The Investment That Does Not Improve Reliability

Over the last decade, the CMMS market grew from USD 0.8 billion to over USD 3.2 billion globally. Major CMMS and EAM platforms have been implemented in virtually every serious mining operation. And yet, in the 40+ fleet audits Structured reliability assessments have identified across four continents, average fleet mechanical availability at the time of first engagement is 74–78% — against an industry target of 88%.

The correlation is uncomfortable: the more sophisticated the CMMS, the more confident the organisation is that it has a reliability system. That confidence is, in most cases, misplaced.

What a CMMS Actually Does — and Does Not Do

To be clear: a well-implemented CMMS is valuable. It creates the data that a reliability system needs. But it does not, by itself, constitute a reliability management capability. The distinction matters because most organisations believe their CMMS investment has solved the problem it was intended to solve.

Five Things Your CMMS Cannot Do

The Financial Consequence of the Confusion

When an organisation believes its CMMS is a reliability system, it stops looking for the reliability capability it does not have. The consequences are quantifiable:

In a 50-unit mining fleet, these metrics translate to an annual reliability-related financial exposure of $28–60 million — of which 60–65% is preventable through structured reliability governance. The CMMS will record every failure event in that exposure. It will not prevent a single one.

What a Reliability System Actually Looks Like

A reliability management system has five components that a CMMS supports but does not constitute:

The Operational Relevance

For Maintenance Managers: the next time your CMMS shows a recurring failure mode on the same component for the third time in 12 months, ask this question — where in our system is the person whose job it is to make sure this never happens a fourth time? If the answer is "there isn't one," your CMMS is performing its function correctly. Your reliability system is absent.

For Mine Directors and CFOs: the question is not "do we have a CMMS?" The question is "do we have a reliability engineer, a structured RCA process, a condition monitoring programme with a threshold-to-WO workflow, and a monthly governance review that connects these to financial performance?" If any of those four elements is absent, the CMMS investment is recording a problem that the organisation has not yet mobilised to solve.

Why Every Audit Looks the Same

Every new mining engagement begins with the same expectation from the client: "Our situation is unique." And in some operational details — ore type, geography, fleet composition — it is. But when MitWin applies the MRMM (Maintenance Reliability Maturity Model) diagnostic to the maintenance and reliability governance structure, the findings converge with striking consistency.

The average MRMM score at first engagement: 38/80 — firmly in what MitWin classifies as the Fragile zone, with 40–60% probability of a major production-stopping event within 6 months without intervention.

The reason for this consistency is that reliability underperformance is not a technical problem. It is a governance problem. And governance problems repeat themselves because the same five structural failures recur — regardless of how sophisticated the technical environment appears from the outside.

Failure 1 — The Reliability Engineer Vacancy

The single most consistent finding in every structured reliability assessment: there is no formally defined Reliability Engineer role. In a significant proportion of operations assessed, the RE function either does not exist or exists in name only — occupied by someone spending less than 30% of their time on reliability engineering activities.

The consequence is precise and quantifiable. Without a functioning RE role:

No Failure Mode and Effects Analysis (FMEA) is conducted. No Root Cause Analysis (RCA) closes the loop on recurring failures. No Condition-Based Maintenance programme is designed with threshold-to-WO workflow. No MTBF trend analysis is performed at the equipment-class level. The result: an operation that repairs failures it has already repaired, and will repair again.

The correct RE hire — role-designed to MitWin specification, with structured onboarding and authority to act — consistently delivers measurable MTBF improvement within 4–6 months. The financial return is not marginal. In a 50-unit mining fleet, the RE role pays for itself in the first 2–4 days of Year 1 operation.

Failure 2 — The Planning Deficit

Every audit finds the same planning structure: one Maintenance Planner — or more commonly, a supervisor doubling as a planner — managing 35–55 maintenance units simultaneously. The industry benchmark for effective planning is one planner per 20–25 units. Above that ratio, the planner is not planning. They are processing reactive work orders and managing parts emergencies.

The Compounding Effect

Planning underresourcing creates a self-reinforcing failure cycle. When planning capacity is insufficient, the 4-week rolling schedule degrades to a 2–3 day reactive queue. When the schedule degrades, equipment release windows cannot be coordinated with operations. When equipment release fails, planned maintenance is displaced by reactive response. When planned maintenance is displaced, failures occur that the planned maintenance would have prevented — which consumes more reactive capacity.

This cycle is well-documented across asset-intensive operations. The average schedule compliance in a planning-deficit organisation: 47–58%. The annual reactive maintenance premium from this displacement: $3–8 million in a 50-unit fleet at a 2.8× reactive cost multiplier.

Failure 3 — Condition Monitoring That Does Not Monitor

This failure is the most operationally acute because its consequences are invisible until they are catastrophic. Most operations of any sophistication have an oil analysis programme in place. A small number have vibration monitoring. But in the majority of cases, the programme exists in name but not in function.

The definition of a functioning CBM programme: every sample result is reviewed by a qualified analyst within 24 hours of receipt, every threshold breach generates a work order within 48 hours, and the threshold-to-WO workflow is configured in the CMMS. In practice, MitWin finds:

The operational consequence: CBM programme costs are being incurred — lab fees, sampling labour, equipment — but the intelligence generated is not reaching the maintenance system. Every unactioned threshold breach is a developing failure that will eventually become an unplanned breakdown event.

Failure 4 — Governance Without Accountability

The fourth failure is the one that most directly implicates senior leadership. Almost every operation has a weekly maintenance meeting. Most have a monthly maintenance report. But in fewer than 20% of operations does this governance cadence constitute reliability governance — because it lacks the two essential elements: financial framing and named accountability.

A weekly maintenance meeting that reviews which machines broke down and who is fixing them is a reactive coordination session. A monthly maintenance report that lists WO counts and schedule compliance percentages is an activity report. Neither constitutes governance.

Reliability governance has three distinguishing characteristics:

1. Financial framing at every meeting. The first number presented is the production loss from downtime in dollar terms — not hours. Not percentage. Dollars. The Mine Director must understand what reliability performance cost the operation in the last 30 days in revenue terms before any operational detail is discussed.

2. Named KPI ownership. Every reliability KPI — MTBF, availability, schedule compliance, recurring failure rate — has a named human owner who is accountable for performance and escalation when below target. Shared ownership is no ownership.

3. Decision-forcing structure. Every governance meeting closes with decisions, not discussions. "We noted the issue" is not governance. "We agreed the following action, owned by this person, due by this date" is governance.

Failure 5 — The Strategy-Reality Disconnect

The fifth failure is the subtlest and the most expensive. Most operations have a maintenance strategy — typically an OEM-derived PM task library implemented in the CMMS. Many have completed a strategy review at some point. But in the vast majority of cases, the strategy the CMMS is executing was not designed from failure mode analysis, has not been validated against actual failure data, and has not been updated to reflect the specific operating conditions of the site.

The PM interval for an air filter on a Komatsu HD785 haul truck at OEM default is 500 hours. At a site in Northern Nigeria operating in iron-oxide laterite dust with 85% relative humidity, the effective filter life is 170–180 hours. Applying the OEM interval produces a structurally predictable engine failure every 14 months — which is exactly what the CMMS records, every 14 months, as an unplanned event.

The Three Strategy Failures

The Leadership Takeaway

These five failures — RE vacancy, planning deficit, non-functioning CBM, governance without accountability, and strategy-reality disconnect — do not require new equipment, new CMMS systems, or capital investment to address. They require structural and governance decisions that sit squarely within the authority of the Mine Director, COO, and CFO.

The financial consequence of not addressing them is measurable and large. The financial consequence of addressing them — in sequence, with structured methodology — is larger still. In MitWin's engagement data, operations that correct all five failures within 18 months achieve an average of 54% MTBF improvement and 22% maintenance cost reduction against pre-engagement baselines.

The Business Situation

The client — a West African open-cut gold producer operating a 46-unit fleet at a single mine site — approached MitWin following the departure of its third Maintenance Manager in four years. Fleet mechanical availability had declined from 83.4% to 74.5% over 24 months. Maintenance cost per operating hour on the excavator class had increased from USD 112/hr to USD 148/hr over the same period. The operation was producing at 78% of licensed capacity.

The board had commissioned two prior audits — one by a Big 4 advisory firm and one by a specialist mining consultant. Both had produced detailed findings reports. Neither had produced operational change. When MitWin began the engagement, a Maintenance Superintendent showed us a binder containing both prior reports. "The recommendations are good," he said. "We just never implemented them." This single observation defined the engagement objective.

What the MitWin S1 Audit Found

The Fleet Stability & Cost Risk Audit (S1) was completed in 15 working days. The headline findings:

The MRMM (Reliability Maturity Model) scored the operation at 38/80 — with D6 (Failure Elimination) scoring 1/8 and D1 (Reliability Governance) scoring 2/8. There was no Reliability Engineer. One planner managed 46 units. The supervisor-to-technician ratio was 1:19. Oil analysis results were reviewed by the OEM dealer, not by the internal team — and in 3 of the 12 months reviewed, results showing threshold breaches had not triggered any work order.

The Top 5 Recurring Failure Modes

The MitWin Intervention — 90-Day Transformation Programme

The Reliability Transformation Programme (S3) was approved at the S1 presentation. The programme ran for 13 weeks across 5 phases:

Results at Day 90

Year 1 Financial Impact

Against a programme investment of a material amount (S1 Audit + S3 Transformation Programme + S4 Inventory Optimisation initiated in Month 2), the Year 1 return was Strong programme return.

The S6 Governance Partnership commenced at Month 4. At Month 12 — the Annual Review — fleet availability remained at 91.3%, MTBF had continued improving to 248 hours (excavator class), and the Maintenance Manager was running all three governance meetings independently. The Year 1 S6 Strong programme return on the $216,000 annual governance fee.

The Executive Lesson

This operation had been diagnosed correctly — twice — before MitWin arrived. The finding reports from the prior audits were accurate. The recommendations were appropriate. The reason nothing changed was not analytical failure. It was the absence of the structured execution framework and the external accountability mechanism that made change obligatory rather than aspirational.

Three elements distinguished this engagement from the prior two:

Outcome commitments before Day 1. Three specific KPI targets — MTBF +25%, schedule compliance >80%, recurring failure rate −60% — were signed by both parties before the programme began. This was not a deliverables contract. It was an outcomes contract. The distinction created a fundamentally different organisational posture.

The Mine Director in the monthly review. From Month 1, the Mine Director attended every Monthly Executive Reliability Review. Not as an observer — as the accountability principal. This single structural decision changed the weight of every action agreed in those meetings.

Governance before tools. The programme installed a governance system before — not after — any technology change. The Power BI dashboard was built in Week 8, after the governance cadence was established. The governance system is what makes the tools useful. Tools deployed into a governance vacuum become expensive dashboards that nobody acts on.

The Programme That Is Not Working

There is a specific type of CBM programme failure that is more common than the absence of any programme at all. The oil analysis samples are taken. The lab report arrives by email. The equipment costs continue to rise. And the next catastrophic engine failure occurs on a machine whose oil analysis data — had anyone reviewed it — was showing Si at 48 ppm against a threshold of 20 ppm, three samples in a row.

In 48% of the structured S1 assessment engagements, at least one oil analysis result containing a threshold breach is found in an unread email or unfiled report at the time of audit. In every case, the relevant component has either already failed or is actively developing toward functional failure. The data arrived in time. The governance did not.

What a Functioning CBM Programme Requires

The definition of a functioning oil analysis CBM programme has four non-negotiable elements. Most mining operations have one or two. The absence of any one of the four makes the programme non-functional — regardless of how good the lab results are.

The Three Unread Reports — An Illustrative Scenario

In this illustrative scenario, a West African gold operation in Q35, three oil analysis results were found in the Equipment Manager's inbox at the time of engagement — all unread, all containing threshold breaches:

Total cost of three unread emails: $1.84 million. Total cost of acting on them within 48 hours: approximately $28,000 in planned interventions. This is not a technology failure. The technology worked. It is a governance failure.

The Four-Step Protocol That Closes the Gap

Step 1 — Assign a named analyst. Every oil analysis result that arrives in any inbox must have a named person responsible for reviewing it against equipment-specific thresholds within 24 hours of receipt. If that person is on leave, a named backup must exist. "Reviewed by the lab" is not internal review. The lab provides data. The analyst provides interpretation.

Step 2 — Post the threshold table where it is used. Equipment-specific threshold tables should be posted in the workshop, in the stores, and available to supervisors digitally. The analysis is only as good as the reviewer's access to the reference data. A threshold table that exists only in the RE's head is not a governance system.

Step 3 — Configure the alert-to-WO trigger in the CMMS. Any threshold breach generates a WO automatically — flagged as "CBM Alert — Priority, Do Not Defer." No human decision required between threshold breach and WO creation. Human judgment is applied at the WO investigation stage — not at the alert stage.

Step 4 — Review CBM alert response at the weekly reliability meeting. Every week, one agenda item: how many CBM alerts were generated last week? How many WOs were created? How many investigations are in progress? This converts the CBM programme from a data collection activity into a governed reliability function.

The Leadership Question

For Mine Directors and CFOs: the question is not "do we have oil analysis?" The question is: "Who reviewed last month's results against equipment-specific thresholds — and what work orders did those results generate?"

If that question cannot be answered in under 30 seconds, the CBM programme is recording data that is not being converted into action. The SOS account is being paid. The lab is doing its job. The failure is in the governance between the result and the work order.

The Reporting Ceremony

A reporting ceremony has specific characteristics. Someone presents data — availability charts, work order counts, schedule compliance percentages, maintenance cost trends. The data is accurate. The presentation is professional. The Mine Director listens, asks a few clarifying questions, and the meeting concludes. Actions may be discussed. A follow-up may be suggested. And then the meeting ends — with no binding decision, no named owner for any commitment, and no mechanism to enforce the follow-up.

This is not an unusual meeting. It is the standard monthly maintenance meeting in the majority of mining operations where this pattern exists. It costs 90 minutes of the Mine Director's time, and it produces no governance.

The Three Tests

Any Mine Director can apply these three tests to their most recent monthly maintenance meeting to determine whether it constituted governance or ceremony:

The Most Important Structural Change

One structural change improves monthly governance meeting quality more than any other: the Mine Director reviews the action register from last month before any KPI data is presented.

This sequence — accountability before information — enforces the culture of the meeting before any new topic creates distraction. When the first question in the room is "what happened to the actions we agreed last month?" — and the answer must be given by named owners in front of the Mine Director — the quality of action commitments in the current meeting changes immediately.

This effect is consistently observed in asset-intensive operations. Where the Mine Director reviews the prior action register first, action completion rate averages 82%. Where the Maintenance Manager chairs the meeting alone and the action register is reviewed last (or not at all), action completion rate averages 34%. Same agenda. Same participants. Same data. Different authority in the room — and a different sequence.

Financial Framing Is Non-Negotiable

The reason most monthly maintenance meetings do not produce decisions is that the data is presented in operational terms — availability percentages, MTBF hours, work order counts — that do not create urgency in a Mine Director whose primary vocabulary is financial.

When the opening slide says: "Fleet availability was 74.5% last month against an 88% target — a gap that cost this operation $2.4 million in production value" — the Mine Director is in a different meeting than when the opening slide says: "Fleet availability was 74.5% against an 88% target."

The data is the same. The framing determines whether it produces a decision or a nod.

The Governance Meeting Standard

A governance event — as distinct from a reporting ceremony — has these five characteristics before anyone leaves the room:

1. A financial consequence figure was the first number presented. 2. The prior month's action register was reviewed with named owners reporting completion. 3. At least one Mine Director-level operational decision was made and documented. 4. Every new action has a named owner, a due date, and a completion standard. 5. Minutes are distributed within 24 hours — not 72.

The Governance Delegation That Is Costing Boards Visibility

The logic of delegating reliability to the maintenance function is understandable: it is a technical discipline, it belongs in operations, and the board cannot govern what it does not understand at the technical level. This logic is correct at the task level and wrong at the risk level. The board does not need to understand hydraulic contamination to govern the financial exposure it creates. The board governs financial performance without understanding every revenue-generating transaction. It governs safety without understanding every operational hazard. Reliability deserves the same framework.

The financial case is straightforward: in a 50-unit mining fleet, every 1% of fleet mechanical availability below target represents $2–8 million in annual production revenue depending on commodity price and utilisation profile. In most mid-tier operations, the availability gap — actual vs target — represents $15–40 million in annual production exposure. This number is almost never on the board agenda.

The Financial Materiality Test

Most boards apply a financial materiality threshold for disclosure and governance attention. The test for whether reliability deserves board-level governance is simple: does the financial exposure from reliability underperformance exceed your organisation's materiality threshold?

If the reliability-related financial exposure at your operation exceeds the threshold at which the board would expect management to report a business risk — then it is board-relevant. The test is not technical. It is financial. And in the majority of mid-tier mining companies, it passes the test by a significant margin.

What Board-Level Reliability Governance Looks Like

Board-level reliability governance does not require the board to understand maintenance engineering. It requires the board to receive, review, and act on a small number of executive-level reliability indicators — in the same way it receives, reviews, and acts on financial KPIs and safety metrics.

The Five Board-Level Reliability Indicators

1. Fleet Mechanical Availability vs Target. The single most important indicator. Expressed as a percentage, with a financial dollar consequence attached: "Fleet availability was 74.5% against an 88% target last quarter — representing the operation's maintenance cost in foregone production value."

2. MTBF Trend Direction. Is the fleet getting more reliable or less reliable? A 12-month MTBF trend line for primary equipment classes tells the board whether the maintenance system is working or deteriorating. Three consecutive quarters of declining MTBF is a board-level signal.

3. Maintenance Cost per Operating Hour vs Industry Benchmark. If the operation's CPH is 40% above the industry benchmark for this equipment class and age, the board is funding an inefficient maintenance system. The gap, multiplied by annual operating hours, is a capital allocation efficiency question.

4. Reliability Maturity Score Trajectory. An annual MRMM (Maintenance Reliability Maturity Model) score gives the board a single number — out of 80 — that reflects the organisational capability of the maintenance and reliability governance system. It tells the board whether the capability to sustain the operation is improving or eroding.

5. Major Reliability Events — Production-Stopping Events. Any event that stops or significantly disrupts production for more than 48 hours should be reported to the board with: root cause, financial consequence, and the corrective action that prevents recurrence.

Three Decisions That Only the Board Can Make

Board-level reliability governance is not only about receiving information. It enables decisions that remain unmade when reliability is managed at the departmental level — because the authority or the budget required sits above the Maintenance Manager's or even the Mine Director's approval threshold.

