10. Spare Parts, MRO Inventory and Supply Networks

Introduction and Context

Maintenance performance depends not only on skilled technicians and sound maintenance procedures, but also on the availability of the right spare part at the right time. A failed bearing, valve actuator, sensor, seal, control module, or motor may be inexpensive in relation to the plant, yet its absence can interrupt production for hours or days. Conversely, excessive stocks of rarely used parts tie up capital, occupy storage space, become obsolete, and conceal weaknesses in planning and data quality.

Spare-parts management therefore differs from ordinary production inventory management. Demand is often intermittent, uncertain, and driven by equipment failure rather than a stable production schedule. Lead times may be long, suppliers may be few, and the cost of stockout can include lost production, safety exposure, environmental damage, emergency freight, and reputational harm. Maintenance, procurement, stores, finance, engineering, operations, and suppliers must consequently be treated as parts of one service network.

The term MRO refers to maintenance, repair, and operations materials. It includes replacement components, consumables, tools, lubricants, safety items, and operating supplies. This chapter develops an integrated approach to classifying, forecasting, stocking, sourcing, repairing, pooling, and governing MRO inventories. The central managerial question is not “How much inventory should be held?” but rather “What level of availability is economically and operationally justified for each item and asset?”

Learning Objectives

After studying this chapter, the reader should be able to:

1. Explain the distinctive characteristics of spare-parts and MRO demand.

2. Classify items according to criticality, demand behaviour, value, and supply risk.

3. Select appropriate forecasting, service-level, and replenishment policies.

4. Evaluate repair-versus-replace alternatives using life-cycle economics.

5. Design supplier, pooling, and supply-network strategies for resilience.

6. Apply digital tools and analytics to improve inventory visibility and decisions.

7. Diagnose common governance failures and recommend practical improvements.

Why Spare-Parts Inventory Requires a Different Logic

Production materials are usually consumed in planned quantities, whereas spare parts may remain unused for months and then be required immediately. Demand may be classified as regular, intermittent, erratic, lumpy, or obsolete. A part with annual demand of two units cannot be managed effectively by simply applying a conventional economic order quantity formula without considering failure consequences and lead time.

Four costs shape the decision. The first is the acquisition cost, including purchase price, transport, inspection, and receiving. The second is the holding cost, which includes capital, storage, insurance, preservation, deterioration, and obsolescence. The third is the ordering or transaction cost. The fourth is the shortage cost: downtime, expedited procurement, contractor charges, production rescheduling, and possible safety or regulatory consequences.

A low-value component can be operationally critical. A high-value component may be non-critical if an alternative asset is available or if it can be obtained rapidly. Thus, inventory value alone cannot determine policy. Criticality is a consequence-based concept: it asks what happens if the item is unavailable when required.

Several terms must be distinguished. A consumable is normally used up during maintenance, such as grease, welding rods, or filters. A repairable or rotable item can be restored and returned to service. A capital spare is a high-value item normally associated with major equipment, such as a turbine rotor or transformer. A common spare can serve several equipment models, while a unique spare has limited interchangeability. Insurance stock is held primarily against severe consequences and uncertain replenishment, not because frequent demand is expected.

A Classification Architecture for MRO Materials

No single classification is sufficient. Mature systems use several dimensions, because value, usage, criticality, predictability, and supply risk describe different managerial problems.

Criticality classification

Criticality analysis evaluates the consequence of non-availability. A practical scale may include:

• Class 1: failure creates immediate safety, environmental, legal, or major production consequences.
• Class 2: failure causes significant production loss or extended repair delay but has a feasible workaround.
• Class 3: failure has limited operational impact and can normally be managed through routine procurement.

The assessment should consider failure severity, likelihood, detectability, repair duration, redundancy, bypass options, and supplier lead time. Criticality must be attached to the asset and the service requirement, not merely to the part description.

