Spare Parts Inventory in Indian Manufacturing: Cutting Stock Without Cutting Uptime

The wrong target is the lowest inventory value

Maintenance spares sit between two costs that appear in different reports. Stock value appears in working capital; a missing part appears later as lost production, overtime, expedited freight and a longer repair. Optimising one without the other simply moves cost.

India's Annual Survey of Industries makes the accounting side explicit. MoSPI defines physical working capital as total inventories including raw materials and components, fuels and lubricants, “spares, stores and others”, semi-finished goods and finished goods. Maintenance stock therefore competes for the same capital as material that can be converted into saleable output. But a stores-reduction target applied evenly across all items is unsafe: one low-value seal and one high-value gearbox do not carry proportionate stockout consequences.

The right objective is:

Minimise the total cost of holding, ordering, stockouts and obsolescence while meeting the service level required by each asset risk.

That requires an item-by-asset view. Purchase value alone cannot supply it.

Why MRO demand behaves differently

Production material demand follows a bill of materials and a production plan. Maintenance, repair and operations demand often does not. Many spares have long periods of zero usage interrupted by one or two issues. A critical insurance spare may have no consumption for ten years and still be correctly stocked. Another part may show frequent issues only because technicians withdraw more than they use and never return the balance.

Five characteristics make MRO inventory difficult:

• Intermittent demand. Average monthly usage hides the spacing and size of rare demands.
• Coupled risk. Need arises from asset failure, planned maintenance or inspection findings rather than sales volume alone.
• Long or uncertain replenishment. The observed lead time includes specification, quotation, approval, manufacture, import or domestic movement, inspection and receipt—not merely the supplier's quoted dispatch time.
• Asset specificity. A part may fit one installed model, a family of models or nothing still operating.
• Repair loops. Motors, drives, gearboxes, valves, electronic cards and assemblies may circulate through issue, removal, repair, testing and return rather than being consumed.

Traditional ABC analysis sorts annual consumption value: unit cost multiplied by annual usage. It is useful for financial control, but dangerous as the stocking policy. A ₹500 component that stops a bottleneck line can be C-class by value and critical by consequence. A ₹10 lakh assembly with redundancy and a reliable repair exchange may be A-class by value but need no plant-level spare.

Start with a clean item-to-asset relationship

Before changing min/max levels, find out what each material record actually represents. Duplicate item masters are common where plants, projects and buyers create descriptions independently: “BRG 6205”, “6205 ZZ”, a manufacturer number and a supplier number may describe the same base item—or materially different clearances and seals.

For each item, retain:

• a controlled noun-modifier description;
• manufacturer and manufacturer part number;
• technical attributes that decide interchangeability;
• approved alternatives and supersessions;
• units of issue and purchase, with conversion;
• linked asset classes, models and bills of material;
• stocked, non-stocked, special-order or repairable status;
• each storeroom, bin and usable balance;
• shelf life and storage conditions where applicable;
• preferred and alternate suppliers;
• actual lead-time observations, not only the catalogue value;
• issue, return, reservation, repair and stockout history.

Do not merge records on similar descriptions alone. A bearing's internal clearance, seal, cage, tolerance or manufacturer restriction can matter. Engineering must approve equivalence, and the old identifier should remain as an alias so historical issues still aggregate.

Asset bills of material need the same discipline. An OEM manual lists what an asset was designed with; the as-maintained BOM records what is installed now. Modifications, local substitutions and control-system upgrades make the difference important. A spare linked to a retired configuration is not protection.

Segment on four dimensions

Every stocking decision should combine four questions.

1. What is the consequence of not having it?

Use the consequence of the stockout during a credible demand, not the general criticality of the parent asset. Ask whether an alternative part, redundant train, temporary repair, cannibalisation route or safe operating workaround exists.

Class	Stockout consequence	Typical policy
Vital	Immediate safety, environmental or bottleneck-production exposure with no adequate workaround	High service target; named ownership; preservation and readiness checks
Essential	Material production or quality loss, but time-limited workaround or redundancy exists	Risk-based safety stock and monitored lead time
Routine	Limited consequence; substitute or rapid supply is available	Low stock or non-stocked procurement

Criticality must be approved jointly by maintenance, operations, engineering and stores. A requester marking everything “critical” transfers no information.

2. How does demand occur?

Measure both demand frequency and quantity variability. Average demand interval (ADI) helps distinguish regular from intermittent issues; squared coefficient of variation (CV²) describes variability in non-zero demand sizes. The precise thresholds are a modelling choice, but the resulting behaviours are useful:

Demand behaviour	What it looks like	Appropriate response
Smooth	Frequent issues, stable quantities	Conventional forecast and reorder point
Erratic	Frequent issues, variable quantities	Higher attention to demand variability and work plans
Intermittent	Long zero-demand periods, similar issue sizes	Intermittent-demand method; avoid simple monthly average
Lumpy	Long gaps and variable quantities	Scenario or risk-based stocking; manual engineering review

Use demand events, not transaction lines. One job issued from two bins and partly returned can create several transactions for one requirement. Conversely, a bulk issue to a technician's locker can hide months of real demand.

