlying assumption. Advanced planning and scheduling changes it: APS plans material and capacity at the same time, respects the constraints of real resources, and produces a schedule that can be executed. The question a buyer should ask is therefore not which product to buy but which arithmetic the planning engine performs.

What does MRP actually calculate, and why does it assume infinite capacity?

MRP begins with a master production schedule stating what finished goods are required and when. It explodes each item through its bill of materials to determine the dependent demand for components and subassemblies. It nets that gross requirement against inventory on hand and orders already placed, producing a net requirement. Finally it offsets each net requirement backward in time by a lead time held against the item, generating planned orders with release dates. The output is a set of instructions: order this quantity of this part on this date.

The method was set out by Joseph Orlicky, an engineer at IBM, whose 1975 book established the approach that professional practice adopted through the following decade. Its insight was to distinguish dependent demand, which can be calculated from a schedule and a bill of materials, from independent demand, which must be forecast. That distinction remains correct and is the reason MRP is still the foundation of manufacturing planning half a century later.

The limitations sit in two assumptions the arithmetic requires. The first is infinite capacity: the calculation does not ask whether the resources needed to execute the planned orders are available on the dates it proposes. It assumes that whatever is scheduled can be produced. The second is fixed lead time: each item carries a lead time treated as a constant, when in practice lead time on a loaded resource depends on queue, and queue depends on load. Both assumptions make the calculation tractable, which is why they were adopted, and both are false on a constrained shop floor.

The visible symptom is a planning output nobody trusts. Planned orders arrive with start dates that cannot be met, the shop floor reorders them by judgment, and the plan and the reality diverge within days. Organizations often respond by inflating lead times in the item master to create slack, which increases inventory and lengthens quoted lead times without addressing the cause. Recognizing that pattern as a structural property of infinite-capacity logic, rather than as a data quality failure or a planner performance problem, is the single most useful diagnostic in this area.

What did MRP II and ERP add, and what did they leave unchanged?

Manufacturing resource planning, usually written MRP II, extended the original method rather than replacing it. It closed the loop, feeding execution results back into planning so that the plan is revised against what actually happened. It added rough-cut capacity planning, a coarse check of whether the master schedule is broadly feasible against key resources. It brought shop floor control and purchasing into the same framework, and it connected the operational plan to a financial view so that the same plan could be expressed in money. These were substantial improvements in how planning was governed.

What MRP II did not change is the material planning core. Rough-cut capacity planning tests the master schedule at an aggregate level; it does not make the detailed MRP netting calculation capacity-aware. The planned orders that emerge are still generated on infinite-capacity assumptions with fixed lead times, which means the improvement is in oversight rather than in arithmetic.

Enterprise resource planning then absorbed MRP II into a wider system spanning finance, procurement, human resources, sales, and more, unified by a shared database. The gain was integration: one record of a purchase order visible to planning, receiving, and accounts payable at once, with fewer reconciliations and less duplicate data entry. That gain is real and is the reason ERP became standard. But the planning module inside ERP is, in most systems, the same calculation Orlicky described, and buyers who believe ERP superseded MRP are usually surprised to find their ERP planning run behaving exactly as MRP always did.

This history explains a persistent confusion in requirements documents. A team writes that it needs better planning and receives proposals for an ERP upgrade, because the vendor hears an integration requirement. If the underlying problem is that plans are infeasible against capacity, an ERP upgrade will not solve it, since the new system will run the same infinite-capacity logic on cleaner data. Naming the arithmetic in the requirement avoids months of misdirected evaluation.

Table 1. The planning approaches compared. Reading down the first two rows is sufficient to place any product a vendor presents, because the capacity assumption is the property that determines what the output can be trusted to mean.

What does APS do differently, and when do I need finite-capacity planning?

Advanced planning and scheduling plans material and capacity simultaneously. Rather than generating orders and then discovering whether they can be produced, it treats resource availability, sequencing rules, changeover times, and material availability as constraints within a single problem, and searches for a schedule that satisfies them. The techniques vary, from mathematical optimization to heuristics, but the defining property is that a constraint is binding: if the resource cannot do the work in the window, the system does not propose that it does.

The consequence is that lead time becomes an output rather than an input. In MRP a planner supplies a lead time and the system schedules around it. In APS the schedule determines when work will actually finish, given everything else competing for the same resources. This is why APS output tends to be more credible to the shop floor, and also why it is more demanding: an APS model is only as good as its representation of routings, capacities, and changeovers, and organizations with weak master data frequently find that the honest schedule exposes data problems the previous system concealed.

Deciding whether the capability is needed is best done with signals rather than with company size. Persistent expediting, a shop floor that visibly reorders the plan by judgment, quoted lead times that keep being inflated, a bottleneck resource whose sequence materially changes throughput, and significant changeover costs that reward grouping similar work all point toward finite capacity. A single-site operation with ample capacity, stable demand, and simple routings can plan adequately with ERP MRP and a capable planner, and adding APS in that setting introduces integration cost and a master data burden for marginal benefit.