ABC classification

ABC analysis ranks items by annual consumption value, commonly calculated as annual quantity issued multiplied by unit cost. A-items account for a large share of value and deserve close financial control; C-items have low annual value and may be managed with simplified procedures. ABC analysis is useful for prioritising review effort, but it is not a substitute for criticality analysis. A low-value fuse may be more important than an expensive spare used on redundant equipment.

Demand and variability classification

A commonly used approach combines demand volume and variability. Smooth, frequent demand is easier to forecast than intermittent demand. Measures such as the coefficient of variation, demand intervals, and the proportion of zero-demand periods help distinguish regular from lumpy demand. For intermittent items, forecasting methods such as Croston-type methods or its variants can be more suitable than ordinary moving averages, although managerial judgement remains essential.

Supply-risk classification

Supply risk reflects supplier concentration, geographic exposure, manufacturing complexity, minimum order quantities, intellectual-property restrictions, transport dependence, and obsolescence risk. A sole-source electronic module with a twelve-month lead time should receive more attention than a locally available standard fastener, even if the latter is used more frequently.

A useful portfolio combines these dimensions. For example, a critical, high-value, long-lead-time item may require executive ownership, validated specifications, preserved stock, and a supplier continuity plan. A non-critical, low-value, readily available item may be governed through a catalogue and simple min–max control.

Forecasting Intermittent Demand

Forecasting spare-parts demand is difficult because historical issues may reflect failures, planned overhauls, cannibalisation, stockouts, or data errors. Zero issues do not necessarily mean zero need. Demand history should therefore be reconstructed from work orders, equipment failure records, planned maintenance schedules, purchase orders, and inventory adjustments.

For regular demand, moving averages, exponential smoothing, seasonal models, or causal models may be appropriate. For intermittent demand, the forecast should estimate both the probability of a demand occurrence and the quantity demanded when it occurs. Croston’s method separates these two components. Its variants attempt to correct biases associated with long intervals between demands and changing demand patterns.

Forecasting should not be treated as an automatic calculation. Engineering changes, asset retirements, maintenance strategy changes, reliability improvements, warranty conditions, and supplier discontinuation notices can make historical demand irrelevant. A forecast review should ask:

• Has the installed asset population changed?
• Is the part still specified in current equipment?
• Has preventive maintenance altered the failure mechanism?
• Are issues genuine consumption or merely stock transfers?
• Is demand suppressed because the item was unavailable?

Forecast accuracy should be measured with methods suitable for intermittent demand. Percentage errors can be misleading when actual demand is zero. Bias, mean absolute deviation, forecast value added, service level, and stockout frequency may provide more meaningful evidence.

Service Levels and Stock Policies

A service level expresses the probability or proportion of demand fulfilled without delay, but its meaning must be specified. The cycle service level is the probability of avoiding a stockout during a replenishment cycle. The fill rate is the proportion of total demand supplied immediately. These measures are not interchangeable.

Safety stock protects against uncertainty in demand and lead time. For relatively stable demand, a simplified expression is:

Reorder point = expected demand during lead time + safety stock

If demand and lead time variability are approximately independent, safety stock may be estimated from the standard deviation of lead-time demand multiplied by a service factor. However, intermittent demand and highly uncertain lead times reduce the reliability of normal-distribution assumptions. Critical items may require scenario analysis, simulation, or a management-determined emergency reserve instead.

Common replenishment policies include:

• Continuous review (s, Q): order quantity Q when inventory position reaches reorder point s.
• Continuous review (s, S): order enough to raise inventory position to target level S.
• Periodic review (R, S): review every R periods and order up to S.
• Min–max policy: replenish when stock falls at or below the minimum, usually to a maximum level.
• Order-on-demand: purchase only when a specific requirement occurs, appropriate for low-criticality or highly expensive items with reliable suppliers.
• Two-bin or kanban control: use a visual replenishment trigger for repetitive consumables.