The Quarterly Board Reliability Summary

MitWin's S6 Reliability Governance Partnership produces a Quarterly Board Reliability Summary — a four-page executive brief that gives the board exactly what it needs to govern reliability at the strategic level: the five board-level indicators, the quarter's major events with root causes and corrective actions, the financial impact summary, and the three strategic priorities for the next quarter.

The summary is deliberately brief. Board members read four pages. They do not read twelve pages of maintenance data. The discipline of producing a four-page summary that covers everything material is what distinguishes board-level governance from departmental reporting delivered to a higher audience.

In MitWin's S6 client portfolio, operations where the board receives a Quarterly Reliability Summary consistently report that the board has approved at least one significant reliability-related capital decision — RE hire, fleet replacement revision, governance programme investment — within 12 months of the first summary. In operations where reliability governance remains below board level, these decisions remain pending.

What the Report Shows. What the Event Cost.

A CAT 793F haul truck hydraulic pump failure at a West African gold operation. The breakdown report: repair cost $82,000, downtime 36 hours. This is the number that appears in the maintenance budget, the KPI dashboard, and the monthly performance report. It is also the number that informs the investment case for the preventive actions that would have prevented this failure.

The true cost of that failure: $866,400. The report was not wrong. It was incomplete — by 91%.

The Six Components of True Breakdown Cost

Every unplanned failure event carries six financial consequences. Maintenance reporting systems are designed to capture the first. The other five are absorbed elsewhere in the financial statements — or simply not counted at all.

Component 1 — Direct Repair Cost

Labour, parts, and consumables for the repair. This is the number that appears on the maintenance budget line item and the breakdown report. It is the only component most operations formally record against the failure event. In the CAT 793F example: hydraulic pump assembly $58,000, labour 12 hours × $120/hr = $1,440, consumables and fluids $1,800, subcontract lifting $780. Total direct repair: $62,020.

Component 2 — Production Value Lost

The most significant component — and the one most consistently absent from breakdown reporting. Downtime hours × production value per operating hour. In the CAT 793F example: 36 hours × $18,000/hr production value (gold operation, primary haul truck, confirmed with Finance Director). Production loss: material at this fleet's production rate. This number does not appear anywhere in the maintenance budget. It appears in the operations report as a production shortfall — with no connection to the maintenance event that caused it.

Component 3 — Emergency Procurement Premium

The hydraulic pump assembly was not in stock. Emergency sourcing from the CAT dealer in the nearest major city: standard cost $58,000 purchased as scheduled → $76,400 purchased as emergency (1.32× premium) + air freight $4,200 = $80,600. Premium above standard procurement: $22,600.

Component 4 — Component Life Consumed

The failure was not a wear-out failure. It was a contamination failure — hydraulic oil at ISO 18/16/13 particle count vs target of 16/14/11. The pump ran contaminated for an estimated 340 hours before failure, consuming component life at 2.4× the designed wear rate during that period. Remaining design life consumed prematurely: estimated 4,200 operating hours × lifecycle cost rate of $14/hr = $58,800 in premature lifecycle value erosion.

Component 5 — Secondary Damage from Rushed Repair

The repair was completed under time pressure. Post-repair inspection at the next scheduled service (6 weeks later) found: hydraulic filter bypassing from incorrect reinstallation torque — estimated contamination exposure during this period sufficient to initiate bearing wear on the replacement pump. Secondary damage repair at next service: $38,000. This cost appeared on the next month's breakdown report as a separate unrelated event. It was not separate.

Component 6 — Planning Disruption Cascade

The breakdown consumed 14 technician-hours of unplanned labour, displacing three planned maintenance tasks that were deferred by 8–12 days. One of those three deferred tasks — a magnetic plug inspection on HD785-11 — missed its window. Six weeks later, HD785-11 suffered a final drive failure from undetected magnetic debris. Conservative planning disruption allocation for the original CAT 793F event: $37,000 (proportional attribution of the downstream cost).

Why the Undercount Persists — and Why It Matters

Reporting systems are designed to capture costs — not consequences. The production loss from a breakdown event is recorded against an operations output target, not against the maintenance event that caused it. The emergency procurement premium is a procurement variance, not a maintenance cost. Secondary damage appears as a new independent event three months later. The connections between these numbers and their common cause are never made — because no one's reporting mandate requires making them.

The consequence: every investment case for reliability improvement is built on a cost basis that understates the problem by 35–91%. A CBM programme that costs $120,000 per year and prevents two hydraulic pump failures — each truly costing $866,000 — generates a true return of $1.61M per year. The investment case built on the maintenance report numbers: $120,000 cost vs $124,000 benefit (two × $62,000 repair cost avoided). The investment case built on the complete numbers: $120,000 cost vs $1,610,000 benefit. Same programme. Different decision.

The CFO Reconstruction — Building the Complete Model

The complete breakdown cost model can be reconstructed from existing data sources. No new systems are required — only cross-functional data integration that is rarely performed.

Production loss (Component 2): Downtime hours from CMMS × production value per hour from Finance. Production value per hour = Annual revenue ÷ annual operating hours ÷ number of production-critical machines. This calculation takes 20 minutes and uses existing data.

Emergency procurement premium (Component 3): Emergency PO value × (emergency price ÷ standard price − 1) + air freight cost. Emergency POs are flagged in most procurement systems. The calculation requires one cross-reference between procurement and CMMS breakdown date.

Component life erosion (Component 4): Estimated remaining component life × lifecycle cost rate. Lifecycle cost rate = (Replacement cost ÷ design life in hours). Requires component life tracking in CMMS — which is also the foundation of the S5 lifecycle programme.

Secondary damage and cascade (Components 5–6): These require a formal RCA for each major failure event — the only way to trace downstream cost attribution accurately. This is precisely what a functioning Reliability Engineer produces.

The Situation: Over-Stocked and Under-Stocked Simultaneously

In this illustrative scenario, the operation's warehouse contained a $22M+ spare parts inventory. The last physical stock count had been conducted 22 months prior. Emergency procurement accounted for 34% of all purchase orders — each carried at an average 2.1× price premium above standard procurement, plus air freight costs that averaged $4,200 per emergency shipment.

During the structured S1 diagnostic assessment, the field consultant conducted a warehouse walk. The observation: bins were unlabelled. Parts were stored by delivery batch — not by equipment class. The storesperson required 24 minutes to determine whether the hydraulic pump assembly for the primary excavator class was in stock. The answer, after 24 minutes: it was not. The lead time from the OEM supplier in Japan: 7 weeks.

This was not a budget problem. The operation was spending $1.2M per year in emergency procurement premiums alone. It was a classification and governance problem — and it had an exact analogue in every dollar of dead stock sitting in the same warehouse.

What the S4 Analysis Found

The MitWin S4 (Spare Parts & Inventory Optimisation) engagement began with a complete inventory audit — the first in 22 months. All 5,847 line items in the inventory management system were reviewed against physical stock counts. Finding: 28% of items in the system did not match physical reality. The CMMS data was unreliable.

The dual classification framework — ABC (by consumption value and frequency) crossed with Criticality (by production consequence of stockout) — produced a 9-cell matrix across all 5,847 items. The C1 category (critical, production-stopping consequence of stockout, no substitute available): 64 confirmed items. Of those 64, only 38 were at or above their minimum reorder point. 26 were below reorder point. 16 were at zero stock.

The Intervention: Ten Weeks, Four Workstreams

Workstream 1 — Classification and Min-Max Design

All 5,847 items classified using the dual ABC × Criticality framework. For all 64 C1 and 186 C2 items, minimum stock levels were calculated using the safety stock formula: (Lead Time days ÷ MTBF hours) × fleet count × component quantity per unit. This produced mathematically defensible minimum stock levels for every critical item — grounded in failure probability and lead time risk, not in historical consumption or budget intuition.

Workstream 2 — ERP Auto-Trigger Configuration

The ERP procurement module was configured to auto-generate a purchase order when any C1 or C2 item fell to its reorder point. No human decision required between stockout risk and procurement action for any critical item. The system governs the stocking level — the procurement team executes. This configuration removed the primary cause of the 16 zero-stock C1 failures: procurement decisions that were deferred, forgotten, or deprioritised by budget pressure.

Workstream 3 — Consignment Stock for Long-Lead Critical Items

For the hydraulic pump assembly — 7-week OEM lead time, $58,000 per unit, primary excavator class — a consignment stock arrangement was negotiated with the OEM's regional distributor. Two units held on-site in the OEM's name; the client pays on consumption (when a unit is used from consignment stock, a replacement is ordered and the used unit is invoiced). Working capital impact: zero. Insurance against a 7-week stockout: guaranteed. This arrangement was extended to 8 other long-lead C1 items with similar lead time profiles.

Workstream 4 — Dead Stock Disposal

All items classified as dead stock (zero consumption in 24 months) were reviewed against three disposal routes: supplier return under stock-rotation clauses (applicable to $1.32M of items from suppliers with return agreements), industry sale to other mining operations in the region ($0.31M recovered), and write-off ($0.53M with tax benefit). Total dead stock disposal: $2.16M in cash and write-offs — releasing the warehouse space and ERP system attention these items were consuming.

Results: Ten Weeks, $4.8M Released

The Zero-Stock Event That Did Not Happen

Fourteen weeks after the ERP auto-trigger was configured, the hydraulic pump assembly on EX-03 (primary excavator) showed a threshold breach on oil analysis — contamination ingress confirmed. Under the prior governance system: no stock, 7-week wait, a material annual cost in production loss. Under the new system: 2 units in consignment stock on-site, pump replaced within 4 hours, machine returned to service same shift. The a significant production loss event was a $58,000 planned replacement event. This single prevented event recovered 68% of the S4 programme investment.

The Executive Lesson

The CFO who approved this engagement asked, before signing: "Why has this not been done before?" The answer: spare parts management at this operation had been treated as a procurement function. Procurement logic — historical consumption, budget efficiency, supplier terms — does not include failure probability, MTBF data, or the production consequence of a stockout on the primary excavator hydraulic system.

The reliability engineering logic that identified the 64 C1 items and calculated their minimum stock levels used data that procurement had access to — CMMS failure records, OEM lead times, fleet size. The data existed. The analytical framework did not. That is the gap S4 closes.

The Age-Based Replacement Fallacy

Age is visible. Economic crossover requires calculation. This difference in effort explains why most mining fleet replacement decisions are made using the wrong variable. A 9-year-old haul truck that is performing at or above its benchmark cost-per-hour profile has significant economic life remaining. A 6-year-old haul truck operating at 2.1× its benchmark CPH has already crossed the economic crossover point. Age alone tells you which is older. It does not tell you which should be replaced.

In a 2024 MitWin engagement, a board had approved a significant maintenance spend fleet replacement programme covering 23 major assets. The approval trigger: "aging fleet — declining reliability and time to refresh." After LVI analysis, 4 of the 23 approved-for-replacement assets had LVI above 0.90 — outperforming their expected lifecycle value. The revised programme: 5 immediate replacements, 6 rebuilds, 12 retained. CAPEX deferred: $47.1 million.

The Lifecycle Value Index — How It Works

The LVI is a single ratio calculated from two numbers that every mining operation already tracks:

The benchmark CPH for each equipment class and age profile is derived from OEM lifecycle data, adjusted for local operating conditions (duty cycle, environment, hours-per-day utilisation). MitWin maintains a reference database across multiple mining engagements covering 12 equipment classes in 8 countries.

The Three Questions Before Approving Any Replacement Programme

The Rebuild Zone — Where the Most Value Is Created

Assets in the LVI 0.70–0.85 range are the most economically important fleet management decision point. Too often, these assets are replaced because they are declining — without modelling whether a correctly-executed rebuild would restore their CPH profile to benchmark. In MitWin's S5 engagement data, rebuild programmes for assets in this LVI zone deliver an average ROI of 3.2–6.8:1 against replacement cost — and extend asset life by 12,000–18,000 operating hours.

The test for rebuild viability: estimated post-rebuild CPH × expected rebuild life (hours) vs replacement EARC × same period. If the rebuild scenario produces lower total lifecycle cost over the modelled period — rebuild. The modelling requires 3–4 hours per asset. The decision it changes is worth $400,000–$2.4M per asset.

Planning vs Scheduling — the Distinction That Determines Everything

A scheduler puts jobs on a calendar. A planner makes jobs executable before they appear on the calendar. A planned job is one where the scope is defined, the parts are confirmed available, the tools are listed, the correct skill level is assigned, the equipment release window is confirmed with operations, and the post-repair verification requirement is documented — before the job is started. A scheduled job is one where a date and a machine number appear in the CMMS.

In most mining operations, the planning function produces scheduled jobs. The distinction explains why the same planned maintenance task costs 2.8× more when executed reactively — not because the task is different, but because everything that should have been done before the job started is now done at speed, under pressure, with the machine down and production stopped.

Why Most Operations Cannot Achieve 4-Week Planning

The structural reasons a planner operates on a 3-day horizon — and they are structural, not individual capability failures:

Three Zero-Cost Structural Fixes

Fix 1: Remove the planner from reactive dispatch. Reactive breakdown coordination is assigned to the Workshop Foreman or a dedicated shift coordinator. The planner does not attend to breakdowns. The planner does not answer breakdown calls. The planner's phone number is not given to supervisors for breakdown coordination. This single change recovers 40–60% of planning time — without hiring anyone.

Fix 2: Establish a weekly equipment release meeting. Thursday afternoon, 45 minutes. Operations Superintendent and Maintenance Planner. Agenda: confirm equipment release windows for next week (Week 1 of rolling schedule). Any machine required by operations in a window previously marked for maintenance: negotiate and document. Output: locked Week 1 schedule before end of Thursday. This meeting enables the planning function — without it, the planner is planning in a vacuum.

Fix 3: Implement a "parts confirmed" gate on every Week 1 WO. No WO is locked into Week 1 unless parts availability has been confirmed in the warehouse. If parts are not confirmed by Wednesday, the job moves to Week 2 and an emergency parts request is initiated. This single gate eliminates the most common cause of planned job failure: the technician arrives at the machine and the required parts are not there.

Measuring the Improvement

The metric to track: Job Plan Coverage Rate — the percentage of planned WOs that have a complete job plan (scope, parts, tools, skill, time, safety) before execution begins. Target: above 80%. Most first-engagement MitWin operations: 12–22%. In operations where the three structural fixes are implemented, Job Plan Coverage Rate typically reaches 60–70% within 4 weeks and 80%+ within 8 weeks — without adding headcount.

What a Supervisor Actually Does — and Cannot Do at 1:18

A Maintenance Supervisor has five quality-critical functions: directing active repair work, verifying completed repairs before WO closure, enforcing contamination control during service, coaching technicians through non-standard diagnostic situations, and documenting failure observations accurately before the machine is returned to service.

At a supervisor-to-technician ratio of 1:18, a supervisor managing three 8-hour shifts cannot physically observe more than 30% of active repair work in any given shift. Post-repair verification — the supervisor checking that the work is actually correct before the WO is closed — is the first function to disappear. When verification disappears, contamination gets reintroduced during hydraulic service. Components are reinstalled with incorrect torque. Failure codes are filled in from memory, not observation. These are not technician failures. They are supervision failures — produced by a structure that makes supervision impossible.

The Financial Calculation

At a 46-unit mining fleet operating at a 1:18 supervisor-to-technician ratio, MitWin estimates the quality failure cost from under-supervision at $380,000 per year — from three sources:

The Industry Benchmark — and Why It Is Different Underground

For underground operations: the benchmark is stricter — 1:8 maximum — because underground repair environments are more complex, lighting is limited, contamination risks are higher in humid acidic environments, and the consequence of a quality failure underground is potentially a production-stopping event in a single-access mining section.

Three Structural Responses

Response 1: Supervisor hire with technical credibility. The additional supervisor must have the technical authority to reject a technician's repair and require it to be redone. A supervisor who cannot diagnose the failure mode cannot assess repair quality. The hire brief must specify minimum years of hands-on experience at the relevant trade level — not just management experience.

Response 2: Post-repair verification as mandatory WO closure step. Configure the CMMS so that Class A WOs cannot be closed without a supervisor sign-off field completed. This enforces the verification step within the existing system — at zero additional cost beyond the configuration change.

Response 3: Shift redesign to align supervisors with Class A repair windows. If the supervisor ratio cannot be immediately corrected through headcount, redesign shift coverage to ensure supervisor overlap during the highest-risk repair windows — typically the first 4 hours of day shift when most Class A repairs are started. This concentrates supervision on the highest-consequence work.

The Design Assumption No One Questions

When a maintenance planner configures a PM schedule in the CMMS, they typically start with one data source: the OEM operation and maintenance manual. This is rational — the manual is authoritative, it is available, and it represents the manufacturer's best knowledge of the equipment's maintenance requirements. It is also based on an operating environment that almost no real mining site exactly replicates.

The OEM PM interval is not the correct interval for a specific operation. It is the interval that the OEM expects will satisfy the warranty obligation for a broad range of customers — designed conservatively enough to prevent catastrophic failures across the widest possible operating context. In a highly corrosive underground acid-rock drainage environment, the OEM hydraulic seal specification is incorrect. In a high-oxide laterite dust open-cut, the OEM air filter interval is incorrect. The equipment is correct. The interval is wrong.

The Five Most Common Interval Mismatches in Mining

What RCM-Based Strategy Design Produces

A Reliability-Centred Maintenance approach to strategy design answers seven questions for every significant failure mode in every equipment class:

(1) What are the functions and performance standards? (2) In what ways can it fail? (3) What causes each failure mode? (4) What happens when it fails? (5) Does it matter — and how much? (6) Can the failure be predicted or prevented? (7) What if it cannot?

The answers to these questions — applied to a specific equipment class in a specific operating environment — produce a strategy matrix: Failure Mode × Detection Method × Task × Interval × Condition for Interval Change. The strategy matrix is the document that governs what the CMMS executes. The CMMS enforces the strategy — it does not create it.

The P-F Interval — the Most Important Concept in CBM Design

The Potential Failure to Functional Failure (P-F) interval is the time between when a developing failure first becomes detectable and when it produces functional failure. It is the most important concept in CBM programme design — and the one most consistently absent from mining maintenance operations that have an oil analysis programme but no RE to design it correctly.