Table 2. Signals on each side of the decision. The last row is a precondition rather than a signal: an organization whose routing and capacity data is unreliable should fix that before buying a system that depends on it.

Have ERP planning modules closed the gap with standalone APS?

Partly, and the honest answer is that the question is now about configuration rather than about category. Major ERP vendors offer constraint-based planning capability, sometimes as a distinct module and sometimes as an option within a planning suite. It is therefore inaccurate to claim that ERP cannot do finite capacity. What remains accurate is that ERP planning defaults to infinite capacity, and that finite capability is generally a separate module, a separate license, and a separate implementation effort rather than a setting.

The test that settles it in an evaluation is narrow and worth writing into requirements. Ask whether the tool can treat opening capacity, adding a shift, or moving work to an alternate resource as decision variables, and whether it will refuse to schedule work that exceeds available capacity. A system that reports an overload after the fact is performing capacity requirements planning, which is a check. A system that will not produce the infeasible plan in the first place is performing finite-capacity scheduling. Both are useful, but only the second changes what the planner receives.

Two practical considerations follow for buyers weighing an ERP module against a specialist product. Integration favors the module, since the planning engine sits on the same data and the interfaces are already built, and that advantage is real rather than marketing. Depth of modeling frequently favors the specialist, particularly where sequencing rules, changeover matrices, or campaign planning are complex, because that is the problem specialists have spent decades on. The decision therefore rests on how complex the scheduling problem actually is, which is a question about the operation rather than about the software.

Is DDMRP a real methodology or a branded framework?

Both descriptions are partly right, and the fair answer requires holding them together. Demand driven material requirements planning is a defined methodology positioning strategic buffers at decoupling points in the flow, sizing those buffers dynamically, and generating supply orders based on buffer penetration rather than on a forecast exploded through the bill of materials. Its aim is to reduce the amplification of demand variability that conventional MRP can transmit through a chain.

It is also a trademarked, certified framework owned and promoted by the Demand Driven Institute, which sells training and certification and has a commercial interest in adoption. That should be disclosed whenever the method is discussed, and it means that material published by the Institute or its affiliates ought to be read as advocacy rather than as neutral evidence.

The independent literature is where the useful answer sits, and it is more mixed than either camp usually acknowledges. Peer-reviewed simulation work has found advantages for the method over conventional MRP, particularly under high demand variability, which is a real result and should not be dismissed. Later independent work has qualified it: benefits appear contingent on how tight due dates are, dynamic buffer sizing has been found in at least one study to perform no better than simpler fixed buffers, the standard buffer-penetration dispatching rule has been criticized, and reviewers have noted that the method is under-specified on how its parameters should be set. Taken together, that is the profile of a method with genuine merit and an unsettled evidence base, not a breakthrough and not a marketing exercise.

A buyer's practical position follows from this. Treat DDMRP as one option for managing variability rather than as a replacement for planning logic, note that its most cited supportive study is not fully at arm's length from the body that promotes the method, and be skeptical of any implementation quote that presents the benefits as settled. Where variability is the dominant problem, the approach merits evaluation on its own terms alongside alternatives.

Frequently asked questions

Is MRP obsolete?

No. The distinction between dependent and independent demand that MRP formalized remains correct, and the netting calculation is still the foundation of manufacturing planning. What is dated is the assumption that capacity is unlimited, which is why the method is frequently supplemented rather than replaced.

Does my ERP already include MRP?

Almost certainly yes. ERP absorbed MRP as a planning module, and in most systems the calculation runs substantially as it always has. The more useful question is whether your ERP includes finite-capacity planning, which is usually a separate module rather than part of the standard planning run.

What is the difference between MRP and MRP II in one sentence?

MRP calculates what to order and when, while MRP II wraps that calculation in closed-loop feedback, rough-cut capacity checks, shop floor control, and a financial view, without making the material planning core capacity-aware.

Can APS replace my ERP?

No. APS is a planning engine, not a transactional system of record. It depends on ERP for master data, inventory positions, and order records, and it writes its results back. Treat it as a specialist layer above the transactional system rather than as an alternative to it.

Where does APS sit relative to MES?

APS decides what should be made, in what sequence, on which resources. A manufacturing execution system governs and records what is actually happening on the floor. They are complementary, and the interface between them matters, since a schedule that execution cannot report against is difficult to keep current.

Do I need APS if I am a single-site manufacturer?

Not necessarily. Site count matters less than whether capacity binds. A single site with a genuine bottleneck, complex sequencing, and costly changeovers may need it more than a multi-site operation with ample capacity and simple routings.