Inventory position includes on-hand stock plus on-order and repairable quantities, less reservations and backorders. Ignoring these elements causes duplicate ordering and inflated stock. Policy parameters must also account for shelf life, preservation requirements, packaging, substitution, and minimum order quantities.

The correct service level is economically differentiated. A safety-critical component may justify a very high availability target, while a low-criticality item may rationally tolerate a stockout. Setting every item at the same target creates either excessive inventory or unacceptable risk.

Repair, Replace, Cannibalise, or Pool?

A failed part may be purchased new, repaired, refurbished, cannibalised from another asset, borrowed from a sister site, or replaced by an approved alternative. The decision should use total cost and total risk rather than purchase price alone.

Relevant repair-versus-replace factors include:

• direct repair and testing cost;
• turnaround time and transport;
• probability of successful repair;
• expected post-repair life and warranty;
• diagnostic and teardown costs;
• availability of replacement units;
• technological obsolescence;
• safety and regulatory approval;
• environmental disposal implications; and
• opportunity cost of the equipment being unavailable.

An illustrative expected-cost expression is:

Expected repair cost = repair cost + failure probability after repair × consequence of repeat failure + delay cost

Repair is attractive when the unit is technically recoverable, the repair process is reliable, and replacement lead time is long. Replacement is attractive when repair quality is uncertain, the technology is obsolete, or the price difference is small relative to downtime risk. Cannibalisation can provide emergency relief but may transfer failure risk to another asset and should be controlled through asset records and subsequent restoration plans.

Repairable pools require closed-loop management. Each unit should have an identity, condition status, location, repair history, warranty status, and expected return date. Without such visibility, organisations often purchase new units while repairable stock remains lost in transit or in an external workshop.

Supplier Management and Supply-Network Design

Spare-parts supply networks often contain original equipment manufacturers, authorised distributors, specialist repair houses, local fabricators, logistics providers, and internal warehouses. Supplier selection should consider technical conformity, capacity, lead time, responsiveness, financial stability, quality systems, cybersecurity where relevant, and continuity plans.

A single-source arrangement may be justified by proprietary design or warranty constraints, but it creates concentration risk. Mitigation may include approved alternate sources, second-source qualification, last-time buys, technical drawings and bills of material, framework agreements, vendor-managed inventory, consignment stock, and strategic safety stock. Dual sourcing is not automatically superior: splitting demand may reduce economies of scale or prevent either supplier from maintaining sufficient competence.

Contracts should define more than price. They may include lead-time commitments, emergency response, test certificates, obsolescence notification, minimum shelf life, repair turnaround, warranty, change-control notification, counterfeit prevention, and data-sharing requirements. Supplier performance should be reviewed using measures such as on-time delivery, lead-time reliability, quality acceptance, emergency response, fill rate, and total cost.

Supply resilience requires visibility beyond the immediate supplier. A critical control board may depend on a particular semiconductor, software licence, or specialist manufacturer several tiers upstream. Scenario analysis can test the effects of transport disruption, natural hazards, geopolitical restrictions, cyber incidents, labour shortages, and demand surges. Resilience is not unlimited redundancy; it is the capacity to sustain or quickly restore required maintenance service at an acceptable cost.

Pooling, Standardisation, and Network Decisions

Pooling spare parts across plants can reduce total safety stock because independent demand risks partially offset one another. The benefit is strongest when parts are interchangeable, demand is not perfectly correlated, and transportation time is acceptable. Pooling becomes less valuable when sites experience common-cause failures, such as a regional weather event, or when urgent transport eliminates the response advantage.

A network may use a central distribution centre for expensive and infrequently demanded items, regional hubs for critical fast-response parts, and local stores for frequently used consumables. The design should compare inventory reduction with transport time, handling cost, customs exposure, and the operational consequence of central-stock unavailability.

Standardisation is a powerful but often overlooked inventory strategy. Reducing the number of equivalent filters, sensors, fasteners, or motor specifications can increase interchangeability and improve bargaining power. However, standardisation must not compromise equipment performance, certification, maintainability, or life-cycle cost. Engineering governance is required before substituting a part.