If the P-F interval for hydraulic pump contamination failure is 800 operating hours (detectable via oil analysis at 800 hours before breakdown), then the oil analysis programme must sample at least every 400 hours — half the P-F interval. If the programme samples every 500 hours, approximately 40% of developing failures will reach functional failure before the next sample. The programme is not wrong — the frequency is insufficient for the P-F interval of the failure mode it is designed to detect.

The Six-Month Strategy Transition

Transitioning from OEM-default to RCM-based strategy in a producing operation does not require a shutdown. MitWin's S2 programme operates on a 6-month rolling transition: prioritise the highest-frequency failure modes first, validate interval corrections with field data from the first 3 months, update the CMMS task library monthly, and train supervisors on changed task specifications as each failure mode's strategy is updated. The result: a strategy that improves every month, rather than one designed once and never reviewed.

The Procurement Logic Gap

Procurement logic is optimised for cost efficiency and supplier relationship management. It answers: how do we buy this at the lowest per-unit cost, on the best payment terms, with the most reliable supplier? These are legitimate questions. They are also the wrong questions for determining whether a hydraulic pump assembly should be held in stock for a primary excavator in a remote mining operation with a 7-week OEM lead time.

The right question is a reliability engineering question: given this component's failure probability over the next 90 days (derived from its MTBF and the current fleet hours), the production consequence of a stockout (primary excavator down for 7 weeks = a material annual cost in production loss per event), and the lead time risk (7 weeks), what minimum stock level is required to maintain a >98% probability of having this component available when it fails?

That question cannot be answered from a purchase order history. It requires MTBF data, fleet composition data, lead time data, and production value data — none of which procurement manages.

The Dual Classification Framework

The governance fix is a classification system that brings reliability logic into stocking decisions. The dual classification framework crosses two dimensions:

The 9-cell matrix produced by crossing these two dimensions tells every stocking decision: a C1/A-class item (critical, high-value) requires consignment stock, aggressive min-max, and active monitoring. A C3/C-class item (standard, low-value) is governed by pure procurement logic — order as needed, no safety stock required. The same governance system cannot apply to both.

The Safety Stock Calculation

For every C1 and C2 item, safety stock is calculated using:

This formula requires three inputs that procurement does not manage: MTBF, fleet count by class, and production operating hours per day. It requires one input procurement does manage: lead time. The calculation is a 10-minute exercise per C1 item — once the data access is established. The total calculation for 64 C1 items takes approximately one working day for a qualified analyst.

The ERP Auto-Trigger — Removing Human Decision From Critical Reorder

The most common failure in C1 spare management is not incorrect min-max calculation — it is that the reorder trigger is known but not acted on. Budget pressure defers the order. The reorder alert is generated but not approved. The critical spare runs to zero — not because the system failed, but because the human decision to approve the PO was delayed by a procurement cycle or a budget freeze.

The fix: ERP auto-trigger for all C1 items. When the stock level falls to the reorder point, a purchase order is generated automatically — not requiring human approval — and issued to the pre-approved supplier at the pre-agreed price. For C1 items, the cost of the component is always less than the cost of the production loss from the stockout. Human approval authority on C1 reorders is a governance risk, not a governance control.

The Financial Case — Two Operations, Same Industry

The Three RE Role Definitions That Fail

Before specifying what the right RE looks like, it is worth naming the three versions of the role that produce no reliability improvement — because all three are common, and all three are hired from well-intentioned but under-specified job descriptions.

The Six Non-Negotiable Technical Competencies

These six competencies must be demonstrable at interview — not claimed in a CV. The interview should be designed to test each one. A candidate who cannot demonstrate any one of the six should not be hired regardless of experience, credentials, or interpersonal skills.

The 30-Minute Assessment Exercise

Give every RE candidate the following exercise — 10 minutes to review, 20 minutes to discuss. This exercise distinguishes capable from claimed in a way that no CV screening can.

The Role Design Foundation

The RE must report to the Maintenance Manager — not to a supervisor. Reliability engineering authority requires direct access to maintenance strategy decisions. An RE reporting to a supervisor will be subordinated to reactive priorities within 60 days.

The RE must have protected time. Minimum 60% of working hours on reliability activities (FMEA, RCA, CBM monitoring, MTBF analysis, strategy review). The remaining 40% may include meeting participation, reporting, and training support — but must not include reactive coordination. The RE who spends 70% on reactive is a supervisor with a reliability title.

The RE must have corrective action authority. When the RE issues a corrective action — "install desiccant breather on all 6 excavators before this week's hydraulic service" — supervisors must implement it. If the RE's corrective actions are suggestions that can be deferred, the role has no reliability engineering function.

The Handover Nobody Planned For

The operator had known for 18 months that the maintenance contract would not be renewed. What they had not done — and what almost no operations do — is plan the capability transfer that a contractor-to-owner transition requires. At contract end: the contractor's CMMS data (7 years of failure records) was legally the contractor's intellectual property and was not transferred. The contractor's PM task library was proprietary. The maintenance team that had been employed by the contractor was offered positions with the operator — but the institutional knowledge they carried could not be transferred as a document.

The Mine Director described Month 1 as "maintenance by feel — we know roughly what needs to be done, but we have no organised basis for doing it, no data to plan from, and no reliability system of any kind." Fleet availability had been 84.2% in the contractor's final quarter. At Month 1 of self-managed maintenance: 68.4%. The gap: 15.8 percentage points. The financial consequence at this coal operation's production rate: $2.4M per month in production loss above the contractor baseline.

What MitWin Found at Engagement

With no historical failure data accessible, any strategy redesign had to be built from field observation, OEM data, and rapid failure pattern identification during the first 6 weeks of engagement. MitWin field consultants conducted structured physical inspections of all 52 units in the first two weeks — documenting visible wear patterns, fluid conditions, and maintenance evidence. This observation-based approach replaced the data analysis that would normally anchor the strategy design.

The Integrated S2 + S3 Programme — 16 Weeks

Months 1–4 (S2 Priority — Strategy Rebuild): A new PM task library was built for all 6 equipment classes from scratch — using OEM manuals as the starting point, calibrated to Southeast Asian operating conditions (high humidity, laterite dust, wet season haul road degradation) from MitWin's regional reference database. Every PM interval adjusted for local conditions. Oil analysis programme launched in Week 4 — SOS account opened, sampling schedule configured, threshold limits set for each equipment class. Operator pre-shift inspection routes designed and rolled out with technician training in Weeks 5–6.

Month 3 onward (S3 Execution — Governance and Discipline): Planning discipline installed — planner removed from reactive dispatch, 4-week rolling schedule built, equipment release protocol established with Operations Superintendent. Daily stand-up from Week 10. Weekly reliability review from Week 10. First formal RCA conducted Month 3 on the highest-frequency failure mode (haul truck engine air filter bypass — confirmed as OEM interval mismatch for the laterite operating environment). Monthly executive review established Month 4 — Mine Director attending.

MRMM tracked monthly throughout the programme: 31/80 → 38/80 (Month 4) → 44/80 (Month 8) → 51/80 (Month 10) → 54/80 (Month 12).

Results at 12 Months

The Executive Lesson — Contracts Need Exit Strategies

The contractor who managed this fleet for 7 years did their job. Their CMMS data, their PM task library, and their institutional knowledge were legitimately their property under the contract terms. The operator had agreed to those terms. What neither party had planned — and what no one required — was a knowledge transfer protocol that would give the operator the capability to maintain their own fleet at the point of transition.

The lesson is not that contractors are inappropriate. It is that every contract maintenance arrangement should include, from Day 1, a data sharing and knowledge transfer protocol that gives the operator access to failure records, PM task libraries, and CBM data throughout the contract period — so that when the contract ends, the operator holds the institutional knowledge required to sustain maintenance independently. MitWin now includes this recommendation in every S1 audit where contract maintenance is identified as part of the current model.

The Wrong Hire: When Role Design Failure Compounds Rapidly

In this illustrative scenario, eight months before the structured intervention, the Maintenance Manager had convinced the Mine Director to approve a Reliability Engineer hire. The job description had been written by HR from a generic template: "minimum 5 years maintenance experience, knowledge of reliability principles, CMMS proficiency, good communication and problem-solving skills." The description attracted 34 applicants. The selected candidate had 7 years of process plant maintenance experience — well-regarded, articulate, and genuinely motivated to work in a reliability role.

At Month 2, the Maintenance Manager began to suspect the hire was not producing. At Month 3, the RE had attended 6 breakdown callouts, produced 2 monthly MTBF reports, and conducted zero formal RCAs. When asked to interpret the most recent SOS oil analysis result for the primary LHD fleet — showing Fe at 38 ppm against a threshold of 25 ppm — the RE's response was: "I would recommend an oil change at the next service." The threshold had been exceeded for 6 weeks. The correct response was a work order within 48 hours and an investigation into the source of the iron wear. The RE did not know the difference. The hire lasted 4 months.

In the 8 months between the failed hire and MitWin engagement, the 11 recurring failure modes continued at their prior rate — costing an estimated $854,000 per month in preventable recurring failure events. Total cost of the wrong hire decision: significant annual cost in lost prevention value, plus $68,000 in hiring and salary costs.

The S7 Engagement — Designing the Role Correctly This Time

the S7 workforce advisory was commissioned with a specific mandate from the Mine Director: "Do not let us hire the wrong person again." The 6-week engagement produced four outputs.

Output 1 — RE Role Profile to MitWin Specification. Role purpose: eliminate recurring failures, not report on them. Six non-negotiable technical competencies specified with testable evidence criteria (FMEA application, RCM decision logic, oil analysis interpretation, MTBF trend analysis, CBM programme design, failure elimination project management). Reporting line: direct to Maintenance Manager. Protected time requirement: minimum 60% on reliability engineering activities.

Output 2 — Competency-Based Interview Framework. Five structured questions, each with a scoring rubric distinguishing genuine from claimed competency. The oil analysis question: "I am going to hand you a real SOS report from our LHD fleet. Tell me what it says, what thresholds have been exceeded, and what you would do in the next 24 hours." Candidates who say "change the oil at the next service" fail immediately.

Output 3 — 30-Minute Assessment Exercise. Real CMMS failure data from the site (anonymised unit IDs). 24 months, 11 recurring failure modes. Candidates given 10 minutes to review, asked to: identify the highest-frequency mode, hypothesise the root cause, specify the CBM programme element that would detect it before breakdown, and quantify the annual cost. The assessment distinguishes analytical thinking from experience claims.

Output 4 — 90-Day Onboarding Plan. Week 1: fleet profile, underground environmental brief (ARD pH 2.8–3.6, seal specifications, contamination risks), CMMS access. Month 1: first supervised RCA completed and presented to MM. Month 3: CBM programme designed for LHD class and launched; MTBF tracked weekly. The plan was given to the new hire on Day 1.

The Replacement Hire — Month 3

Using the MitWin brief, a specialist mining technical recruiter sourced 8 candidates. MitWin reviewed CVs against the competency framework — not against years of experience. Three candidates were selected for interview. The assessment exercise was administered to all three. One candidate scored significantly above the others: correctly identified the hydraulic contamination root cause in the exercise data, proposed sealed desiccant breather installation and particle count programme as the primary corrective action, and calculated the annual cost correctly. Offer accepted. Onboarded Month 3.

Results at Month 12

The Board Decision That Was Based on the Wrong Variable

The Mine Director's replacement programme recommendation to the board was built on two inputs: equipment age and a general perception of declining reliability performance. The CFO approved it because the Mine Director recommended it. The Board approved it because both the Mine Director and CFO supported it. The approval process took 6 weeks. No one asked: "Is the economic crossover point confirmed for each of these 23 assets — or are we replacing some assets that still have significant economic life remaining?"

At the time of board approval, the programme included 4 assets with LVI above 0.90 — outperforming their expected lifecycle value profile. These assets were on the replacement list because they were old — not because they were economically exhausted. Replacing them would have destroyed $12–18M in residual lifecycle value that the operation was already capturing.

What the S5 LVI Analysis Found

The remaining 5 assets — LVI below 0.70 — were confirmed as requiring replacement. 3 of these were confirmed past economic crossover (LVI below 0.60), with marginal maintenance cost exceeding equivalent annual replacement cost. 2 were approaching crossover and had a 60-day replacement timeline to prevent operating below economic crossover for a full additional year.

The Rebuild Economics — Six Assets in the Borderline Zone

For the 6 assets in the LVI 0.70–0.80 zone, MitWin modelled rebuild vs replace economics for each. The key question: does a correctly-executed component rebuild restore CPH performance to benchmark level, and for how many additional operating hours?

Total rebuild investment across 6 assets: $1.32M. Total equivalent replacement cost for the same 6 assets: $8.2M. Value of rebuild decision vs replacement decision: $6.88M in capital not deployed — plus the residual lifecycle value of the retained assets estimated at an additional $3.4M over the rebuilt life extension.

The Revised Board Programme

MitWin produced an 8-chart board presentation for the MD, CFO, and COO — presenting the LVI analysis for all 23 assets, the rebuild economics model for the 6 borderline assets, and the revised programme recommendation.

Original board-approved programme: 23 replacements, significant maintenance spend capital commitment.
Revised programme: 5 immediate replacements ($14.2M) · 6 rebuilds ($3.1M) · 12 retained with lifecycle extension programme · Total capital deployed: $17.3M.

CAPEX deferred: significant capital deferral. Value of deferral at the operation's 9.5% hurdle rate (NPV of capital not deployed over the 3-year asset life extension): the operation's maintenance cost. CPH improvement from rebuild fleet (contributes to ongoing EBITDA improvement): estimated $2.8M over rebuild life. Total Year 1 value: significant. Programme investment: $175,000. Strong programme return.

The Board's Response

The CFO's question after reviewing the presentation: "Why wasn't this analysis done before we approved the original programme?" The Mine Director's answer: "We didn't have a framework for it." The Board approved the revised programme unanimously. The significant capital deferral in deferred capital was redirected to the operation's expansion programme — a project that the original replacement spend would have delayed by 18 months.

The Mine Director subsequently implemented a standing requirement: no fleet replacement programme to be submitted to the board without a completed LVI analysis for all assets on the proposed replacement list. This requirement costs approximately 4 hours of analyst time per asset. It prevents the capital allocation error that almost consumed significant capital of the operation's expansion capital.

The Correct Components in the Wrong Configuration

The prior strategy redesign had produced a technically competent PM task library — calibrated to the Moroccan operating environment, incorporating local dust and humidity conditions. The CMMS had been configured correctly — work order types, cost centres, equipment hierarchy, and reporting templates were all functional. An RE had been hired between the two engagements — 14 months in post at MitWin engagement, with a genuine background in reliability engineering from a process plant environment.

None of these components were connected. The strategy document existed in PDF. The CMMS was executing the original OEM-default intervals — 38% of the redesigned PM tasks had not been entered into the CMMS. The oil analysis programme was generating results through a lab in Casablanca, but the alert-to-WO workflow had not been configured in the CMMS. The RE was producing monthly availability reports — and had conducted zero formal RCAs in 14 months of employment.

MRMM at MitWin engagement: 44/80 — higher than the average first-engagement score of 38/80, because the structural elements existed. But D6 (Failure Elimination): 2/8. D7 (Condition Monitoring): 2/8. D5 (Execution Discipline): 3/8. The governance connective tissue between the strategy, the CMMS, the RE, and the leadership was absent.

The MitWin Approach — Activation, Not Rebuild

The engagement was scoped as coordination and activation rather than a rebuild — because the prior work was good. The problem was not the quality of the strategy or the CMMS configuration. It was the absence of the connections between them.

Weeks 1–2 — Gap Analysis. MitWin compared the strategy document PM task library against the CMMS task list. Finding: 38% of the redesigned tasks were not in the CMMS. The oil analysis programme was generating results that were being sent directly to the Equipment Manager's email — no integration with the CMMS, no alert threshold configured, no WO auto-generation. The RE's time allocation: 22% on reliability activities, 78% on reactive coordination and reporting administration.

Weeks 2–4 — Strategy-to-CMMS Activation. All 38% of missing PM tasks entered into the CMMS and activated. Oil analysis alert thresholds configured in the CMMS — any result exceeding threshold generates a WO automatically, flagged "CBM Alert — Priority." This single configuration change activated the CBM programme that had been generating data with no downstream governance for 8 months.

Weeks 3–6 — RE Coaching. The RE's capability was genuine but untested in mining environments. MitWin Reliability Lead conducted structured coaching: oil analysis threshold interpretation (3 sessions), 5-Why RCA methodology with worked examples from the site's CMMS data (2 sessions), first supervised RCA at Week 5. The RE conducted their first independent RCA at Week 7 — correctly identifying contamination as the root cause of the excavator hydraulic recurring failure mode and specifying a corrective action that eliminated it within 30 days.

Weeks 4–8 — Planning Discipline. The planner had been operating without a forward schedule — producing scheduled WOs from a reactive queue rather than from a 4-week forward plan. Planner removed from reactive dispatch. 4-week rolling schedule built. Equipment release protocol with operations established. Job plan coverage: 9% → 54% in 8 weeks.

Weeks 6–13 — Governance Activation. Daily stand-up from Week 6. Weekly reliability review from Week 7. Monthly executive review — Mine Director attending — from Month 2. This was the first time the Mine Director had attended a structured reliability review meeting in 14 months. By Week 11, the RE was co-leading the weekly review independently.

Results at Month 6

Faster-than-average maturity progression because the structural elements already existed — strategy designed, CMMS configured, RE in post. They needed to be connected, activated, and governed. Year 1 financial recovery: significant at this fleet scale. Programme investment: $210,000. Year 1 Strong programme return.

The Lesson About Sequencing

The two prior engagements at this operation were not wasted. They produced genuinely useful outputs — a competent strategy document and a correctly configured CMMS. But they were not coordinated with each other, and neither addressed the governance layer that would have made both of them functional. The result: two correct investments in the wrong sequence, producing an operation that looked like it had a reliability system but did not have one.

The correct sequence for a mining operation building reliability capability from a fragile baseline: (1) Diagnostic — understand the current state and quantify the financial exposure (S1). (2) Strategy — design the correct maintenance approach for this environment and equipment (S2). (3) Transformation — implement the strategy, install execution discipline, and establish governance (S3). (4) Governance — sustain what was built with monthly governance, KPI tracking, and on-call advisory (S6). The operation had completed steps 2 and part of step 3 — but had skipped step 1 and had never implemented step 4.

The Paradox of the Growing Maintenance Budget

In 2014, the average maintenance cost per operating hour for a 50-unit open-cut mining fleet was approximately USD 92/hr. In 2024, the same fleet configuration costs approximately USD 134/hr — a 46% increase over 10 years. Over the same period, industry-wide fleet mechanical availability has declined from approximately 83% to 78% on average. More is being spent. Less is being achieved.