Is DDMRP the same as MRP?

No. It uses strategic buffers at decoupling points and generates supply orders from buffer penetration rather than from a forecast exploded through the bill of materials. It is a different planning logic that addresses variability, and it is a trademarked methodology promoted by an interested body.

What does finite capacity scheduling mean in plain terms?

It means the system will not plan work that the available resources cannot perform in the time allowed. If a machine has eight hours available and the work requires twelve, a finite-capacity system moves work rather than reporting an overload and proceeding.

Why do our planned orders always look infeasible?

Usually because the planning engine assumes infinite capacity and fixed lead times, so it schedules against dates rather than against resources. Inflating lead times in the item master is the common response, and it raises inventory and quoted lead times without addressing the cause.

What should we fix before buying APS?

Routing accuracy, capacity definitions, and changeover data. A constraint-based engine models what you tell it, so weak master data produces a confident schedule built on wrong assumptions. Most implementations that stall do so on data rather than on the algorithm.

Methodology, caveats, and sources

Methodology

• The description of the MRP calculation and its assumptions follows Orlicky's founding text and standard practitioner definitions rather than vendor product descriptions.
• The assessment of DDMRP rests on independent peer-reviewed literature, cited below, rather than on material published by the body that owns the methodology. Where a study is not fully at arm's length from that body, this is stated.
• Process framing follows the SCOR Digital Standard maintained by the Association for Supply Chain Management, a professional body rather than a regulator.
• Supply Chain Research is independent and vendor-neutral. We accept no payment from the vendors or categories covered, and this page names no products.

Caveats

• SCR publishes no benchmark for what APS delivers. Percentage improvements in inventory, on-time delivery, or utilization circulating in this area come from vendor case studies selected by the vendor, without sample or method, and should not be used in a business case.
• The academic literature cited on DDMRP is largely simulation-based. It establishes how the method behaves under modeled conditions; it does not measure what typical organizations achieve in production.
• Figure 1, Table 1, and Table 2 are structural illustrations and decision aids rather than measured data.
• This page addresses planning logic only. Which vendor to buy from, how to structure the contract, and how to sequence an implementation are separate questions covered elsewhere in SCR's library.

Where to go deeper

Readers whose question is whether to buy a suite or a specialist should read the SCR guide to ERP versus best-of-breed, which addresses the architecture decision this page deliberately sets aside. The supply chain network design guide covers the structural decisions that sit above planning and bound what any engine can achieve. Readers scoping across categories should start with the SCR supply chain software category map, those building a business case should read the SCR software ROI method, and those preparing to evaluate products should read the SCR selection framework before contacting vendors.

Sources

1. Orlicky,Joseph. MaterialRequirements Planning, 1975. Archived edition.Primary founding text.
2. Orlicky,Plossl and Wight. Orlicky'sMaterial Requirements Planning, later edition.Primary.
3. Miclo,Lauras, Fontanili, Lamothe and Melnyk. DemandDriven MRP: assessment of a new approach to materials management.International Journal of Production Research.Peer reviewed and supportive of DDMRP. Not fully independent: aprincipal of the body that owns the methodology was associated withthe underlying research.
4. Thurer,Fernandes and Stevenson. Onthe meaning and use of excess capacity in demand driven MRP.International Journal of Production Research.Peer reviewed and independent; qualified validation contingent on duedate tightness.
5. Martin,Lauras and Baptiste. Assessmentof dynamic buffer adjustment in demand driven MRP. Computers andIndustrial Engineering.Peer reviewed and independent; finds dynamic buffer sizing may notoutperform fixed buffers.
6. Fernandes,Thurer and Carmo-Silva. Onthe dispatching rule used in demand driven MRP. Operations ManagementResearch.Peer reviewed and independent; critiques the standardbuffer-penetration rule.
7. Damand,Lahrichi and Barth. Parameterisationof demand driven MRP. International Journal of Production Research.Peer reviewed and independent; notes the method is under-specified onparameter setting.
8. Orue,Lizarralde and Kortabarria. Demanddriven MRP: a systematic review. Journal of Industrial Engineeringand Management.Peer reviewed and independent.
9. Bahu,Bironneau and Hovelaque. DemandDriven MRP: literature review and research issues. HAL working paper.Independent and explicitly skeptical.
10. DemandDriven Institute. DDMRPmethodology overview.Interested source: owns and certifies the methodology and sellstraining.
11. Associationfor Supply Chain Management. SCORDigital Standard overview.Professional body.
12. Associationfor Supply Chain Management. SCORDS model reference.Professional body.

Supply Chain Research is an independent, vendor-neutral research platform for supply chain and technology leaders. We accept no payment from the vendors, consultancies, or firms discussed. This article is analysis, not legal, procurement, or investment advice, and its conclusions should be validated against your own circumstances before any decision.