Digital Systems, Analytics, and Data Governance

An enterprise asset management or computerized maintenance management system should connect asset hierarchies, bills of material, work orders, inventory balances, reservations, suppliers, and repair status. A reliable part master is fundamental. Duplicate descriptions, inconsistent units, missing manufacturer numbers, and obsolete records produce both overstock and stockouts.

Useful analytics include criticality heat maps, slow-moving and obsolete inventory analysis, stockout prediction, lead-time variability dashboards, repair-loop ageing, supplier risk scoring, and what-if simulation. Machine learning may identify relationships between operating conditions and component failures, but it should support—not replace—engineering judgement. Sparse and biased failure data can produce false confidence.

Digital tools also create governance risks. Access controls should prevent unauthorised changes to item specifications and reorder parameters. Audit trails are important for high-value or safety-critical materials. Cybersecurity requirements apply to connected warehouses, supplier portals, automated storage systems, and maintenance platforms.

Illustrative Case: Regional Pooling of Critical Pump Spares

Background

A fictional chemical-processing company operated four plants within one country. Each plant maintained its own inventory of mechanical seals and pump cartridges. The parts were technically interchangeable for a major pump family, but item descriptions differed across sites.

Situation and problem

The company held 38 seal kits valued at approximately 1.2 million currency units, yet two plants experienced emergency purchases during the same year. An audit found that six kits were approaching shelf-life limits, three were awaiting inspection after repair, and several work orders had reserved stock that was not actually required. Each plant used different reorder parameters.

Analysis

A cross-functional team reviewed failure history, pump criticality, lead times, shelf life, repair turnaround, and transport routes. Demand was intermittent, but the consequence of a missing kit on one solvent recovery pump was severe. The team found that demand correlation was low under normal conditions but could become high during a common process upset. Therefore, complete centralisation would create excessive common-cause risk.

Approach and solution

The company created a common part master and approved interchangeability rules. Two high-criticality kits were retained at each plant, while lower-frequency variants were pooled at a regional hub. A repair loop with serial tracking was introduced. Reorder points were recalculated using lead-time demand, shelf-life constraints, and a specified emergency response time. The supplier contract added minimum remaining shelf life, technical-change notification, and emergency dispatch provisions. Monthly reviews examined stockouts, ageing, and obsolete inventory.

Outcome

In this illustrative case, the policy reduced duplicated holdings while preserving local protection for the most critical pumps. The main improvement did not come from a mathematical formula alone; it came from standardisation, accurate records, repair visibility, and explicit decisions about which risks should be shared and which should remain local.

Learning points

The case demonstrates that pooling is a network decision, not merely a warehouse decision. It also shows why criticality, interchangeability, common-cause risk, and data quality must be analysed together. Inventory reduction is sustainable only when the maintenance service requirement remains visible.

Challenges, Risks, and Common Mistakes

Several recurring mistakes weaken MRO performance. Treating all items with the same service level ignores consequence differences. Relying solely on ABC analysis overlooks low-value critical items. Using annual average demand for intermittent parts can generate misleading reorder points. Counting on-hand stock without checking condition, shelf life, location, or reservation status overstates availability.

Other failures arise from organisational behaviour. Maintenance personnel may hoard parts after previous stockouts. Procurement may select the lowest purchase price while ignoring downtime and repair quality. Finance may impose blanket inventory reductions without distinguishing strategic spares. Stores may create duplicate item codes because searching is difficult. Engineering changes may not be transferred to the parts master.

There are also technical risks. Counterfeit or non-conforming components can compromise safety. Poor preservation can damage seals, electronics, and lubricants before use. Unapproved substitutions may invalidate certification or warranties. Excessive emergency buying may conceal chronic planning failures and produce supplier dependency.