The explanation is not inflation alone, though input costs have risen. It is structural: maintenance budget growth is being allocated to the reactive consequences of reliability failure — emergency labour, emergency parts, emergency repair subcontracts — rather than to the proactive systems that prevent failure. The spending is accelerating. The failure is accelerating with it.

Where the Budget Is Actually Going

In the average mining operation structured reliability assessments, the maintenance budget breaks down as follows:

The pattern is consistent: 56–69% of maintenance spend is consumed by reactive failure and its consequences. The activities with the highest ROI — reliability engineering, CBM programme operation — receive 3–7% of the budget. Increasing the total budget at this allocation ratio produces more reactive spend. It does not produce more reliability.

The Three Reallocation Decisions

The CFO Conversation

The right conversation is not "our maintenance budget is too small." It is: "our maintenance budget is allocated 57% to the consequences of failures that our reliability engineering function — currently at 2% of budget — could prevent." The ask is not a larger budget. It is a different allocation of the existing one. That framing changes the conversation from a cost request to a capital efficiency argument.

What Wrench Time Measures — and Why It Matters

Wrench time is the percentage of available shift hours that a maintenance technician spends in direct productive work on equipment. It is the most direct measure of maintenance labour efficiency — and the most consistently disappointing metric when it is measured honestly.

Industry benchmark for world-class mining maintenance operations: 55–65% wrench time. Industry average in structured reliability assessments: 24–34%. The gap between these two figures represents the lost productive capacity of the maintenance workforce — paid for, rostered, present on site, and not productively engaged for half their available working hours.

In a 50-technician fleet operation paying an average blended rate of $85/hour (inclusive of overhead): every 10 percentage points of wrench time improvement recovers approximately $1.85M per year in labour productivity on a 40+ unit fleet at this blended rate — without hiring a single additional person.

The Three Friction Points That Consume 80% of Lost Wrench Time

Measuring Wrench Time Without a Formal Study

A formal wrench time study requires work sampling methodology. A useful proxy can be constructed in one afternoon: observe the workshop and active work sites for 4 hours. Record, at 15-minute intervals, how many technicians are in direct productive contact with equipment vs waiting, travelling, or in administration. The ratio is a reliable approximation of wrench time. If the result is below 35%, the three friction points above are almost certainly the primary causes.

What a Failure Code Actually Is — and Is Not

A failure code is not an administrative field on a work order. It is an engineering data point that describes the failure mechanism of the component that was repaired. The distinction matters because the failure mechanism is what drives the corrective action. "Pump failed" is not a failure code. "Pump failure — internal seal wear from hydraulic fluid contamination (Si above threshold)" is a failure code. One produces a record. The other produces intelligence.

Most CMMS systems have a failure code library with 50–200 coded options. Most maintenance teams use the same 8–12 codes for 80% of their work orders — typically the broadest, most general options available. "Mechanical failure." "Wear." "Hydraulic." These codes are not wrong. They are so general as to be analytically useless for identifying failure modes, tracking recurring patterns, or building an accurate picture of what is actually causing downtime.

Why Technicians Use Wrong Failure Codes

The Financial Cost of Bad Failure Codes

The cost of poor failure code data is measured in lost analytical capability — which translates to delayed identification of recurring failure modes, delayed RCA, and continued preventable failure costs.

In a 50-unit mining operation with a 28% recurring failure rate and a poor failure code library: the Reliability Engineer cannot identify recurring failure modes from CMMS data without manual investigation of each work order. This manual investigation consumes 40–60% of the RE's available analytical time. Time that would otherwise be spent on RCA, CBM programme design, and MTBF analysis is consumed by data cleaning. Conservative estimate of the delay cost: 3–6 months of additional recurring failure expenditure before each failure mode is identified and addressed — at an average cost of $45,000–$180,000 per additional month per active failure mode.

The Three-Step Fix

Step 1: Rationalise the failure code library. Remove codes that are never used and those that are too general to be analytically useful. Build a library of 40–60 codes that map directly to the failure mechanisms occurring in this fleet — reviewed and approved by the RE. Every code should be connected to a specific failure mode.

Step 2: Train every technician on failure code usage in a single 2-hour session. Connect each code to an example failure they have seen or repaired. Post a laminated quick-reference guide in the workshop and at underground working areas. This converts failure codes from an administrative mystery to a shared analytical language.

Step 3: Measure and report WO Documentation Compliance Rate monthly. Include it in the weekly reliability review. The Supervisor whose team consistently closes WOs with incomplete codes should understand that they are destroying the analytical capability of the reliability system — not failing an administrative requirement.

The Underground Environment Is Not Standard

The OEM maintenance manual for a Sandvik LH514E underground LHD was written for an underground equipment profile — which is an improvement over surface OEM intervals applied underground. But even Sandvik's underground design assumptions are based on an environment with: pH-neutral groundwater, ambient humidity below 85%, silica dust (not sulphide ore dust), and ramp grades below 12%. In many underground hard-rock and sulphide ore operations, none of these assumptions hold.

The consequence: an underground LHD maintained at OEM intervals in a West African underground zinc-lead sulphide operation will experience hydraulic pump failure approximately 2.4× more frequently than the OEM MTBF data predicts. The equipment is not faulty. The standard is wrong for the environment.

Four Underground Factors That Accelerate Failure

Three Failure Modes That Are Directly Caused by Surface-Standard Underground Maintenance

1. Recurring hydraulic pump contamination failure. Standard OEM hydraulic breather (not sealed for ARD environment) + high humidity + sulphide contamination = hydraulic fluid degradation at 2.4× expected rate. A sealed desiccant breather rated for the underground environment eliminates this failure mode at a materials cost of $210 per unit.

2. Recurring final drive seal failure. Standard OEM seal specification (pH > 4.5 rated) in ARD environment (pH 2.8–3.6) = seal failure every 60–90 days. Viton seal upgrade + pH monitoring of underground sump water = failure mode eliminated. Cost per drive: $640 in materials. Cost of continued failure per event: $316,000 average.

3. Recurring drill rotation head bearing failure. Standard OEM grease interval (250hr) in 95% RH underground environment = bearing running dry between services. Interval reduction to 150hr + EP2 lithium complex grease = failure mode eliminated. Cost: 40% more grease per unit per year ($380 additional materials). Cost of continued failure per event: $203,000 average.

Why the Opening Matters More Than the Content

The purpose of a reliability review meeting is to produce decisions — maintenance decisions, operational decisions, resource decisions. The content of the meeting (KPIs, failure reports, work order summaries) is evidence that informs those decisions. The opening sequence of the meeting determines whether the evidence is reviewed in a governance posture (decision-making) or a reporting posture (acknowledgment and observation).

The most common failure of the monthly reliability review is that it begins with the data — the slide deck, the KPI report, the availability chart — before it establishes the accountability context. When the data comes first, the meeting is about the data. When accountability comes first, the meeting is about decisions. These produce fundamentally different outcomes.

The Six Opening Questions — In This Order

The Timing

These six questions take 12–18 minutes when the meeting is well-prepared. Action register: 8–10 minutes. Financial figure: 1 minute. Trend: 2 minutes. Risk asset: 1 minute. Recurring failure: 2–3 minutes. Decision required: 1 minute. Everything after these questions is the evidentiary detail that supports the decisions already identified. The data becomes purposeful — not performative.

The PdM Promise and the PdM Reality

The vendor pitch for predictive maintenance technology is compelling: detect failures before they occur, reduce unplanned downtime by 30–50%, extend equipment life, and reduce maintenance costs. The technology behind these claims is often genuinely capable. The implementation reality in most mining operations is significantly less compelling.

MitWin's engagement data across operations that had deployed PdM technology before our involvement shows: a significant proportion of PdM programmes were generating sensor data that was not being converted into maintenance actions. 42% had alert fatigue — so many alerts being generated that technicians and supervisors had stopped responding to them. 31% had deployed technology on equipment where the fundamental CBM programme (oil analysis, inspection routes) was not yet functioning. In each case, the technology was working. The governance was not.

Why Most PdM Programmes Fail Within 18 Months

Failure Mode 1 — Alert Without Action

A vibration monitoring system on a crusher bearing generates an alert when vibration amplitude exceeds a configured threshold. The alert is sent to the Maintenance Supervisor's email. The Maintenance Supervisor receives 47 emails that day. The alert is read on Day 3. A work order is raised on Day 5. The bearing fails catastrophically on Day 4. The technology detected the failure 72 hours before it occurred. The governance failed to act within that window. The technology is blamed. The governance failure is invisible.

Failure Mode 2 — Alert Fatigue

When alert thresholds are configured incorrectly — too sensitive, not calibrated to equipment-specific baseline performance — the system generates more alerts than the maintenance team can investigate. Within 60–90 days, supervisors and technicians learn to filter alerts as background noise rather than reliability intelligence. The system continues generating alerts. No one investigates them. The programme continues to be funded. No failure is detected before breakdown. Alert fatigue is the most common and most destructive PdM governance failure.

Failure Mode 3 — Technology Deployed Before Basic CBM Is Functioning

An operation that does not have a functioning oil analysis programme — with correct sampling frequency, threshold limits configured, and alert-to-WO workflow active — is not ready for vibration analysis or AI-driven maintenance prediction. The analytical sophistication of the technology must match the analytical capability of the organisation. Deploying Level 4 monitoring technology into a Level 1 reliability governance structure produces a Level 1 outcome at Level 4 cost.

The Four Governance Prerequisites for PdM Success

The PdM Readiness Assessment

Before approving any PdM technology investment, a Mine Director or CFO should require the answer to four questions:

1. Is our oil analysis alert-to-WO workflow configured and actively managed — or are results going to email and not being reviewed? 2. Do we have a Reliability Engineer with the training to interpret PdM output data independently — without relying on the vendor? 3. Have we collected equipment baseline data for the monitored components in healthy operating condition? 4. Is our failure detection rate from existing CBM above 35% — indicating the organisation can absorb and act on condition monitoring intelligence?

If any of these four answers is "no" — the PdM investment will fail. Not because the technology is wrong, but because the organisation is not ready to use it effectively. The investment should be deferred until the prerequisite is met — typically 3–6 months of targeted S2/S3 programme work.

Why Distance Is a Reliability Multiplier

Every governance failure in maintenance has a financial consequence proportional to the time it takes to resolve. A critical spare stockout at a site 45 minutes from a major distributor is a 4-hour problem. The same stockout at a remote site 380km from the nearest parts stockist — on roads that are impassable during the wet season — is a 7–28 day problem. The same failure code quality gap that delays an RCA by 2 weeks at an accessible operation delays it by 8 weeks at a remote site where the RE can only access site data through a satellite connection.

The governance framework required for a remote mining operation is not a simplified version of what a well-resourced accessible operation requires. It is a more disciplined version — because every gap is more expensive, every delay is longer, and every knowledge deficit is harder to fill through external support.

Four Remote Operation Governance Adaptations

Adaptation 1 — Compressed C1 Spare Governance

At a remote operation with 28-day average parts lead time, the standard safety stock calculation (Lead Time ÷ MTBF × Fleet Count) produces significantly larger safety stock requirements than at an accessible operation. This is not an inventory cost problem — it is a risk management calculation. The cost of holding 2 additional hydraulic pump assemblies at $58,000 each ($116,000 working capital) is precisely comparable to the cost of one pump stockout event that extends downtime by 28 days at $18,000/hr production value: $12.1M. The safety stock is not expensive. Not holding it is.

Additionally: consignment stock arrangements with OEM distributors become operationally critical at remote sites. Supplier-held stock on-site (paid on consumption, not on holding) eliminates the working capital objection to adequate safety stock while maintaining the physical availability guarantee.

Adaptation 2 — On-Site Technical Authority

At accessible operations, a Reliability Engineer who is stumped by a complex failure mode can call a specialist, receive a site visit within 48 hours, or access reference materials from a well-connected knowledge base. At a remote site, none of these options are reliably available. The RE at a remote site must have broader technical authority and more comprehensive analytical training than an RE at an accessible operation — because the backup that an accessible operation can call on does not exist.

The hiring brief for an RE at a remote operation must specify: experience in the relevant ore type and mining method, exposure to the specific equipment fleet, and demonstrated ability to diagnose and resolve complex failure modes independently. The "I will consult our specialist team" response that a city-based consulting firm can offer does not work when the specialist team is 1,200km away and the satellite internet has a 2GB daily data cap.

Adaptation 3 — Redundant Governance Cadence

Remote operations often experience governance cadence disruption from: key personnel on R&R rotation (mine director off-site for 2 weeks), communications infrastructure failures (satellite outage during wet season storm), and emergency response events that consume leadership bandwidth. A governance system designed around a single monthly meeting is fragile in this environment.

The MitWin adaptation: governance is conducted at two levels simultaneously — the primary governance meeting (monthly executive review) and a secondary operational review (weekly reliability review) that can maintain decision-making continuity when the primary meeting is disrupted. The weekly review is owned by the Maintenance Manager and the RE — it does not require the Mine Director's physical presence and can be conducted by satellite link. The monthly review requires Mine Director attendance and is conducted in person when the Mine Director is on-site, and held as a synchronised video call when they are off-site. The governance cadence continues regardless of R&R cycles.

Adaptation 4 — Embedded Knowledge Transfer

In remote operations with 35–60% annual workforce turnover — common in developing economy locations where the operation is the primary local employer — knowledge transfer is not a nice-to-have. It is an operational continuity requirement. When the Maintenance Planner who built the 4-week rolling schedule leaves the site after 6 months, the scheduling system must survive the departure.

The adaptation: every governance system element — daily stand-up format, weekly review agenda, monthly KPI Pack template, CBM alert response protocol — is documented as a process, not as a person's knowledge. The documentation is kept in the CMMS, on the workshop noticeboard, and in the site induction package. Every new hire receives the documentation in Week 1. The system survives individual departures because it is encoded in the process, not in the practitioner.

Why Level 1 to Level 2 Is Relatively Fast

The Level 1 to Level 2 transition (MRMM 20–35 → 36–50) is driven by structural installation: governance meetings established, planning function activated, CBM programme launched, RE role filled. These are discrete actions with clear completion criteria. When MitWin's S3 programme installs a daily stand-up, a weekly reliability review, a monthly executive review, and a 4-week rolling schedule — the operation moves from Level 1 to Level 2 within 60–90 days. The improvements are visible, measurable, and motivating. This phase of the improvement journey is the one most operations describe as transformational.

The Level 2 to Level 3 transition is not driven by installation. It is driven by integration — the process by which each governance element begins to inform and strengthen every other element. This integration takes longer, requires deeper organisational capability, and plateaus more easily.

What the Level 2 Plateau Looks Like

A Level 2 operation has a governance cadence that functions — meetings occur, KPIs are tracked, actions are recorded. But the system is not improving. MTBF is stable but not growing. Availability has improved from the Level 1 baseline but has stopped improving. Recurring failure rate has declined but a core set of failure modes persists. The RE is producing good RCAs but the corrective actions are not being consistently implemented. Schedule compliance is at 75–80% — better than Level 1 but unable to break through to 85%+.

This plateau is not caused by a missing governance element. All the elements are present. It is caused by the absence of three integration qualities that distinguish Level 3 from Level 2:

Integration Quality 1 — Failure Intelligence Drives Strategy Updates

At Level 2: the RE conducts RCA on recurring failures. Corrective actions are implemented. The maintenance strategy task library remains unchanged. At Level 3: the RCA output feeds directly into the PM task library — if an RCA identifies a failure mode caused by an incorrect PM interval, the interval is changed in the CMMS within the month. The failure intelligence system and the strategy system are integrated. One informs the other continuously rather than in isolated projects.

Integration Quality 2 — CBM Alerts Drive Planning Input

At Level 2: the CBM programme generates alerts, which generate work orders, which are added to the reactive queue. At Level 3: CBM alerts feed into the planning function as a 4-week forward input — a developing failure detected at 800 hours before breakdown is planned for correction at 600 hours. The CBM system and the planning system are integrated. Condition monitoring intelligence becomes a planning input rather than a reactive trigger.

Integration Quality 3 — Governance Meetings Drive Strategy and Resourcing Decisions

At Level 2: the monthly executive review reviews KPIs and agrees actions. At Level 3: the monthly executive review produces at least one strategic or resourcing decision per month — a maintenance strategy change, a spare parts policy update, a workforce structure adjustment, or a capital allocation decision — in response to the reliability data presented. The governance system is integrated with operational decision-making rather than operating as a separate reporting cadence.

The Three Interventions That Break the Level 2 Plateau

The Undifferentiated Fleet Problem

The logic of treating all fleet assets equally is appealing in its simplicity: every machine matters, every failure is important, every technician should give every job their best work. This logic is operationally correct at the individual job level. It is strategically wrong at the governance level.

A 62-unit mining fleet contains: primary production equipment (excavators, primary haul trucks) whose failure stops production entirely; secondary production equipment (secondary haulers, drills) whose failure reduces production rate; support equipment (graders, water trucks, light vehicles) whose failure disrupts logistics without stopping production. Managing all three categories with identical priority — identical spare stock governance, identical PM compliance focus, identical CBM programme coverage — misallocates every maintenance resource toward the lowest common denominator.

The Three-Tier Criticality Framework

What Changes When Criticality Classification Exists

Workshop resource allocation. When the primary excavator and the grader both need workshop time, the Supervisor who has a criticality classification can make the resource allocation decision based on objective criteria — not on which supervisor asked first or who makes more noise. Tier A always takes priority over Tier B, which always takes priority over Tier C. The decision is pre-made by the classification system.

Spare parts governance. C1 spares (maximum safety stock, ERP auto-trigger) are only required for Tier A and selected Tier B components. Tier C components are managed by standard procurement logic. This concentration releases working capital from non-critical spare provisioning and redirects it to critical spare pre-positioning — without increasing total inventory budget.

CBM programme coverage. Full oil analysis programme coverage at 250-hour intervals is justified for Tier A equipment where downtime costs $18,000/hr. The same programme for a Tier C grader — whose failure costs $200/hr in logistics disruption — cannot be justified at the same cost and frequency. Criticality classification allows the CBM programme to be calibrated to the production consequence of each equipment class rather than applied uniformly to the entire fleet.

RE time allocation. In the average 50-unit fleet, Tier A equipment comprises 20–30% of units and accounts for 60–75% of production loss from downtime. An RE who allocates their analytical time proportionally to criticality is spending 60–75% of their time on the failures that matter most. An RE without criticality guidance allocates time by failure frequency or supervisor priority — which may be entirely unrelated to production consequence.