3. How risky is replenishment?

Record the full lead-time distribution. The average alone does not protect against variability. Separate:

• internal specification and approval time;
• supplier quotation and order confirmation;
• manufacture or supplier processing;
• transit and clearance where applicable;
• incoming quality inspection;
• repair turnaround for repairables.

Measure promised and actual dates by supplier and part family. A nominal 45-day item delivered between 30 and 140 days is not a 45-day risk. Also record minimum order quantity, pack size, price breaks, expiry and whether the supplier still supports the installed model.

4. Is it consumed, repaired or insured?

Classify the material flow:

• consumable: used and not economically recovered;
• repairable or rotating: removed unit returns through a repair loop;
• insurance spare: rarely demanded, held for high consequence and long replacement time;
• project or shutdown material: bought for a named scope and date;
• standard reusable item: shared across many assets or plants.

Applying one min/max formula to all five creates both excess and shortage.

Reorder points that reflect reality

For regular consumables, the basic relationship is:

Reorder point = expected demand during replenishment lead time + safety stock.

The inventory position that triggers an order should normally include usable on-hand stock plus open orders minus reservations and backorders. Counting quarantined, expired, damaged or already reserved stock as available creates a false service level.

IBM Maximo's documentation describes the same core control: a reorder point is intended to keep the balance from falling below safety stock during lead time, and reorder calculation uses safety stock, lead time and economic order quantity. The formula is only as good as those fields. If lead time has not been updated since commissioning and issues bypass work orders, automatic replenishment automates the error.

Set service levels by consequence segment. Higher service requires more buffer, but do not translate “critical” into an arbitrary six months of usage. For intermittent vital spares, a simulation using observed failure demand and lead-time variability is often more honest than a normal-distribution safety-stock formula.

Review ordering constraints after the theoretical result:

• round to the supplier pack or minimum quantity;
• respect shelf life and storage capacity;
• include hard reservations for approved shutdowns and planned jobs;
• subtract confirmed parts recoverable through repair before need date;
• consider a lateral transfer from another plant before external purchase;
• expose the override and its approver.

Insurance spares need an economic case

An insurance spare may be correct despite zero historical usage. Its case depends on the probability of demand during the replenishment horizon and the consequence if demand occurs without stock.

Test these questions:

1. Is the failure mode credible, and does it require full replacement?
2. Can the asset operate degraded, switch to redundancy or accept a temporary repair?
3. Can another installed unit be cannibalised safely, and how long would that take?
4. Is a supplier exchange, repair contract or shared spare available?
5. What is the verified time from failure to a serviceable unit on site?
6. Will the spare remain compatible after planned upgrades?
7. What preservation, testing and controlled storage does it require?

Compare annualised holding and preservation cost with expected avoided loss, but show the uncertainty. Low failure probability multiplied by enormous production loss can generate a misleadingly neat number. Record the decision, assumptions, accountable owner and next review trigger.

Insurance spares also fail while stored. Shafts brinell, elastomers age, batteries discharge, moisture enters windings, firmware becomes incompatible and electronic components become obsolete. A spare that has never been inspected is inventory value, not assured availability.

Repairables: control the loop before buying more

For repairable items, the relevant balance is not only serviceable stock. It is:

Serviceable on hand + units under repair + recoverable failed units − reservations − condemned units.

A plant may buy an additional motor because the store shows zero while two repairable cores sit untagged in a workshop and a third is overdue from a vendor. Give every serialized or rotating unit its own identity and status: installed, serviceable, reserved, in transit, awaiting inspection, under repair, awaiting test or condemned.

Set the pool size from simultaneous demand risk and repair turnaround, then attack turnaround before adding stock. Common delay sits before the purchase order, after the repaired unit returns, or in an unclear repair specification—not on the repair bench itself.

Repair history must record fault found, work performed, components changed, test results, warranty and cost. Repeated repair of a unit that returns quickly to the workshop is not inventory availability.

Pool across plants selectively

Multi-plant manufacturers can remove duplicate insurance stock by holding a shared spare, but only where transfer time is shorter than the tolerable outage and the item is genuinely interchangeable.

Before pooling, confirm:

• common technical specification and installed interfaces;
• transport method, lifting, preservation and packaging;
• 24/7 release authority and transfer documentation;
• ownership and cost-transfer rules;
• the risk of simultaneous demand across sites;
• a tested dispatch route and realistic door-to-door time;
• reservation rules for planned outages.

Fast-moving consumables often belong close to use. Large, rare and portable insurance spares offer the strongest pooling case. A dashboard showing another plant's balance is not a sharing arrangement unless someone can release and move it at 02:00.

Planned work should create visible future demand

Inventory optimisation fails when the material forecast sees only past issues. Approved maintenance plans, shutdown scopes, campaigns, projects and known modifications are future demand and should create time-phased reservations.

Use hard reservations for approved work with a required date and soft reservations for likely but uncommitted demand. IBM distinguishes the two on this basis. Protecting hard reservations prevents a routine issue from consuming a shutdown-critical part, while including them in replenishment gives procurement time to act.

