Plan a Last-Time Buy for Discontinued Spare Parts

A last-time buy is not simply annual usage multiplied by the years an aging fleet might remain in service. The decision begins with an authenticated discontinuation notice, then combines remaining asset exposure, usable stock, repair and alternative options, and low, base, and high demand scenarios. The result is a documented range with residual risks, not a guaranteed forecast.

The broader discussion of parts for an aging fleet covers general fleet support. This workflow applies when a specific orderable part is being discontinued. Nothing on KTSU's public site establishes a last-time-buy program, discontinuation notice, available stock, replacement approval, or support commitment.

Authenticate the notice and its deadlines

Obtain the notice from the manufacturer or another source whose authority can be verified. Record the issuer, notice or product-discontinuation number, revision, publication and effective dates, and contact. Map every affected orderable part number, revision, package or configuration. A family name alone may hide unaffected variants or leave a superseded number in the plan.

Keep dates in separate fields: last order or last-time-buy date, last date to change or cancel an order when stated, final production date, last shipment or delivery date, and any later repair or spare-support period. Record minimum quantities, pack sizes, price conditions, non-cancellable or non-returnable terms, and allocation or capacity limits only when the actual notice or quote states them.

A current NXP product discontinuation notice illustrates why this precision matters: it distinguishes notice, last-time-buy, and last-delivery dates and lists affected orderable numbers. It also states that a listed replacement is not a warranty of equal form, fit, function, or performance until the customer qualifies it. This semiconductor notice is an example of notice structure; its dates and commercial terms do not apply to mechanical parts.

The DEIF product life-cycle page provides another manufacturer-specific example of final-order, final-delivery, later spare-parts phases, replacement statements, and non-cancellable terms. Do not transplant its support periods or policy. Each supplier's current notice and acknowledgment control its own offer.

Monitor the notice source for revisions and obtain supplier acknowledgment of the exact order. An internal deadline needs enough time for engineering, maintenance, quality, procurement, finance, and legal or regulatory review where applicable. A revised notice or rejected order can change the feasible range even after initial approval.

Define the remaining service exposure

Build the horizon from assets and obligations, not a fixed number of years. List installed machines by model, serial range, configuration, location, and status: active, stored, scheduled for retirement, or decommissioned. Record the quantity and position of the affected part per asset, planned refresh dates, and any contract, warranty, regulatory, or mission support obligation that an authorized reviewer confirms.

Separate consumable, repairable, and condemnation behavior. Use dated issues and installations, removals, repair inductions, successful repairs, scrapped units, and turnaround times. Zero-demand periods are observations, not proof that future demand is zero. Cannibalized or reclaimed parts need identity, condition, approval, and yield records before they can reduce planned supply.

Reconcile current usable stock, quarantine, damaged or expired stock, in-transit material, open orders, repair work in process, and shared pools. Do not count the same unit as both repairable and usable. Assign owners to data gaps and conflicts. A site-wide annual average can mislead when only some configurations use the discontinued part.

The March 2024 SD-22 DMSMS guidebook treats availability as a changing snapshot and compares multiple resolution paths with engineering, logistics, risk, cost, storage, and assembly-level considerations. It is U.S. defense guidance, not a civilian contract requirement or a quantity formula for KTSU parts.

A declining installed base can reduce exposure while aging, poor repair yield, longer turnaround, or a campaign can increase it. Historical demand alone cannot decide which effect dominates. Record the evidence date, data owner, missing periods, exceptional bulk events, and the reason each assumption enters a scenario.

Compare alternatives before sizing the final order

Treat a life-of-need purchase as one option among approved inventory, remaining authorized supply, a manufacturer-listed candidate, second-source development, repair, refurbishment, reclamation, a higher-assembly replacement, redesign, production extension, or accelerated retirement. Compare schedule, technical work, cost, evidence, and residual risk for each path.

A cross-reference or notice-listed replacement remains an engineering candidate. The responsible authority must examine form, fit, function, performance, interfaces, machine model and serial range, software or controller interaction where relevant, and the drawing and documentation changes required. Qualification, test, inspection, and field-introduction evidence must match the actual application.

The official IEC 62402:2019 page describes an obsolescence-management process spanning planning, resolution selection and implementation, and improvement across an item's life cycle. The detailed standard is not freely reproduced here, and its scope does not supply a universal last-time-buy formula or establish compliance by KTSU.

For residual-market supply, verify source authorization, traceability, condition, storage history, revision, inspection, and counterfeit controls appropriate to the product. Do not call a source authorized or a part genuine without evidence. For repair and reclamation, use observed yield, turnaround, condemnation criteria, and capacity. For an alternative still under qualification, treat completion date and probability as assumptions and consider a bridge quantity rather than counting future approval as certain.

Build low, base and high scenarios

Use the same transparent calculation structure across three scenarios while changing the assumptions that truly drive exposure. Each scenario should identify the horizon, installed-base trajectory, usage or failure pattern, condemnation rate, repair yield and turnaround, planned campaigns, current usable stock, open supply, stored-stock attrition, alternate-qualification timing, pack size, minimum order, and capacity constraint.