Implementing the Classification — The Four-Step Process

Step 1: List all fleet assets and assign each to Tier A, B, or C using the production consequence criteria. This is a 2-hour exercise with the Mine Director, Operations Superintendent, and Maintenance Manager. Step 2: Enter the classification in the CMMS asset hierarchy — every WO on every Tier A asset is flagged at creation as Priority Critical. Step 3: Review spare parts inventory against the classification — all long-lead components for Tier A equipment become C1 classification items requiring safety stock. Step 4: Present the classification to the board as the basis for the maintenance priority allocation model. This is the governance artefact that converts maintenance resource allocation from intuitive to evidence-based.

The Benchmark Trap

A mining CFO receives a maintenance benchmarking report showing that the operation's CAT 793F haul truck maintenance cost per operating hour is USD 142/hr — against an industry benchmark of USD 96/hr. The CFO concludes there is a maintenance cost problem and requests an explanation. The Maintenance Manager provides one that the CFO cannot evaluate. The conversation is circular, produces no action, and consumes three meetings.

The correct response to a raw benchmark comparison is not an explanation — it is a question: "Is this benchmark adjusted for the five factors that make our operation different from the benchmark population?" If the answer is no, the benchmark comparison is not yet meaningful. It is a starting point for a more specific analysis, not a conclusion.

The Five Benchmark Adjustments

Three Situations Where a "High" Benchmark Is Not a Problem

Situation 1: You are operating an aging fleet in a high-duty cycle environment. If your average fleet age is 8 years and the benchmark population averages 5 years, your CPH will be structurally higher — and should be. The question is whether your CPH is appropriate for your fleet age, not whether it matches an age-unadjusted benchmark.

Situation 2: You are in the early phase of a reliability improvement programme. During S3 programme execution — when reactive failure costs are being eliminated systematically — CPH may initially increase as deferred maintenance is executed and as the cost of eliminating recurring failures (which involves repairing the root cause, not just the symptom) is incurred. This is the cost of transition, not the cost of failure. It normalises within 6–9 months.

Situation 3: Your operation is in a high-cost parts supply environment. If benchmark data is derived from Australian or North American operations and you are comparing an Indonesian or West African operation, the parts cost index differential alone accounts for 15–30% of CPH difference. This is not a maintenance performance issue. It is a supply chain context issue that no amount of reliability improvement will eliminate.

The Situation — Productive Time Invisible to Management

The Mine Director of this operation had approved two additional technician headcount positions in the prior budget cycle — both unfilled due to recruitment difficulty. The assumption: the maintenance team did not have enough capacity for the maintenance workload. The finding from MitWin's S3 Phase 1 assessment: the team had more than enough capacity. It was consuming it on waiting, travel, and administration rather than on maintenance.

A 3-day work sampling study, conducted by the MitWin Field Consultant across all three shifts, produced the following breakdown of a standard 10-hour shift:

The Mine Director reviewed these findings and asked: "How is it possible that we are spending 2.1 hours per technician per shift waiting for parts?" The answer: the planning function operated on a 2-day horizon with no parts pre-staging protocol. Every job that started without pre-staged parts produced a waiting event.

The Eight-Week Intervention

Week 1–2 — Parts Pre-Staging Protocol. Every job scheduled for the following week requires parts confirmed available in stores by Wednesday 15:00. Any job without confirmed parts is removed from the Week 1 schedule and replaced with a job where parts are available. The Planner is the enforcement mechanism. The result: parts waiting time reduced from 2.1 hours to 0.4 hours per shift within 3 weeks.

Week 2–4 — 4-Week Planning Horizon. Planner removed from reactive coordination. Equipment release meeting established Thursday 14:00 with Operations Superintendent. 4-week rolling schedule built for the first time. Reactive task switching (where supervisors re-allocated technicians mid-job to respond to breakdowns) reduced from 1.4 hours to 0.6 hours per shift — because the planning system could absorb some reactive events without disrupting the full planned schedule.

Week 3–5 — Briefing Compression. Shift briefing redesigned from a 40-minute informational session to a 15-minute stand-up format: what broke, who is on it, what is planned today. Pre-task analysis forms moved to digital completion at the work face. Shift handover structured to 12 minutes with a standard handover template. Administrative overhead reduced from 1.6 hours to 0.9 hours per shift.

Week 6–8 — Satellite Parts Staging. A parts delivery run was established from the main warehouse to the primary excavator working area twice daily (07:30 and 13:30). Technicians assigned to the excavator area no longer need to travel to the main workshop for parts retrieval — the parts travel to them. Travel time for excavator team reduced from 1.8 hours to 1.1 hours per shift.

Results at Eight Weeks

The Investment That Was Not Producing Return

The Mine Director's question to MitWin was direct: "We spent $1.8M on predictive maintenance technology. Eighteen months later, we cannot demonstrate a single prevented failure. What is wrong with the technology?" The answer, after a 2-week diagnostic, was: nothing. The vibration sensors were accurately measuring bearing condition. The AI analytics engine was correctly identifying anomaly patterns. The technology was working as designed. The organisation was not ready to act on what it was detecting.

Three diagnostic findings confirmed the governance failure:

Finding 1 — Alert fatigue was complete. In Month 12, the system generated 340 alerts across the monitored fleet. The Maintenance Supervisor responsible for alert review — a role that had been added to his existing responsibilities without time allocation — had reviewed 23 of the 340 alerts. The other 317 were marked "read" and archived. The Supervisor was not failing at his job. He was being asked to review 11 alerts per day alongside his existing supervisory responsibilities. It was structurally impossible.

Finding 2 — No alert-to-WO workflow existed. For the 23 alerts that were reviewed, there was no configured pathway from an alert assessment to an automatically generated work order. The Supervisor had to manually create a WO, assign it, and add it to the schedule — a 20-minute administrative task per alert. Of the 23 reviewed alerts, 4 generated WOs. Of those 4, 2 were scheduled for execution within 2 weeks and 2 were deferred to the next maintenance window (4–6 weeks away).

Finding 3 — No baseline data existed for alert threshold calibration. The system's alert thresholds had been configured using OEM default values rather than equipment-specific operational baselines. The the operation site operating environment — high temperature, high vibration from haul road surface, high dust loading — produced vibration profiles that triggered OEM-threshold alerts on healthy equipment. The alerts were not wrong. The thresholds were uncalibrated.

The Ten-Week Governance Activation

Weeks 1–2 — Alert Threshold Recalibration. The MitWin Reliability Lead worked with the technology vendor to recalibrate all alert thresholds from OEM defaults to equipment-specific operational baselines — derived from 4 months of sensor data from known-healthy equipment in the current operating environment. Alert volume reduced from 340/month to 28/month — all 28 now representing genuine anomalies rather than baseline operating noise.

Weeks 2–4 — CBM Specialist Role Defined and Filled. A dedicated CBM Specialist role — reporting to the RE, with 80% of working time allocated to alert review and CBM programme management — was defined and filled from an internal candidate (a senior technician with 12 years experience who had expressed interest in a reliability pathway). The specialist received 2 weeks of structured training from the technology vendor and from MitWin's oil analysis interpretation module. Alert review became a dedicated function rather than an add-on to a supervisor's existing responsibilities.

Weeks 3–5 — Alert-to-WO Workflow Configuration. The CMMS was configured so that any alert above a defined severity threshold (vibration amplitude >2.5× operational baseline) automatically generated a WO flagged "PdM Alert — Priority — Do Not Defer." Below that threshold, the CBM Specialist reviewed and either generated a WO or documented "within normal parameters" with a signature. Alert-to-WO time reduced from an average of 5.4 days (for the 11 alerts that had generated WOs) to same-day for auto-generated, and 24 hours for Specialist-reviewed.

Weeks 5–10 — PdM Alert Review Integration into Weekly Reliability Meeting. A PdM alert summary became a standing agenda item at the weekly reliability review: how many alerts were generated, how many generated WOs, how many interventions were conducted, and — most critically — how many failures did the system detect before breakdown vs after. The failure detection rate became a trackable, reportable KPI for the first time.

Results at Ten Weeks and at Month 12

The Cold Weather Gap

The operation's maintenance strategy had been designed by the OEM during the initial mine development — before the full impact of the arctic operating environment was understood. Six years into operation, the maintenance team had accumulated extensive experience of what failed, when it failed, and under what conditions. What they had not done — and what required an RE with cold-weather mining experience to accomplish — was connect the failure patterns to the specification mismatches that caused them.

Five failure modes had been recurring for more than 3 years without elimination. Each was treated as an environmental inevitability — "of course hydraulic systems have more problems in the cold, that is just what happens up here." The MitWin finding: none of the five failure modes were inevitable. Each was a direct consequence of an OEM specification that had not been adapted to the operating environment.

The Five Specification Mismatches

Results at Six Months

The RE Who Could Diagnose but Could Not Change Anything

The RE at this operation was technically competent. The CMMS analysis she had produced was accurate — correctly identifying 6 of the 10 active recurring failure modes with plausible root cause hypotheses. The problem was structural: she had no authority to implement corrective actions. Supervisor compliance with her recommendations was voluntary. Her corrective action recommendations had been presented at 3 consecutive monthly maintenance meetings — noted, acknowledged, and not implemented.

When MitWin arrived, the RE's assessment was candid: "I know what needs to change. I cannot make anyone change it. The supervisors agree in the meeting and then continue the same practices the next week. I have no idea what to do next." This is one of the most common failure modes of the under-structured RE role in mining: technical capability with no authority pathway.

The Governance Framework MitWin Installed

The Recurring Failure Register — Formal and Named. All 10 active recurring failure modes were entered into a formal register — each with: failure description, equipment affected, occurrence count, estimated annual cost, named RE as investigation owner, and target elimination date. The register was presented at every weekly reliability review and every monthly executive review. Any mode that had not advanced in a 30-day period was escalated to the Mine Director with an explanation of the blocker.

Corrective Action Authority — Written Policy. The Mine Director signed a written policy confirming that corrective actions issued by the RE — documented in the register with the RE's signature — were mandatory for all supervisors to implement within the timeframe specified. A supervisor who did not implement a corrective action without a legitimate reason (material unavailability, safety concern, technical alternative) was accountable to the Mine Director directly. This was the structural change that broke the "noted and not implemented" cycle.

Monthly Review — Register on Slide 1. The monthly executive review was restructured to present the Recurring Failure Register as the first item — before any other KPI data. This established the accountability posture of the meeting before new information produced new distractions. The Mine Director opened every monthly review by asking: "Which modes have been eliminated since last month and which have not advanced? Why?"

The Ten Failure Modes and Their Status at 14 Months

The Governance Decay Pattern

Reliability governance improvement erodes without active maintenance — this is the fundamental reason S6 exists. The rate of erosion is determined by three factors: the depth of institutional embedding of each governance element, the frequency of staff turnover in key governance roles, and the presence or absence of an external accountability mechanism. At this operation, all three factors compounded against the improvement.

The original programme had not included S6 governance partnership — the Mine Director at the time had been confident that the internal team could sustain the system independently. The Maintenance Manager who had been the programme's champion was an exceptional individual — and his departure removed the institutional memory that had been carrying the governance system. His replacement, a capable operations professional from the pit management side, did not know what the governance system was, what it required, or why it had been implemented.

What the New Mine Director Found

Arriving 18 months after programme close, the new Mine Director commissioned a review of the maintenance system. The review found:

Every governance element that had been installed during the programme had partially or fully reverted. The equipment was the same. The CMMS was the same. The maintenance team was largely the same (with some turnover). But the system that had connected these elements into a functioning reliability governance structure had quietly disintegrated over 18 months of unchecked erosion.

The Six-Week Governance Restart

The re-engagement was structured as a restart, not a rebuild. The documentation from the original programme still existed — meeting templates, KPI definitions, action register formats, PM task library, CBM programme specifications. The task was to reinstate what had existed rather than recreate it from scratch. This compressed the timeline significantly.

Week 1: All 6 governance elements reinstated simultaneously. Daily stand-up resumed Monday of Week 1. Weekly reliability review resumed Wednesday of Week 1. Monthly executive review booked for end of Month 1. 4-week rolling schedule rebuilt from scratch (the planning system had not been maintained). CBM alert-to-WO workflow reconfigured in the CMMS — all outstanding oil analysis alerts reviewed and those with threshold breaches generating immediate WOs.

Week 2–4: Recurring failure register rebuilt — all 12 active modes documented, named owners assigned, investigation timelines set. RE given corrective action authority by written policy from new Mine Director. First 3 corrective actions implemented (desiccant breathers, air filter interval correction, magnetic plug protocol). Governance calendar for the next 12 months locked in all relevant calendars.

Week 5–6: S6 Governance Partnership approved — monthly KPI Pack production, monthly executive review facilitation, quarterly board summary. New Mine Director committed to sustaining the governance cadence with MitWin external accountability.

Results at 90 Days

What Gets Lost at Every Shift Change

Shift handover in most mining operations consists of a verbal conversation — often less than four minutes — and a logbook entry that describes what broke, not what was developing. What the outgoing crew knows but rarely transfers: subtle behaviour changes in equipment that have not yet triggered a fault code, service interventions that were performed but not yet logged in the CMMS, near-misses and operator observations that did not warrant a formal report, and the sequence and context of events that experienced technicians use to build a diagnostic picture.

Without structured handover, the incoming crew starts with a zero-state situational awareness on every machine. Developing failures that the outgoing crew was monitoring — informally, from experience — are invisible to the incoming crew. The machine continues operating. The failure progresses.

The Five Categories of Information That Must Transfer

The Financial Calculation

MitWin's shift handover loss estimate for a 44-unit fleet is built from three cost categories. First: developing failures that progress through two additional shifts due to information non-transfer, producing breakdown events that would have been PM interventions if detected at the first-shift observation. Average: 3.8 events per month at $28,000 per event = $107,000 per month. Second: repeated diagnostic time — each incoming technician spending 15–25 minutes re-establishing the situational awareness the outgoing crew already had. At 3 crews × 8 technicians × 22 working days × 20 minutes = 88 hours per month of recoverable diagnostic time. Third: repeat interventions — services performed twice because the first performance was not transferred. These three categories sum to $380,000–$440,000 per year across a 44-unit fleet — scaling proportionally with fleet size.

The Structured Handover Document — Five Fields, Eight Minutes

The fix is not a longer briefing. It is a structured document with five mandatory fields, completed by the outgoing Supervisor before leaving site, reviewed by the incoming Supervisor in the first 8 minutes of shift. The five fields map directly to the five categories above. Total completion time for the outgoing Supervisor: 8–12 minutes. The incoming Supervisor review time: 6–8 minutes. The diagnostic value: eliminates the zero-state situational awareness problem that costs $380,000 per year.

The document is configured in the CMMS as a mandatory shift-end entry — not a paper logbook. Mandatory means it cannot be bypassed: the outgoing Supervisor cannot complete end-of-shift sign-off without submitting the five-field handover form. Within 90 days of implementing this requirement at a 44-unit operation, MitWin typically observes a 40–55% reduction in the repeat-diagnostic-time category and an 18–24% reduction in developing-failure-to-breakdown conversion rate.

What Leaves When an Expert Leaves

The departing Superintendent knows which haul truck always develops a steering pull before the steering box fails. They know which section of the underground drive produces the most aggressive tyre sidewall damage and why. They know which component in the CAT 793 rear axle emits a specific sound pattern approximately 180 operating hours before bearing failure — a pattern that is not in any OEM manual and is not measurable by any current sensor in the fleet. They know which supplier's hydraulic filter media is marginal for the site's contamination profile. And they know which junior technician is developing the diagnostic capability to eventually replace them — and which one is not.

None of this knowledge is in the CMMS. None of it is in any document. All of it exists in pattern recognition built across 18 years of working on specific equipment in a specific environment — and all of it departs on the last day of employment.

The Three-Part Knowledge Capture Protocol

The Cost of Not Doing This

In MitWin's engagement data, operations that lose a long-tenure senior technical expert without knowledge capture experience a statistically measurable deterioration in MTBF for the equipment classes that expert specialised in — typically within 90 days of departure. The mechanism: failure modes that the departing expert would have flagged from pre-failure observation begin reaching functional failure without intervention. Average MTBF impact: 12–18% deterioration in the primary equipment class over the subsequent 6 months. At a haul truck fleet with an average breakdown cost of $28,000 per event, a 15% MTBF deterioration across 12 units equals approximately $580,000 in additional breakdown cost over 6 months.

The three-part protocol described above requires approximately 40 hours of structured time from the departing expert and 20 hours from the Reliability Engineer — over a standard 4-week notice period. The cost: two salaries for four weeks of structured knowledge work. The value: prevention of $580,000 in MTBF-deterioration-related breakdown cost, plus the ongoing value of preserved institutional knowledge in the CMMS.

Why Road Construction Is Different From Mining

Mining operations run equipment in a fixed location with consistent environmental conditions. A road construction fleet moves from project to project — each with different material characteristics, climate, haul distances, gradient profiles, and operating hours per day. The maintenance strategy that was correct for the last project may be materially incorrect for the next one. In practice, most road construction contractors apply the same OEM-based PM schedule to every project, adjust it informally based on supervisor experience, and absorb the cost of premature failures on high-stress projects as "project variability" rather than as a strategy failure.

MitWin's road construction engagement data shows that over-servicing on low-stress projects adds 8–14% to maintenance cost without improving reliability. Under-servicing on high-stress projects produces component failures 20–35% earlier than OEM-expected intervals — increasing breakdown frequency and emergency part procurement. Both are symptoms of an undifferentiated strategy applied across differentiated conditions.

The Project Stress Classification (PSC) Framework

Before mobilisation to any new project, MitWin recommends a 2-hour PSC assessment that scores the project on five dimensions and assigns each piece of mobilised equipment to a stress class — determining whether the standard OEM interval should be applied, shortened, or extended for that specific project context.

Total PSC score: 5–25. Low stress (5–10): standard OEM intervals apply. Medium stress (11–17): reduce fluid and filter intervals by 20–30%. High stress (18–25): reduce by 40–50% and add a condition monitoring check at 50% of each interval. A motor grader on a basalt highway project in East Africa at 38°C with 13-hour operating days scores 22–24 — requiring air filter replacement at 200 hours vs the OEM default 500 hours and engine oil change at 150 hours vs 250.