Return unused material promptly against the work order. Record the reason: job cancelled, quantity overestimated, wrong part, reusable balance or scope change. Chronic returns reveal poor job plans and inflate apparent demand if the issue is counted without the reversal.

A safe stock-reduction sequence

Do not begin with a percentage target. Use a controlled sequence:

1. Freeze uncontrolled item creation. Route new records through duplicate and specification checks.
2. Clean balances. Separate usable, quarantined, expired, damaged, reserved and consignment stock.
3. Link items to active assets and work history. Flag stock supporting retired or modified equipment.
4. Segment criticality, demand, replenishment and flow type. Assign an owner to exceptions.
5. Correct lead times and units. Use actual receipts and repair cycles.
6. Recalculate policy. Reorder point, safety stock, order quantity or non-stocked status by segment.
7. Challenge excess safely. Cancel open orders, transfer, return to supplier, redeploy, sell or dispose through authorised routes.
8. Pilot by plant or commodity. Watch service and downtime before scaling.
9. Review periodically and on change. Asset retirement, supplier change, BOM modification and failure are triggers.

Never dispose of a slow-moving part until engineering confirms applicability and consequence. “No issue in 36 months” may describe obsolete stock—or a valid insurance spare that has performed exactly as intended.

Measure cash and service together

Measure	What it reveals	Important control
Inventory value	Capital held at standard or moving-average cost	Split usable, repairable, excess and obsolete
Service level	Demands fulfilled completely by required time	Measure by criticality, not one blended percentage
Stockout events	Unavailable part when a valid demand arose	Link lost hours and workaround
Inventory turns	Annual issues relative to average inventory	Do not use as the target for insurance spares
Excess and obsolete value	Stock above policy or with no supported application	Require engineering disposition
Emergency purchase rate	Requisitions raised outside normal planning	Separate genuine failure from planning failure
Repair turnaround	Removal to tested serviceable return	Show waiting time by stage
Record accuracy	System balance matching verified physical stock	Report by criticality and value
Reservation fulfilment	Planned jobs with parts ready by need date	Prevents stock cuts from damaging schedule compliance

Pair every cash-release claim with stockouts, maintenance delay and lost-production impact. If inventory falls while emergency purchases and cannibalisation rise, the business has not saved money; it has changed where the cost is recorded.

India-specific controls that matter

Indian manufacturing ranges from single-site businesses to multi-state groups with central procurement and diverse installed equipment. The method must work at the operating level rather than rely on one nationwide parameter set.

• Preserve GST-relevant purchase and transfer records in the financial system; keep technical applicability and maintenance reservations in the EAM, with controlled identifiers between them.
• Measure actual supplier performance by ship-from location and receiving plant. A vendor's performance to Pune does not establish the same replenishment risk to every site.
• Maintain approved alternatives carefully where local manufacture can shorten lead time. Qualification must cover technical fit, material, performance, warranty and required inspection—not only matching dimensions.
• For standards-dependent parts, verify the current applicable Indian Standard and amendments through the Bureau of Indian Standards' Know Your Standard service rather than copying an old specification from a purchase order.
• Protect shutdown demand from routine consumption and late scope growth. Time-phased reservations and a controlled material-freeze date are more useful than filling a separate project warehouse “just in case”.

FAQ

What is a good target for reducing spare-parts inventory? There is no defensible universal percentage. Establish current service and risk, correct data, then quantify excess above approved policy by item. A blanket target invites removal of cheap critical parts and retention of expensive obsolete ones.

Should an item with no usage for three years be disposed of? Not without checking asset applicability, failure consequence, replenishment time, alternatives and insurance-spare status. Zero usage is a review trigger, not a disposal decision.

Is ABC analysis enough? No. ABC controls annual consumption value. Add criticality, demand behaviour, lead-time risk and repairability before setting service levels or stocking policy.

How often should reorder points be updated? On a defined cycle and after material changes in demand, lead time, supplier, asset population, maintenance plan or pack size. Vital items need more frequent exception review than routine items.

Can AI forecast spare-parts demand? It can help identify intermittent patterns, anomalies and policy options where history is adequate. It cannot infer an unrecorded asset modification, decide whether two parts are technically interchangeable or repair missing issue and return discipline.

Where a system helps

The useful system is not a digital stock ledger alone. It connects each part to the assets it protects, the work that created demand, the reservation that protects planned work, the supplier and repair lead time, and the downtime consequence of a stockout. That lineage makes it possible to release working capital item by item while proving that uptime risk remains controlled. See eAMS for asset management.

Related reading: What Is a CMMS? (KB-129) and The Work Order Lifecycle: Nine Statuses, and the Two Where Data Quality Dies (KB-500).

Sources

• MoSPI, Annual Survey of Industries 2022–23, Volume I
• Bureau of Indian Standards, Know Your Standard
• IBM Maximo Manage, Inventory module
• IBM Maximo Manage, Reordering inventory items
• ISO 14224:2016, Collection and exchange of reliability and maintenance data for equipment