Scenario	Conditional assumptions	Decision use
Low	Faster fleet retirement, stable repair yield, timely approved alternative, and reconciled usable stock	Shows surplus exposure if support demand falls sooner
Base	Current approved retirement plan, central failure and repair assumptions, and expected alternative milestone	Provides the working case, not a promise
High	Longer fleet retention, lower repair yield, later alternative approval, or a defined campaign	Shows shortage consequence and capacity pressure

A 2026 open-access Scientific Reports study on automotive end-of-life demand demonstrates a data-rich approach using consumption history, installed-base attrition, warranty or failure information, and uncertainty. The authors also identify dataset and model limitations. Its accuracy figures, decay parameters, horizon, and buffers cannot be transferred to undercarriage parts without equivalent data and validation.

Show the range and its sensitivity drivers. Recalculate when fleet retirement accelerates, repair yield falls, demand includes a one-time campaign, or the alternate qualifies later. Where data permit, test the method against an earlier period. Preserve the calculation owner, version, inputs, exclusions, and unit conversions. Do not convert missing data to zero or multiply one noisy annual average by an arbitrary horizon.

Shortage and surplus consequences should both be visible. A critical shortage can idle equipment or force a redesign under time pressure. Excess stock can tie up capital, degrade, lose traceability, or become stranded after fleet retirement. The chosen range reflects an approved risk policy; the scenario method alone cannot choose it.

Test commercial, quality and storage feasibility

Ask the supplier to confirm whether it can accept the proposed order, production lots, schedule, final shipment, packaging, and quality evidence. Review the drawing and specification revision, material and process certificates, traceability, inspection and acceptance plan, first- or final-production risks, discrepancy process, and receiving capacity. A quantity that exists in a spreadsheet may not be producible within the window.

Review price, currency, taxes, duties, freight, payment, cancellation, non-returnable terms, ownership, and risk transfer with the responsible commercial and legal teams. Record the warranty start and end basis from the actual offer. Do not assume a manufacturer example defines any of these conditions for the proposed order.

Assess how packaging, seals, coatings, rubber, lubricants, corrosion protection, and records will remain serviceable for the horizon. Specify the approved storage environment, inspections, preservation renewal, stock rotation, access, security, insurance, and retrieval process from verified technical and contractual information. Do not invent a shelf life. Consider lot diversification and the risk that a single final production lot contains a common problem.

Include warehouse space, handling, periodic inspection, finance cost, insurance, deterioration, disposal, and documentation maintenance in the comparison. Assign an owner for receiving discrepancies and quarantine. If the preservation or traceability plan cannot support the intended horizon, reduce the assumption or resolve the gap before treating the quantity as feasible.

Approve the decision and set review triggers

Record the decision date and version, chosen range and scenario, selected mix of purchase, repair, alternative and retirement actions, and the supporting evidence. Obtain the actual approvals required from maintenance, engineering, procurement, quality, finance, and legal or regulatory reviewers. Show funding and purchase-order status separately from analytical approval.

List open assumptions and residual shortage or surplus risk. A decision may remain conditional until the supplier accepts the order or engineering approves an alternative. Maintain enough buffer before the stated cutoff to resolve those conditions, but derive that buffer from the case rather than copying a vendor policy.

Set review triggers: a revised notice, supplier capacity change, order rejection, changed retirement plan, demand or failure shift, repair-yield change, inventory discrepancy, alternate milestone, cost change, or storage problem. Assign a cadence and escalation owner, and compare later outcomes with the scenario inputs. Approval is a controlled snapshot, not a permanent truth.

Use the scenario register to expose three states

Decision field	Normal	Missing	Conflict
Notice and dates	Issuer, ID, revision, affected parts, order and shipment dates verified	Order cutoff or affected revision cannot be confirmed	Supplier messages or notice revisions give different dates
Exposure and stock	Installed base, horizon, demand, repair yield, usable stock and open supply reconcile	Configuration, usable inventory, failure or repair data unavailable	Warehouse, ERP and repair records count the same units differently
Alternative	Candidate status, evidence, owner and milestone are explicit	No qualification plan or approval authority	Notice lists a replacement that engineering has rejected for the application
Scenario and feasibility	Low/base/high inputs, range, commercial terms and storage controls are documented	Capacity, preservation, cost or sensitivity driver unresolved	Approved quantity assumes an alternate date that the project plan contradicts
Decision and review	Approvals, PO status, residual risk and triggers have owners	Funding, supplier acceptance or closure owner absent	Decision is marked final while a required approval remains open

A normal record supports a conditional range tied to its evidence. A missing record cannot yield a precise quantity. A conflict must remain visible until the authorized owner resolves it. The register does not issue an automatic buy or no-buy result; it makes the decision, uncertainty, alternatives, and next evidence reviewable before the opportunity closes.