Pre-Mobilisation Spares Calculation

The PSC score drives the pre-mobilisation spares calculation. A high-stress project with reduced PM intervals will consume 40–60% more air filters, hydraulic filters, and engine oil per operating month than the same equipment on a low-stress project. The contractor who mobilises to a remote site with a standard spares kit and a high-stress project profile will be emergency-procuring at air freight rates within 6 weeks. Emergency procurement on a remote road project — typically 12–18 hours from the nearest parts supplier — costs 2.8–4.2× the standard part price when logistics, urgency premium, and downtime are included.

The PSC-driven spares calculation: multiply each standard monthly consumption by the stress adjustment factor (Low=1.0, Medium=1.3, High=1.6), add 20% buffer for remote location risk, and build the initial mobilisation kit from this number. The calculation takes 30 minutes per equipment class. The alternative — discovering the consumption shortfall at Week 6 on a remote project — takes three times longer and costs significantly more.

Symptom vs Cause — The Diagnostic Value Gap

A symptom is an observable condition: "oil leak," "machine overheating," "excessive vibration," "won't start." A cause is the engineering explanation for why the symptom occurred: "external abrasion of hydraulic hose from contact with unprotected frame edge," "thermostat failure — wax element degradation," "propshaft universal joint — needle bearing collapse," "battery positive terminal — corrosion buildup exceeding contact resistance threshold." The difference between these two levels of description is the difference between a repair log and a reliability database.

When the CMMS contains 24 months of symptom-level WOs for a 44-unit fleet, the Reliability Engineer running an analysis on hydraulic failure frequency will find: 47 entries coded "hydraulic leak." They will not find: 22 entries coded "hose abrasion — frame contact" and 25 entries coded "fitting fatigue — inadequate support clamp." They will not know that the 22 hose abrasion events are all concentrated on machines 04, 11, and 17 — all of which have the same frame modification from the 2021 maintenance campaign. They will not know that the 25 fitting fatigue events are a supplier quality problem with the batch purchased in Q3 2023. Two distinct, solvable problems appear as one undifferentiated category. Neither is eliminated.

Why Symptom Coding Happens

Technicians code symptoms rather than causes for three structural reasons — none of which are carelessness or incompetence. First: they are not required to identify the cause before closing the WO. The CMMS allows WO closure with a symptom-level code. Remove this permission and the behaviour changes. Second: identifying the root cause requires diagnostic analysis that the technician may not have the time or tools to complete at the point of repair — especially under time pressure to return the machine to service. Third: the failure code taxonomy in most CMMS systems is built around component categories, not failure mechanisms. "Hydraulic system — leak" is a valid code. "External abrasion — inadequate hose routing" is not in the taxonomy. Technicians code what the system allows, not what the situation requires.

The Four-Field WO Closure Standard

The Downstream Value

With four-field WO closure standard applied consistently for 6 months, the RE's analysis changes fundamentally. The 47 "hydraulic leak" entries become two distinct failure modes — each with a clear root cause, a specific affected machine population, and a corrective action that eliminates the failure entirely. The CMMS transitions from a cost ledger to a reliability engineering database. Failure elimination becomes possible from data analysis rather than from informal observation and experience alone.

How Most Rebuild Decisions Are Made

In practice, the workshop rebuild vs replace decision is made on three informal criteria: what the Foreman has seen before ("last time we rebuilt this transmission it lasted 6 months — replace"), what the parts availability looks like today ("reman unit is in stock — take it"), and what the time pressure is ("we need this machine back in 6 hours — what is faster?"). All three criteria are legitimate operational inputs. None of them are economic inputs. None of them answer the question: which option delivers the lower total lifecycle cost over the next 18 months?

The consequence: operations systematically over-replace when rebuild would have been the correct economic decision. A transmission rebuild at $28,000 that delivers 14,000 hours of additional life is cheaper than a reman unit at $64,000 delivering the same life — by $36,000. Across 12 such decisions per year on a 44-unit fleet, the systematic preference for replacement costs approximately $1.6M more than the rebuild-when-viable alternative.

The Three-Question Rebuild Economics Test

When to Override the Economics

Three legitimate override criteria: (1) The component's condition makes rebuild technically infeasible — bore wear beyond machining tolerance, crack detected in critical stress zone, casting damage beyond repair spec. (2) The machine's remaining fleet life is shorter than the rebuild life — rebuilding a component to last 14,000 hours on a machine that will be retired in 8,000 hours is not economic. (3) The rebuild turnaround time exceeds the operational availability requirement and no loaner or substitute unit is available. These three exceptions are real — but they are encountered in approximately 20–25% of rebuild decisions, not the 60–70% rate at which most operations default to replacement.

Implementing the Test Without Slowing the Workshop

The test does not require a spreadsheet or a specialist. It requires a 10-cell calculation table on a laminated card at the Workshop Foreman's desk — the six input fields and four calculated outputs printed in the format the Foreman uses every day. MitWin builds this tool as part of every S5 engagement, pre-populated with the equipment class-specific data for each major component category. Total calculation time: 4 minutes per decision. The Foreman retains full decision authority — the test provides the economic output that anchors it.

What a Backlog Actually Tells You

A maintenance backlog has a correct size — approximately 2–4 weeks of available craft hours. Below this: the planning function has insufficient forward work to keep the team efficiently deployed. Above 4 weeks: work is being deferred faster than it is being completed. The causes of above-target backlog are predictable: reactive work volume displacing planned work execution, planner-to-unit ratio too high for effective scheduling, equipment release windows too narrow or too unpredictable, and — most commonly — WOs being opened for work that the organisation has no current intention of executing.

The 12-week-plus WO is the most diagnostic element of the backlog. Work that has remained unexecuted for 12 weeks typically falls into one of four categories: genuinely deferred work that represents real maintenance liability; work that was raised but the parts are on back-order; work that was superseded by a different repair but was never closed; or work that was raised as a precaution but the condition has since resolved. All four categories require a decision — and the WO that sits for 12 weeks without a decision is compounding operational risk and inflating the backlog number with work that does not represent real, actionable liability.

The Backlog Age Distribution Analysis

The 90-Day Backlog Reduction Protocol

Week 1–2: Triage. Every WO older than 8 weeks is reviewed by the Supervisor and Planner together. Decision required for each: Execute this month / Defer with explicit date / Close — superseded or resolved. The triage eliminates ghost WOs (closed condition, no longer valid) and explicitly deferred WOs (now have a date, not just an age). Target: reduce 12-week+ WOs by 60% through triage alone.

Week 3–8: Dedicated capacity. 20% of available craft hours allocated weekly to backlog reduction — not to new reactive work. This requires operations agreement: 20% of maintenance capacity is contractually committed to backlog work each week. The Mine Director must sanction this allocation — it will be challenged every time a breakdown competes for the capacity.

Week 9–12: Governance lock. Weekly backlog age report to Mine Director. Any WO crossing the 8-week threshold without a decision triggers an automatic Supervisor alert. Target by Week 12: backlog age distribution within benchmark on all five bands.

The Three Production Growth Levers — Modelled

Consider a standard open-cut coal operation: 44 haul trucks, 76% average availability, 20 operating hours per day, 340 operating days per year. Current production baseline: 18.4 Mtpa. Production target: 21.0 Mtpa — a 14% increase required.

The Financial Comparison

Why Boards Choose Fleet Expansion First

The fleet expansion decision is tangible. A new truck can be photographed. It appears on the asset register. It represents visible capital deployment that shareholders can interpret as growth investment. Availability improvement is intangible — it does not add an asset, it recovers the output of assets that already exist. This perceptual asymmetry drives boards toward the option that is visible, delayed, and capital-intensive — and away from the option that is invisible, immediate, and capital-light.

The CFO who reframes availability improvement in language the board understands — "recovering 4.2 unit-equivalents of productive capacity for $320,000 vs adding 4.2 physical units for $22M" — changes the conversation. The availability case does not compete with fleet expansion on emotional terms. It competes on financial terms — and it wins on every metric except physical visibility.

The Sequencing Recommendation

MitWin's recommendation for any operation facing a production growth requirement: exhaust the availability improvement lever first. Commission S1 to quantify the availability gap and its financial value. Commission S3 to recover it. At the 12-month mark, assess remaining production gap after availability improvement. If further growth is required: the fleet expansion decision is now made with a reliable, well-governed fleet as the foundation — rather than adding units to a fleet with structural availability problems that will simply scale the downtime proportionally.

The Risk Allocation Illusion

Most operators believe that contracting out maintenance transfers the maintenance risk. This belief is frequently incorrect. A contract that includes force majeure clauses covering "abnormal operating conditions," penalty caps at 5% of contract value, and availability guarantees measured as fleet average rather than critical-unit uptime creates the appearance of risk transfer while preserving most of the financial exposure for the operator. The contractor's revenue is protected. The operator's production is not.

Ten Clauses That Create Operator Liability

Five Provisions That Genuinely Protect the Operator

The CBM Technology Evaluation Framework

Every CBM technology must justify its deployment cost against three questions: (1) What specific failure modes does it detect that other methods cannot? (2) What is the P-F interval it provides — the warning time between first detection and functional failure? (3) What analyst competency, sampling frequency, and threshold calibration does it require to deliver that P-F interval reliably? A CBM technology deployed without the competency and frequency its P-F interval requires is not a CBM programme — it is an inspection activity with intermittent diagnostic value.

Oil Analysis — The Baseline Technology

Oil analysis detects internal wear through elemental spectroscopy (iron, copper, lead, chromium — each sourced from specific component wear surfaces), contamination through particle count and silicon measurement, and lubricant degradation through viscosity, TAN/TBN, and water content. Its strength is its directness: wear metals in oil are physically present in the sample, making the analysis less dependent on analyst interpretation than vibration or thermography.

Its limitation: it cannot detect structural failures, electrical faults, or external damage modes. A haul truck that develops a cracked frame rail will not produce an oil analysis signal. A final drive that develops an external housing crack from impact damage will not appear in the SOS report until the lubricant is contaminated — at which point failure is imminent rather than developing. Oil analysis is necessary but not sufficient for a comprehensive CBM programme on heavy mobile equipment.

Deployment quality requirement: sample frequency must be calibrated to the P-F interval of the target failure mode — not to a convenient service schedule. For engine wear metal monitoring with a P-F interval of 400 hours, sampling at 500-hour service intervals misses approximately 20% of developing failures before the next sample opportunity.

Vibration Analysis — The Most Under-Deployed and Most Misapplied

Vibration analysis on rotating equipment can detect bearing defects 200–1,400 hours before functional failure (bearing type and load dependent), gear damage 300–800 hours before failure, and rotor imbalance from the first measurement cycle if correctly calibrated. It offers the longest P-F intervals of any CBM technology — which makes it the highest-value prevention tool for major rotating components.

Its deployment challenge: the diagnostic value is entirely dependent on analyst competency. A vibration spectrum collected from a CAT 793 wheel motor requires an analyst who can distinguish bearing defect frequencies from normal operating noise at the specific speed and load conditions of the measurement. Without this competency — which requires 200–400 hours of supervised practice on mining equipment — vibration analysis produces data that is collected, filed, and never acted on. The measurement programme cost is incurred. The prevention value is not captured.

MitWin's recommendation: vibration analysis is justified on rotating equipment where bearing or gear failure consequence exceeds $80,000 per event — typically the wheel motors, final drives, and swing circles on primary loading equipment. Below this consequence threshold, oil analysis and scheduled inspection provide adequate detection at lower deployment complexity.

Thermography — The Most Correctly Deployed

Thermal imaging on electrical connections, brake systems, and hydraulic circuits is the CBM technology most consistently deployed at correct frequency and with adequate analyst competency in mining operations. The P-F interval is short (hours to days for electrical connection overheating) — which makes monthly or quarterly inspection routes appropriate and makes the detection event clearly actionable. A thermal image showing a brake caliper at 180°C above ambient requires no specialist interpretation: the brake is overheating and must be investigated immediately.

The deployment gap: most thermography programmes are applied to electrical switchgear and ignored on hydraulic circuits and brake systems — where failure consequences are often higher. A motor grader hydraulic cylinder operating at 30°C above normal temperature for the circuit is developing a bypass condition that will produce a steering or blade control failure within 200–400 operating hours. Thermography detects this. Most thermography programmes do not include hydraulic circuit mapping as a standard route item.

The Integrated CBM Architecture

A mining operation's CBM programme should be designed as an integrated architecture — each technology deployed for the specific failure modes and equipment classes where its P-F interval and detection capability justify its cost. MitWin's standard architecture for a 44-unit heavy mobile equipment fleet: oil analysis on all major fluid systems at calibrated intervals; thermography quarterly on electrical systems and monthly on brake and hydraulic circuits; ultrasound on compressed air systems and high-value bearings; vibration analysis on wheel motors, swing circles, and final drives of primary loading equipment; motor current analysis on all critical pump and conveyor motors above 75 kW.

What a Job Plan Actually Is — and Why Most Operations Don't Have One

A job plan is a document that makes a maintenance task executable without improvisation. It contains: a defined work scope (what exactly is to be done — not "service engine" but "drain engine oil, replace oil filter, replace air filter primary and secondary elements, check and record valve clearances against spec, replace coolant additive at prescribed rate, complete post-service oil sample SOS"), a parts list with part numbers and quantities confirmed available in warehouse, a tools list including any special tools required and their location, a skill requirement specification (what trade level and what minimum experience), an estimated duration in hours, a safety requirement (permits required, isolation points, relevant JSA reference), and a post-completion verification step (what the supervisor must check before signing the WO as complete).

In most operations, a WO contains: a description field with "engine service" and a PM type code. That is not a job plan. It is an instruction — and an incomplete one. Technicians executing from incomplete instructions improvise. Improvisation introduces variability. Variability produces quality failures. Quality failures produce repeat work. The planning failure that produced the inadequate WO is invisible — the cost appears in the CMMS as a breakdown or a callback, attributed to a technician or a part, not to the planning function that made the failure structurally predictable.

The Planner's Three Core Outputs

What Consumes Planner Time — and Shouldn't

In MitWin's planning function assessments, the following activities consistently consume 50–70% of planner time at operations where Job Plan Coverage Rate is below 25%: answering supervisor phone calls about parts availability for reactive work; attending breakdowns to coordinate labour and parts for unplanned events; generating reports and KPI summaries for management review meetings; updating the CMMS with data entry that should be done by the technician or supervisor closing the WO; and attending operations coordination meetings that do not require planner presence.

Every hour spent on these activities is an hour not spent on job plans, parts confirmation, or the 4-week schedule. The structural fix is not working faster — it is removing these activities from the planner's responsibility list. Each belongs to a different role: supervisor, coordinator, administrator, or systems user. Restoring planner time to planning work recovers $800,000–$1.6M per year in maintenance execution quality on a fleet of 40+ units — without a single additional hire.

The Job Plan Coverage Rate — the Planning Function's Single KPI

Job Plan Coverage Rate (JPCR): the percentage of planned Class A WOs that executed from a complete job plan (scope, parts, tools, skill, time, safety, verification) in a given week. Target: >80%. Below 50%: planning function is not functioning as designed. Below 25%: planning role is performing a different function — likely reactive coordination. Above 80%: the planning function is delivering its designed value. The JPCR is tracked weekly and reported to the Maintenance Manager. It is the earliest leading indicator of planning function degradation — a declining JPCR precedes MTBF deterioration by 6–10 weeks.

Why Acquisition Cost Is the Wrong Primary Criterion

Acquisition cost is precisely known at procurement stage — it is the number on the OEM quote. Total lifecycle cost is estimated — it requires modelling assumptions about fuel consumption, maintenance cost per hour, parts pricing over 10+ years, and residual value. Because the acquisition cost is certain and the lifecycle cost is estimated, procurement decisions systematically overweight the certain number and underweight the estimated one. The discipline required to override this cognitive bias — in a capital allocation process that involves Finance, Operations, and the Board — is significant. But the financial consequence of not overriding it is larger.

Example: two haul truck models in a 12-unit procurement decision. Model A: acquisition cost recurring failure cost/unit, fuel consumption benchmark 148L/hr, maintenance CPH $84, expected life 38,000 hours, residual value 8%. Model B: acquisition cost $6.4M/unit, fuel consumption benchmark 138L/hr, maintenance CPH $71, expected life 42,000 hours, residual value 11%. Model A is $600,000/unit cheaper to acquire. Over the expected fleet life at 5,500 operating hours/year, Model B is significant emergency procurement cost/unit cheaper in total lifecycle cost. For a 12-unit procurement: Model A saves $7.2M at acquisition. Model B saves $14.9M over fleet life. The acquisition cost decision costs $7.7M.

The Four-Input Lifecycle Cost Model

Five Procurement Clauses That Protect Total Lifecycle Cost

Clause 1 — CPH Guarantee for the First 15,000 Hours. OEM or dealer warrants that maintenance CPH will not exceed the procurement model assumption by more than 10% in the first 15,000 operating hours. Shortfall compensated through extended warranty or parts credit.

Clause 2 — Parts Availability Commitment for Minimum 15 Years. OEM commits to maintain new parts availability for a minimum of 15 years from last unit delivery in the procurement batch. Critical components identified in the procurement model are named specifically.

Clause 3 — Component Rebuild Support Agreement. OEM or authorised rebuilder commits to component rebuild support (engine, transmission, final drive) for the fleet life — at pricing agreed at procurement stage, not at time of rebuild.

Clause 4 — Fuel Consumption Guarantee Under Defined Duty Cycle. OEM warrants fuel consumption under the specific duty cycle parameters (grade, payload, ambient temperature range) defined at procurement. Exceedance above 8% triggers a diagnostic review at OEM cost.

Clause 5 — Fleet Life Extension Option. Operator retains the right to extend fleet life beyond the original modelled life — with OEM committing to the availability of component rebuild support, parts, and field technical expertise for the extension period. This prevents the fleet exit being forced by support withdrawal rather than by economic crossover.

What 36 Months of Stalled Availability Actually Means

A 180-unit fleet stuck between 72% and 76% availability for 3 years is not a technical plateau — it is a governance equilibrium. The maintenance execution team is performing consistently with the governance structure around them. They have been selected, trained, and conditioned to operate within that structure. When the structure is not designed to improve reliability systematically, it produces consistent — not improving — results. Replacing the Maintenance Manager three times does not change this equilibrium. It adds personnel disruption to a structural problem.

At the engagement commencement meeting, the Mine Director described a pattern that MitWin recognises immediately: "We generate the weekly availability report. We discuss it in the Monday meeting. We identify the machines that were down. We assign action items. The same machines are down the following Monday." This is not a maintenance execution failure. It is a reliability intelligence failure — the meeting discusses symptoms, assigns reactive responses, and produces no mechanism for recurring failure elimination. The governance cadence is active. It is producing no systematic improvement.