Governance, Ethics, and Professional Responsibility

Spare-parts decisions involve safety, environmental protection, fairness, and financial stewardship. A manager must not reduce stock below a safe level merely to meet a working-capital target. Conversely, retaining obsolete inventory without review misuses organisational resources. Criticality assessments should be documented and traceable.

Supplier relationships require transparent specifications, consistent evaluation, conflict-of-interest controls, and protection against counterfeit materials. Procurement records should support auditability. Where parts affect regulated equipment, certificates, traceability, calibration, and approved repair procedures must be preserved. Ethical practice also requires that maintenance personnel report unsafe substitutions and that management does not pressure staff to bypass technical controls during a production emergency.

Current Developments and Best Practices

Important developments include additive manufacturing for selected obsolete or low-volume parts, condition-based replenishment linked to asset-health indicators, digital twins for scenario analysis, and collaborative inventory agreements across sites. Additive manufacturing can shorten supply time but requires control of material properties, qualification, dimensional accuracy, and regulatory acceptance.

Best practice is characterised by the following principles:

• classify by criticality as well as value and demand;
• maintain a governed, searchable, duplicate-free part master;
• connect asset bills of material with work-order and inventory data;
• use differentiated service levels and replenishment policies;
• distinguish genuine demand from stock transfers and data corrections;
• manage repairables through identity, condition, and turnaround visibility;
• standardise parts where technically and economically sound;
• qualify suppliers and monitor lead-time reliability, not only price;
• pool inventory when interchangeability and response time justify it;
• test resilience against common-cause disruptions; and
• review parameters after asset, process, supplier, or maintenance-strategy changes.

Chapter Summary

Spare-parts and MRO inventory management is a reliability and service problem as much as an inventory problem. Demand is often intermittent, consequences are uneven, and supply risks can be severe. Effective management combines criticality, ABC value analysis, demand behaviour, and supply-risk classification. Forecasts must be interpreted with engineering and asset information. Reorder points, safety stocks, and service levels should be differentiated according to operational consequence.

Repair-versus-replace decisions require life-cycle cost, turnaround, quality, and failure-risk analysis. Supplier contracts should address continuity, technical conformity, obsolescence, and emergency response. Pooling and standardisation can reduce inventory while improving availability, provided common-cause risk and response time are controlled. Digital systems provide value only when master data, governance, and process discipline are sound. The strongest programmes integrate maintenance, engineering, procurement, stores, operations, finance, and suppliers around an explicit maintenance service requirement.

Key Terms

MRO inventory: Materials used for maintenance, repair, and operations.

Spare part: A replacement component kept to restore equipment function.

Criticality: The consequence of a part or equipment failure being unresolved.

ABC analysis: Classification based on annual consumption value.

Intermittent demand: Demand separated by periods of zero usage.

Insurance stock: Inventory held against severe consequences and uncertain supply.

Repairable: An item that can be restored and returned to service.

Rotable: A repairable component routinely exchanged and refurbished.

Capital spare: A high-value, strategically important replacement item.

Service level: A specified measure of the probability or proportion of demand served.

Fill rate: The proportion of total demand supplied immediately from stock.

Cycle service level: The probability of avoiding stockout during a replenishment cycle.

Safety stock: Additional inventory held against demand or lead-time uncertainty.

Reorder point: Inventory position at which replenishment is triggered.

Inventory position: On-hand, on-order, and repairable stock less reservations and backorders.

Pooling: Sharing inventory among locations or assets.

Cannibalisation: Removing a usable component from one asset for another.

Obsolescence: Loss of usefulness or supportability of an inventory item.

Part master: The controlled database record describing inventory items.

Supply resilience: The ability to maintain or restore required supply after disruption.

Review Questions

Short Answer Questions

1. Why can a low-value spare part be operationally critical?

2. Distinguish cycle service level from fill rate.

3. What information should be included in inventory position?

4. State two reasons why spare-parts demand is difficult to forecast.

5. What is the purpose of a repairable-parts tracking loop?

Long Answer Questions

1. Explain how criticality, ABC, demand behaviour, and supply risk can be combined into an MRO classification architecture.