S1 Findings — MRMM 47/80, CREM — significant annual exposure

The S1 audit identified 14 distinct governance failures across the five reliability architecture layers. The four highest-value: (1) CBM alert-to-WO workflow absent — oil analysis and vibration data generated at industry-leading frequency but not triggering work orders automatically; (2) Failure elimination function absent — 23 recurring failure modes active for 12+ months with no systematic elimination programme; (3) Reliability review meeting producing no decisions — meeting attended by 18 people, producing action items assigned to absent parties; (4) Planning function consumed by reactive coordination — Job Plan Coverage Rate measured at 11% against 80% target.

S6 Governance Architecture — 90-Day Installation

Priority 1 — CBM Integration (Days 1–30): The CMMS configured to auto-generate WOs from oil analysis threshold exceedances. Vibration analysis alert-to-WO workflow configured for all 24 excavator swing circles and wheel motors. First automated WO generated Day 8. Within 30 days: 34 developing failures detected and addressed before breakdown — estimated 2,100 machine-hours of unplanned downtime prevented in Month 1 alone.

Priority 2 — Failure Elimination Programme (Days 14–60): RE team (3 REs in post, previously generating reports) redirected to active failure elimination. 23 recurring failure modes assigned to named RE owners with 60-day elimination targets. First 5 eliminated by Day 45 — all hydraulic contamination modes across the haul truck fleet, resolved through sealed desiccant breather installation. Recurring failure cost for these 5 modes: $4.2M/year. Eliminated.

Priority 3 — Reliability Review Redesign (Days 7–21): Meeting attendance reduced from 18 to 6. Agenda restructured: 10 minutes MTBF trend review (RE presents), 15 minutes top-3 recurring failure mode status (RE presents, decision required), 10 minutes CBM alert review (RE presents), 5 minutes planning horizon confirmation (Planner presents). Meeting duration: 40 minutes. Decision rate per meeting: 3–5 vs previously 0.

Priority 4 — Planning Discipline (Days 21–60): Planner removed from reactive dispatch. Job Plan Coverage Rate measured weekly from Day 21. JPCR at Day 21: 11%. At Day 60: 54%. At Day 90: 74%.

Results at 12 Months

The Financial Position — Maintenance Cost as an Existential Question

At mid-scale gold operations, the relationship between maintenance cost and profitability is brutally direct. A Tier-1 producer can absorb a 38% maintenance-to-EBITDA ratio — it is high, but the absolute EBITDA size provides margin for performance improvement. At EBITDA in the target range, the annual maintenance cost leaves $13.7M for corporate overhead, finance costs, exploration, and shareholder return. At a mid-scale operation with typical corporate overhead of $5–7M, the margin for maintenance cost inefficiency is approximately zero.

The Mine Director's question — structural or recoverable — is the correct question. Structural maintenance cost is the irreducible minimum for the equipment class, fleet size, and operating environment. Recoverable maintenance cost is the portion attributable to governance failure: reactive maintenance premium, recurring failures that should have been eliminated, emergency procurement at logistics premium, and over-servicing from undifferentiated PM strategy. The S1 structured assessment found that $5.4M of the the annual maintenance cost was recoverable — 64%.

S1 Findings — Where the $5.4M Was Hidden

The 12-Month Programme — S1 + S2 Priority Sequencing

Programme investment constraint: the Mine Director had a budget of $200,000 for the combined engagement. MitWin scoped S1 + S2 with a constrained S3 — prioritising the highest-value recoverable items within the investment constraint. Sequencing: S1 in Weeks 1–4 (audit and findings). S2 priority in Weeks 5–12 (strategy correction for the three highest-value failure mode corrections and criticality-adjusted interval redesign for auxiliary fleet). S3 governance installation in Months 3–6 (planning discipline, CBM workflow, monthly executive review). S3 was partially self-delivered by the Maintenance Manager — coached by MitWin remote advisory for 6 months post-engagement at $4,200/month.

Results at 12 Months

The Mechanisation Without a System Problem

Small-scale mining operations that transition from artisanal or semi-mechanised extraction to full mechanisation frequently make a consistent error: they invest in equipment without investing in the maintenance system that sustains it. The equipment is procured correctly. The operators are trained. The haul road is designed. The maintenance infrastructure — CMMS, PM schedule, CBM programme, planned maintenance workforce, spare parts holding — is deferred because it seems like it can be established after operations begin. It cannot. Within 6–12 months of mechanised operation without a maintenance system, the fleet begins deteriorating from deferred maintenance, and the cost of recovering it from degraded condition exceeds the cost of building the system correctly at mechanisation.

At this operation, the Managing Director's description of the maintenance situation was direct: "We change the oil when the machine stops smoking. We have no idea what condition the other components are in. Two machines have been unserviceable for 6 weeks and I cannot get a reliable assessment of whether they are worth repairing or if I should buy replacements." This is the state of a fleet maintained without a system. The MD was not asking a maintenance question — he was asking a capital question that could not be answered without a maintenance baseline.

What MitWin Found at Engagement

The S1 assessment adapted to the small-scale context: MitWin used a compressed 10-day format rather than the standard 15-day engagement, focusing on physical fleet condition assessment, equipment recovery prioritisation, and foundational system design rather than the full MRMM scoring applicable to larger operations.

The 90-Day Foundation Build

Days 1–14 — Fleet Recovery. Physical assessment of all 8 units. Excavator: hydraulic pump seal failure — recoverable at $18,400. Dump truck: engine seized from extended oil interval — beyond economic repair. Recommendation: source reman engine at $28,000 vs replacement unit at $114,000. Both units returned to service at Day 22. Total recovery cost: $46,400 — 51% below the MD's estimate from informal contractor quotes.

Days 7–30 — PM System Installation. PM schedule designed for all 8 units using OEM manuals calibrated to East African laterite and high-humidity operating environment. Service intervals for engine oil, filters, hydraulic fluids, and greasing routes specified. Oil analysis programme established with a regional laboratory in Nairobi — quarterly sampling for all major fluid systems. Pre-shift inspection checklist designed for operators — 12-point daily check, documented on paper form digitised weekly into a shared spreadsheet (appropriate for the operation's scale and IT infrastructure).

Days 21–60 — Spare Parts Foundation. Critical spares holding calculated for the 8-unit fleet: C1 items (failure produces immediate total fleet capacity loss) — stocked. C2 items (failure reduces capacity significantly) — min/max established with monthly replenishment cycle. Emergency procurement process defined: maximum 3 days to sourcing decision, approved supplier list for each equipment class. Pre-engagement emergency procurement rate: 88% of all parts purchases. At Day 90: 24%.

Days 45–90 — Workforce and Supervision. Two technicians hired at MitWin's recommended specification — both with CAT equipment experience and hands-on diagnostic capability (tested with structured assessment). Workshop Foreman role defined and filled internally from existing team — most experienced technician promoted with a 30-day coaching programme delivered by MitWin Reliability Lead on-site for 2 weeks and by video call for remaining 6 weeks.

Results at 12 Months

The Prior Project — A Reference for What Was Coming

The contractor's prior project history was the most useful diagnostic tool available before mobilisation. Four of the prior project's top-5 failure modes — engine air filter bypass in fine volcanic dust environment, hydraulic hose abrasion on highland switchback haul sections, motor grader circle drive contamination from road surface water infiltration, and dozer final drive contamination from stream crossing operations — were structurally predictable from the operating environment characteristics. None had been anticipated in the maintenance strategy. All four were also present in the current project's operating environment profile.

Pre-mobilisation analysis means projecting failure modes from environment characteristics before the first machine starts. It requires a framework — MitWin's Project Stress Classification — and reference data from comparable operations. It does not require equipment to be on site. The contractor who performs this analysis before mobilisation prevents failures that the contractor who does not perform it will pay for in the first 6 months of the project.

The Pre-Mobilisation S1 — 8 Weeks Before First Machine on Site

MitWin's pre-mobilisation engagement format: 2-day project environment assessment (desktop analysis of route geology, climate data, haul profile, stream crossings, altitude profile), 1-day prior project failure data analysis, 1-day current fleet condition assessment at staging yard, and 3-day strategy calibration and spares kit redesign. Total elapsed time: 8 working days over 3 weeks.

Environment Assessment Findings: Highland volcanic basalt — highly abrasive, angular particle profile. Altitude range 800m–2,200m — significant cooling system and engine performance implications. 34 stream crossings on the 280km route — final drive and seal exposure risk. Average ambient: 28°C at lower elevation, 14°C at upper — thermal cycling risk for hydraulic systems on machines moving between elevation zones. Annual rainfall 1,840mm concentrated in 5-month wet season — road surface water infiltration and mud contamination significant throughout programme.

Fleet Condition Assessment: 68 units assessed at staging yard over 2 days. 14 units identified with deferred maintenance items that, if mobilised without correction, would fail within the first 90 days of project operation. Total pre-mobilisation repair cost: $186,000. Total projected failure cost if mobilised with these defects: $1.1M (based on failure cost reference data from comparable prior project). The pre-mobilisation repair decision — investing $186,000 to prevent $1.1M — required one conversation with the Project Director.

Strategy Calibration for Highland Operating Environment

Mobilisation Spares Kit Redesign

The contractor's standard mobilisation kit — sized for a standard flatland project — was inadequate for 5 of the 8 equipment classes assessed. Air filter consumption at calibrated Highland intervals: 2.8× standard kit quantity. Final drive oil: 1.9× standard. Hydraulic seals: 2.4× standard (stream crossing exposure). Pre-mobilisation kit was redesigned and procured at standard pricing — 8 weeks before project commencement. Emergency procurement premium avoided across the 24-month project: estimated $420,000.

Results — Measured at Project Month 12

What Makes Nickel Laterite Different

Nickel laterite ore bodies present a maintenance environment that is fundamentally distinct from hard-rock mining in three material characteristics: the ore is a fine, highly plastic clay-rich laterite that penetrates every seal and joint in an equipment system with greater efficiency than any coarser material; it carries a significant chrome-nickel particulate in suspension that is far more abrasive per gram than standard silica; and the associated groundwater chemistry — typically pH 5.8–6.8 with elevated nickel and magnesium ion concentrations — is mildly but consistently corrosive to standard hydraulic and final drive seals over extended exposure periods.

Equipment operating in nickel laterite will develop hydraulic contamination, seal failures, and undercarriage wear at rates 30–60% higher than the same equipment in hard-rock operations with equivalent operating hours — if the maintenance strategy is not calibrated to these specific material characteristics. Most nickel laterite operations deploy an uncalibrated strategy — because the OEM manual does not include a "nickel laterite" adjustment and no regional maintenance reference exists outside of MitWin's engagement database.

Three Strategy Corrections That Changed Everything

Results at 12 Months

The Mid-Project Collapse — What Actually Happened

Fleet availability deterioration from 79% to 58% over 12 months on a live project does not happen suddenly. It happens through a compound process: the first recurring failure mode begins producing downtime at Month 3. The maintenance team responds reactively — parts replaced, machine returned. The failure recurs 6 weeks later. Reactive response again. By Month 6, three additional recurring failure modes have appeared. The maintenance team is now managing four active recurring patterns reactively — with no time for planned maintenance, no analytical resource to investigate root causes, and no planning horizon beyond the next breakdown.

By Month 12, the maintenance team is in full reactive mode: 81% of maintenance spend is reactive, scheduled PM compliance is at 34%, and the two most experienced technicians are permanently assigned to breakdown response. The planned maintenance programme has effectively ceased. Equipment condition deteriorates faster than reactive repair can address it. Availability collapses.

Four Recurring Failure Modes — Identified and Eliminated

Governance Restoration — 4 Weeks

With the four recurring modes eliminated, the maintenance team's reactive workload fell 64% in 30 days. This freed capacity for planned maintenance — PM compliance recovered from 34% to 72% in 6 weeks. MitWin installed a lightweight governance cadence appropriate for a project environment: daily 15-minute stand-up (Foreman + planner), weekly 40-minute reliability review (PM compliance, open failures, parts status), and a 2-week rolling schedule replacing the prior day-to-day reactive scheduling. The planning horizon that had collapsed to 24 hours recovered to 14 days.

Results at Project Completion

The Business Situation

In this illustrative scenario, a mid-scale coastal refinery operating a population of roughly 120 API 610 pumps in critical service was experiencing repeated unplanned trips traceable to rotating-equipment failure. Mechanical seal and bearing failures accounted for the majority of unplanned mechanical work orders. A small group of "bad-actor" pumps — fewer than ten units — was generating a disproportionate share of emergency intervention, and two of them sat on the feed circuit to a primary process unit, where an unplanned stop carried both margin loss and flaring exposure.

Maintenance was competent but reactive. Vibration data was collected on a manual route, yet seal-plan and alignment defects were rarely closed before the next failure. The reliability conversation lived in the workshop, not at the leadership table.

What the MitWin S1 Audit Found

The Fleet Stability & Cost Risk Audit (S1) classified the rotating-equipment population by criticality and built a bad-actor ranking from work-order and downtime history.

Mechanical seal failures dominated, and most were traceable to seal-support-system (flush plan) defects and residual misalignment rather than to the seal faces themselves. Vibration alarms were being recorded but not consistently converted into corrective work orders.

Concentration of Recurring Failure

The MitWin Intervention

The engagement combined an S1 diagnostic with an S2 Maintenance Strategy redesign focused on the critical rotating-equipment population.

Indicative Results

Figures are indicative of the scale of opportunity in a refinery rotating-equipment population of this size and should be validated against the specific asset base.

The Executive Lesson

The exposure was not spread evenly across 120 pumps — it was concentrated in fewer than ten, and in two repeating failure mechanisms. Governing the small, recurring population is where reliability value in rotating equipment is recovered. The seals were not the problem; the absence of a system to close seal-support and alignment defects was.

The Business Situation

In this illustrative scenario, a regional container terminal operating a fleet of ship-to-shore (STS) quay cranes and rubber-tyred gantry (RTG) yard cranes was missing berth windows because of unplanned crane downtime. In a terminal, crane availability is not a maintenance statistic — it is berth productivity, and berth productivity is the commercial product. Every hour an STS crane is down during a vessel call is an hour of berth capacity that cannot be re-sold.

The terminal tracked gross moves per hour (GMPH) closely but did not connect the productivity shortfall to a small set of recurring equipment failures in the crane drive and reeving systems.

What the MitWin S1 Audit Found

The audit linked crane availability and GMPH shortfall to equipment failure history, ranking failure modes by their impact on berth productivity rather than by repair cost alone.

Wire-rope and spreader-system defects, hoist-gearbox condition, and festoon / cable-management faults recurred — and condition data on the drive and gearbox systems was not being used to plan intervention ahead of failure.

Failure Concentration by Berth Impact

The MitWin Intervention

An S1 diagnostic was followed by an S2 strategy focused on the systems that interrupt berth productivity.

Indicative Results

Figures are indicative of the opportunity at a terminal of this scale and should be validated against vessel-call profile and crane configuration.

The Executive Lesson

Crane availability and berth productivity are the same number viewed from two seats. When reliability is governed against berth impact rather than repair cost, the maintenance function stops competing for budget and starts protecting revenue. The cranes did not need replacement — the failure concentration needed governing.

The Business Situation

In this illustrative scenario, a combined-cycle power facility was carrying an elevated equivalent forced-outage rate (EFOR) that exposed it to lost dispatch revenue and availability-linked capacity obligations. The forced outages were not random: a recurring set of balance-of-plant failures — boiler-feed pumps, fan bearings, and auxiliary valves — was responsible for a disproportionate share of unplanned megawatt-hours lost.

The plant maintained its major-equipment outage schedule well, but balance-of-plant reliability was managed reactively, and the link between those failures and the availability factor that drives revenue was not visible at the leadership level.

What the MitWin S1 Audit Found

The audit related forced-outage history to specific failure modes and quantified each in terms of lost availability rather than repair cost.

Boiler-feed-pump and fan-bearing condition, plus auxiliary-valve reliability, recurred across the forced-outage record — and condition data was being collected without a disciplined path into planned corrective work.

Forced-Outage Concentration

The MitWin Intervention

An S1 diagnostic informed an S2 strategy and a governance cadence centred on forced-outage reduction.

Indicative Results

Figures are indicative of the opportunity at a facility of this configuration and should be validated against the unit’s dispatch profile and contractual availability terms.

The Executive Lesson

Availability is the revenue product of a generator, and forced-outage exposure was concentrated in balance-of-plant equipment that sat below the leadership line of sight. Governing those recurring modes — and reporting EFOR as a leadership metric — converts reliability from a maintenance cost into protected dispatch revenue.

The $340,000 Surprise That Was Not a Surprise

A CAT 793F haul truck at a copper operation in southern Africa suffered a catastrophic torque converter failure at 38,640 hours. The repair: $340,000, 22 days off-fleet, $1.4M in production loss. When the structured diagnostic assessment reviewed the maintenance history, the torque converter had been replaced at 19,200 hours — its design life is 18,000–22,000 hours from new. At 38,640 total machine hours, it had been running for 19,440 hours since installation: at or beyond its second design life. Nobody was tracking it. Nobody was expecting it to fail. It was, from a component life perspective, a completely predictable event that appeared on the breakdown report as "unplanned."

What Component Life Tracking Covers

For heavy mobile equipment, the components that warrant individual life tracking — because their failure is catastrophic and their replacement is planned-replaceable — typically include:

How to Start — With Data You Already Have

Step 1: Pull every major WO from the last 36 months for each equipment class. Identify all component replacement events. Record the machine hour at installation. That is your starting inventory of current component life clocks.

Step 2: For each component currently installed, calculate: current machine hours minus installation hours = hours on component. Compare against OEM design life. Flag any component within 20% of design life as an upcoming planned replacement.

Step 3: Create a component life register in your CMMS — one row per component per unit, recording: component type, unit ID, installation date, installation hours, design life, current hours-on-component (auto-calculated from current machine hours), and % life consumed.

Step 4: Generate a 90-day replacement forecast from the register. Any component reaching 85% of design life in the next 90 days becomes a planned replacement job — budgeted, parts pre-ordered, equipment release window confirmed with operations.

The Financial Transformation

What Happens on Night Shift

At a West African gold operation — in this illustrative scenario — in 2024, a review of 18 months of breakdown data revealed a pattern that the Maintenance Manager had not noticed: 38% of all recurring failure callbacks occurred on components that had been repaired on night shift. Night shift accounted for 33% of total maintenance hours — but 38% of callbacks. The quality differential was not dramatic. But it was real, consistent, and financially consequential.