2. Discuss the major replenishment policies used for spare parts and the conditions under which each is appropriate.

3. Develop a repair-versus-replace decision framework for a high-value failed component.

4. Explain how supplier contracts and network design contribute to spare-parts resilience.

5. Examine the role of digital systems and master-data governance in MRO inventory performance.

Analytical / Application Questions

1. A plant has a low-cost sensor with a twelve-month supplier lead time and severe production consequences if unavailable. Design a classification and stocking policy, explaining your assumptions.

2. Two sites each hold slow-moving interchangeable pump cartridges, but transport between them takes eight hours. Analyse whether pooling is appropriate and identify the risks that must be controlled.

3. A repairable electronic module costs 40% of a new unit to repair, has a 70% probability of successful repair, and a failed repair would cause a two-day outage. Explain what additional information is required before deciding whether to repair or replace.

22. Multiple Choice Questions with Answers

Question 1

Which classification dimension most directly captures the consequence of a stockout?

A. Annual consumption value

B. Criticality

C. Order quantity

D. Supplier discount

Correct Answer: B. Criticality

Question 2

For an intermittent-demand spare, a forecast should primarily distinguish between:

A. Purchase price and transport cost

B. Demand occurrence probability and demand quantity

C. Warehouse size and labour cost

D. Asset age and accounting value

Correct Answer: B. Demand occurrence probability and demand quantity

Question 3

Which expression best represents a reorder point?

A. Unit price plus order cost

B. Annual demand divided by lead time

C. Expected lead-time demand plus safety stock

D. Maximum stock minus obsolete stock

Correct Answer: C. Expected lead-time demand plus safety stock

Question 4

A fill rate measures:

A. The probability of no stockout in one cycle

B. The proportion of demand supplied immediately

C. The percentage of suppliers audited

D. The proportion of parts sent for repair

Correct Answer: B. The proportion of demand supplied immediately

Question 5

Which situation most strongly supports pooling spare parts across sites?

A. Parts are non-interchangeable and failures are perfectly correlated

B. Parts are interchangeable, demand is partly independent, and transport is acceptable

C. Each site has a unique specification and urgent transport is impossible

D. The pooled item has no approved technical standard

Correct Answer: B. Parts are interchangeable, demand is partly independent, and transport is acceptable

Question 6

Why is ABC analysis insufficient as the sole MRO classification method?

A. It ignores annual consumption value

B. It cannot rank any inventory items

C. It may classify a low-value but operationally critical item as low priority

D. It applies only to repairable items

Correct Answer: C. It may classify a low-value but operationally critical item as low priority

Question 7

Which factor would generally favour repair rather than replacement?

A. Uncertain repair quality and obsolete technology

B. Short replacement lead time and low replacement cost

C. Reliable repair process and long replacement lead time

D. Absence of repair records

Correct Answer: C. Reliable repair process and long replacement lead time

Question 8

What is a major risk of cannibalising one asset to restore another?

A. It always reduces technical risk

B. It may transfer the failure risk to the donor asset

C. It eliminates the need for asset records

D. It guarantees lower life-cycle cost

Correct Answer: B. It may transfer the failure risk to the donor asset

Question 9

Which data problem is most likely to cause duplicate purchasing?

A. A controlled part master with standard descriptions

B. Accurate repair-status tracking

C. Duplicate item codes and inconsistent units of measure

D. Validated equipment bills of material

Correct Answer: C. Duplicate item codes and inconsistent units of measure

Question 10

A supplier contract for a critical electronic spare should ideally include:

A. Only the lowest purchase price

B. No change-notification obligation

C. Obsolescence notification, technical conformity, warranty, and emergency response terms

D. An unrestricted right to substitute components

Correct Answer: C. Obsolescence notification, technical conformity, warranty, and emergency response terms