The root cause was not technician capability. The night shift team included some of the operation's most experienced fitters. The root cause was supervision structure. On night shift, the Workshop Coordinator — a Senior Fitter on a coordinator allowance — managed 11 technicians across surface and underground simultaneously. Post-repair verification happened when he had time. WO documentation was completed from memory at the end of the shift. Contamination control during hydraulic repairs was inconsistent because there was no one to enforce the standard.

The Three Night Shift Quality Gaps

Two Structural Interventions — No Additional Headcount Required

Intervention 1: Staggered shift overlap with day supervisor verification window. Start the day shift supervisor 2 hours before night shift ends. The supervisor's first 2 hours: walk every open job on night shift, verify any Class A repair completed in the prior 4 hours, sign off WO documentation before night shift closes it. This gives night shift post-repair verification coverage without a dedicated night supervisor — and creates a knowledge transfer moment that improves night shift quality over time.

Intervention 2: Closed WO gate — Class A jobs cannot close without supervisor sign-off in CMMS. Configure the CMMS so that WOs flagged Class A (engine, transmission, final drive, hydraulic major) cannot be closed without a supervisor sign-off field completed. On night shift, this means the WO stays open and is verified by the incoming day supervisor in the first 30 minutes of shift. The machine can be released operationally — but the WO remains open until verified. This one configuration change enforces quality verification without requiring a supervisor to be present at the moment of repair completion.

Why Contamination Causes More Failures Than Wear

Hydraulic components are designed to operate in clean fluid. The clearances between hydraulic pump pistons and barrel bores are typically 5–10 microns — smaller than a human red blood cell (approximately 8 microns). A particle of silica dust at 20 microns — invisible to the naked eye — is twice the size of the operating clearance. When it enters the fluid circuit, it does not pass through. It abrades. It scores. It generates secondary metallic particles. Those particles abrade further components. The contamination cascade accelerates component wear at 3–8× the design rate.

Across MitWin's fleet audit database, hydraulic contamination accounts for an average of 31% of all hydraulic system failures — more than normal wear, more than seal failure, more than overheating. In operations with no contamination control protocol, the proportion rises to 44–51%.

The Four Contamination Entry Points

The ISO 4406 Cleanliness Code — Reading the Number

The ISO 4406 cleanliness code is expressed as three numbers — e.g., 18/16/13. Each number represents a range of particle count per millilitre at three particle sizes (4 micron, 6 micron, and 14 micron). Higher numbers = more particles = dirtier fluid. The target code for most mining excavator hydraulic systems: 17/15/12. A result of 20/18/15 means the system is significantly more contaminated than target — and pump life is reducing at an accelerated rate.

When your oil analysis report shows ISO 4406 cleanliness code, compare it to your equipment-specific target. If the code is 3 or more numbers above target on any of the three digits — contamination investigation is required before the next service, not at it.

What Your Backlog Is Actually Telling You

Most Maintenance Managers know their backlog number — total outstanding work orders, or sometimes total outstanding hours. What most do not have is a backlog analysed by age, by equipment criticality, and by failure mode type. A backlog of 180 WOs is not a problem or not a problem by itself. The question is: what is in those 180 WOs?

The Healthy Backlog Target

Industry benchmark: a healthy maintenance backlog is 2–4 weeks of planned work for available crew capacity. Below 1 week: crew is idle — demand is insufficient for capacity. Above 6 weeks: capacity is insufficient for demand — or reactive work is crowding planned work out of the schedule. The target is a stable backlog in the 2–4 week zone, with no Class A equipment WOs older than 5 working days.

The One Intervention That Makes Everything Worse

When a Mine Director sees a 14-week backlog, the instinctive response is a "backlog blitz" — dedicated overtime, contractor augmentation, and a target to clear the backlog by a fixed date. This intervention clears the symptom and damages the system. It consumes maintenance budget on overtime premium. It interrupts the regular maintenance schedule. It produces a post-blitz period where the cleared backlog refills rapidly because the underlying capacity-demand imbalance was not addressed. The backlog blitz is the wrong response to a structural problem.

Three Interventions That Reduce Backlog Sustainably

Intervention 1: Backlog triage — classify and defer correctly. Not every WO in the backlog needs to be done immediately. A structured triage classifies each WO: safety-critical (must complete within 24 hours), Class A equipment (within 5 working days), Class B/C equipment (within scheduled window), non-critical improvement (defer to next planned opportunity). Triage removes the false urgency from the backlog and allows the planning team to sequence work by consequence — not by age.

Intervention 2: Demand reduction through recurring failure elimination. In operations where 30–40% of all WOs are reactive breakdown responses, recurring failure elimination directly reduces WO demand. Eliminate one recurring failure mode (8–12 events per year) and the backlog reduces by 8–12 WOs per year — permanently. This is the only backlog intervention that produces lasting reduction without adding resources.

Intervention 3: Planner capacity alignment. If the backlog is growing because planned work is being deferred by reactive demand, the root cause is planner overload and reactive intrusion into planning time. The fix is structural (planner-to-unit ratio) and process-based (reactive coordination separated from planning) — not overtime.

The Tyre Cost That Nobody Is Managing

At a mid-tier iron ore operation in West Africa, The S1 assessment found the following tyre situation: 12-unit haul truck fleet, average tyre life 3,840 hours against a benchmark of 5,600–6,400 hours for this equipment class and haul profile. The operation was replacing tyres 32–46% faster than benchmark. Annual tyre spend: USD 2.84M. At benchmark tyre life: USD 1.68M. Annual excess tyre cost: USD 1.16M — from a problem nobody had formally identified because tyre spend appeared as an invoice line, not a performance metric.

The root cause was not tyre quality or specification. The cause was four tyre management failures operating simultaneously: inflation management by feel (not by gauge), haul road surface inspection absent, tyre rotation not practiced, and failure mode recording at zero completeness. The operation had every tyre invoice. It had zero tyre performance data.

Five Tyre Management Disciplines

The Rearview Mirror Problem

MTBF is calculated from completed failure events — it is, by definition, a historical metric. A fleet whose MTBF declines from 180 hours to 140 hours between Q1 and Q2 has already experienced the failures that drove that decline. The response — more PM, more RCA, more CBM investment — will show up in MTBF improvement 60–90 days later, if it works. The 60–90 day lag between intervention and measured outcome means that a governance system based solely on MTBF is always reacting to yesterday's performance while today's failures accumulate.

Leading indicators break this lag. They measure the causal conditions that produce MTBF outcomes — before those conditions produce failures. A leading indicator that degrades today predicts MTBF degradation in 60–90 days. An intervention triggered by the leading indicator prevents the MTBF degradation from occurring at all.

The Four Leading Reliability Indicators

Leading Indicator 1 — CBM Alert Response Rate

What it measures: The percentage of condition monitoring threshold breaches (oil analysis, vibration, ultrasound) that generated a work order within 48 hours of the breach being identified. Target: above 95%. Why it leads MTBF: A CBM alert response rate of 70% means that 30% of all developing failures detected by the CBM programme are not being acted on in the 48-hour window. These unactioned alerts are developing toward functional failure — they will appear in the MTBF figure 4–12 weeks later. A declining CBM alert response rate today predicts declining MTBF in 6–10 weeks. How to measure: Count CBM alerts generated in the period ÷ WOs generated from CBM alerts within 48 hours = CBM Alert Response Rate. Available from CMMS if alert-to-WO workflow is configured. Takes 15 minutes per month to calculate.

Leading Indicator 2 — PM Compliance Rate by Failure-Critical Task

What it measures: The percentage of PM tasks classified as Failure-Critical (tasks directly linked to preventing a specific failure mode) that were completed within 10% of their scheduled interval. Target: above 95% for failure-critical tasks. Why it leads MTBF: PM compliance for failure-critical tasks is the most direct measure of whether the maintenance strategy is being executed as designed. A failure-critical PM task completed at 95% of compliance produces the failure prevention it was designed to deliver. At 70% compliance, 30% of the scheduled failure prevention is not occurring — which predicts an increase in those specific failure modes 1–3 months later. Important distinction: total PM compliance rate (typically 70–85% in most operations) is not the same as failure-critical PM compliance rate. Mixing administrative PMs (fluid checks, greasing, inspections) with failure-critical PMs (valve overhauls, seal replacements, bearing changes) dilutes the metric's predictive power.

Leading Indicator 3 — Active Recurring Failure Mode Count

What it measures: The number of failure modes that have occurred more than once in the trailing 90 days and have not had a formally documented corrective action implemented. Target: zero unresolved recurring failure modes older than 90 days. Why it leads MTBF: Recurring failure modes are the most predictable source of future MTBF degradation — because they have already demonstrated that they recur without intervention. An active recurring failure mode is a scheduled future breakdown with an unknown date. Counting them is the simplest leading indicator of future breakdown frequency. Governance application: The recurring failure mode register — a simple table of failure mode, first occurrence, recurrence count, and corrective action status — is the single most actionable document in the reliability governance system. It should be the first item reviewed at every Weekly Reliability Review.

Leading Indicator 4 — Deferred High-Consequence Work Order Age Profile

What it measures: The number of WOs classified as High-Consequence (failure mode that stops production or causes secondary damage if the defect is not corrected) that are more than 14 days past their scheduled date. Target: zero High-Consequence WOs more than 14 days overdue. Why it leads MTBF: A deferred High-Consequence WO is a developing failure that has been identified but not actioned. Every day past the scheduled date, the failure mode is accumulating deterioration. The deferred WO age profile is a direct leading indicator of forthcoming breakdown events — specifically the ones that would not have occurred if the WO had been executed on schedule. How to measure: Filter CMMS backlog by failure consequence classification and count by overdue duration band. Requires failure consequence classification to be configured in the CMMS — a one-time setup task.

Integrating Leading Indicators into the Governance Dashboard

The four leading indicators above, reported weekly alongside MTBF, transform the reliability governance dashboard from a historical review into a forward-looking management system. The dashboard format: MTBF by equipment class (lagging, 30-day trailing) | CBM Alert Response Rate (leading, 7-day trailing) | Failure-Critical PM Compliance Rate (leading, 7-day trailing) | Active Recurring Failure Mode Count (leading, 90-day trailing) | High-Consequence Deferred WO Count >14 days (leading, current).

When any leading indicator falls below target, the governance response is triggered before the MTBF decline that would otherwise follow. This is the governance model that closes the lag between cause and consequence in reliability management.

Why Surface-Calibrated Strategy Fails Underground

The majority of OEM maintenance manuals are calibrated for surface operating conditions. The majority of industry maintenance benchmarks are derived from surface fleet data. The majority of reliability engineering textbook examples use surface operating contexts. When a mining company develops its underground fleet maintenance strategy from these sources — without systematic environmental adjustment — it is applying a surface-calibrated strategy to an underground environment. The result is a set of predictable failures that the strategy was not designed to prevent, occurring on a schedule that the strategy was not designed to anticipate.

This is not a hypothetical risk. MitWin's underground engagement portfolio — spanning zinc-lead operations in West Africa, copper operations in Central Africa, gold operations in West Australia, and nickel operations in Canada — consistently identifies the same set of surface-to-underground strategy mismatches as the primary source of recurring failure modes at engagement. The mismatches are the same across geographies because the underground physics are the same: humidity, acid, confined space, steep grade, and darkness create a consistent set of maintenance challenges that surface operations do not face.

Eight Underground Environmental Differences and Their Strategy Implications

Difference 1 — Humidity (70–99% Relative Humidity Underground vs 20–60% Surface)

Grease degradation rate at 95% RH: 40–55% faster than at 40% RH. Electrical insulation brittleness: 3× higher failure rate at sustained 90%+ RH. Bearing grease film stability: requires EP (Extreme Pressure) additive package rated for high humidity — standard multi-purpose greases lose EP film at 90%+ RH within 80–100 hours. Strategy adjustment: High-humidity grease specification for all rotating equipment (NLGI Grade 1 or 2 with EP + anti-moisture package). Grease interval reduced by 35–45% from surface equivalent. Electrical connector inspection every 200 hours with anti-corrosion treatment.

Difference 2 — Acid Rock Drainage (ARD) — Sulphide Ore Environments

pH 2.8–4.2 in active underground sump water. OEM seal specifications rated for pH >4.5. ARD-exposed seals in hydraulic systems, final drives, and wheel motors fail at 2.4–3.8× surface rates. Frame steel corrosion at ARD contact points: 5–8× standard rate. Strategy adjustment: Viton or PTFE seal specification for all underground equipment in ARD environments. Frame anti-corrosion treatment programme at 500-hour intervals. Hydraulic sealed breather desiccant specification (not standard OEM vented breather). ARD pH monitoring at 4 underground locations monthly — pH above 4.0 triggers upgraded seal specification review.

Difference 3 — Grade Loading (12–20% Underground Ramps vs 0–8% Surface)

Torque converter thermal loading at 16% grade: 1.8–2.6× standard duty cycle. Final drive loading at 14%+ grade on fully-loaded haul truck: 2.1× standard. Retarder brake heat accumulation on descent: surface cooling systems insufficient for sustained 14%+ grade descent. Strategy adjustment: Torque converter oil change interval reduced by 40–50% for 14%+ grade operations. Torque converter temperature gauge monitoring — any reading above 105°C triggers load reduction protocol. Final drive oil sampling every 200 hours (vs 500 hours surface). Retarder fluid specification for sustained grade duty.

Difference 4 — Confined Space and Limited Equipment Access

Underground equipment access is constrained by drift dimensions, ventilation doors, and equipment density. A PM task that takes 2.5 hours on surface may take 4–6 hours underground due to access constraints, lighting requirements, and equipment positioning. Service vehicle access to specific underground levels may be limited by portal clearance. Strategy adjustment: Underground PM task time estimates increased by 1.6–2.2× surface equivalent. Mobile service capability (service truck or underground service vehicle) assessed and specified for each underground level. Underground PM windows planned around ventilation cycles and blasting schedules — not surface shift patterns.

Difference 5 — Single-Access Production Risk

In many underground operations, each active stope or development heading is accessed by a single equipment route. A machine breakdown blocking that route can stop production from that entire section for the duration of the repair — not just the downtime of the broken machine. MTTR in this context includes tow-out time, portal access time, and re-entry time for the production equipment behind the broken machine. Strategy adjustment: Critical-path underground equipment classified separately from non-critical-path equipment. Critical-path equipment minimum availability target: 92% vs 88% overall fleet target. Tow-out capable vehicle maintained on each underground level. Maximum MTTR target for critical-path breakdowns: 8 hours (vs 14-hour average surface MTTR).

Difference 6 — Ventilation and Air Quality

Underground diesel equipment operates in a ventilation-controlled environment with diesel particulate limits governed by occupational exposure standards. High engine load at high altitude or in poorly-ventilated sections produces diesel particulate concentrations that exceed permitted limits — requiring machine removal from service until ventilation is restored. Engine condition monitoring (oil analysis, particulate filter condition) has direct occupational health consequences in underground — not just equipment reliability consequences. Strategy adjustment: Engine oil analysis sampling monthly (not quarterly) for underground diesel equipment. Diesel particulate filter inspection at 300 hours. Engine tune checks every 500 hours for emissions compliance.

Difference 7 — Tyre and Ground Engaging Tool Failure in Sulphide Rock

Angular sulphide rock fragments (pyrite, chalcopyrite) in stope floors produce tyre sidewall punctures at 2.8–4.2× the rate of limestone, schist, or granite. Standard ply rating LHD tyres designed for smooth-floor applications fail catastrophically on sulphide stope floors. Bucket teeth wear rates in massive sulphide ore: 3.5× sandstone or limestone equivalent. Strategy adjustment: Underground LHD tyre specification: high-ply sidewall (minimum 20-ply) with cut-resistant compound. Tyre inspection after every shift on sulphide stopes. Bucket tooth replacement trigger: visual dull indicator (not weight or cycle count) in sulphide ore.

Difference 8 — Parts Accessibility and On-Site Storage

Underground workshops have constrained storage capacity, limited lifting height (ceiling height constraints), and restricted access for large components (final drives, engines). The logistics of transporting a major component from underground workshop to surface rebuild shop and back adds 1.5–3 days to major component replacement MTTR. Strategy adjustment: Underground on-site critical spare holding: minimum 1 unit of each C1 component that fits within underground workshop. Component change-out protocol: pre-agreed surface repair shop logistics (road route, crane spec, component transport equipment) for each major component class. Major component exchange-in-advance programme: rebuild component arrives before failure, not after.

The Procurement Logic That Creates the Diversity Tax

The procurement rationale for fleet diversity is competitive: "We always buy from multiple OEMs to maintain competitive tension in future procurement negotiations." This rationale is financially sound at the acquisition stage — competitive procurement consistently delivers 5–12% unit cost reductions versus single-source procurement. It is financially unsound over the asset's operating life, because it implicitly assumes that the diversity cost of the mixed fleet is zero. It is not zero. It is, in most mid-tier operations reliability assessments have found, $180,000–$420,000 per year on a 30-unit mixed fleet — and it compounds every year the mixed fleet configuration persists.

Five Components of the Fleet Diversity Tax

Component 1 — Parts Inventory Duplication

A standardised 20-unit Komatsu HD785 haul truck fleet requires one parts inventory for the hydraulic pump assembly — one part number, one safety stock calculation, one supplier relationship. A mixed fleet with 12 HD785s, 5 CAT 793Cs, and 3 Volvo A60Hs requires three independent hydraulic pump inventories, three supplier relationships, three safety stock calculations, and three emergency procurement routes. At a minimum safety stock of 2 units per model, the mixed fleet holds 6 units of hydraulic pump inventory. The standardised fleet holds 2 units, supplying twice the number of machines. Excess inventory investment: 3× the capital requirement. Annual excess holding cost (at 12% holding cost rate on $240,000 in excess inventory): $28,800 on this single component alone.

Component 2 — Technician Training and Competency Breadth

A technician qualified on Komatsu HD785 requires additional training to work on CAT 793C hydraulic systems — the hydraulic architecture, the diagnostic software, the fault code structure, and the service manual nomenclature are different. A mixed-OEM fleet requires technicians with broader competency profiles — which either means more training cost per technician or a larger workforce with narrower specialisation per person. At a 30-unit mixed fleet requiring competency across 4 OEM platforms: estimated additional training cost vs standardised fleet: $28,000–$46,000 per year in formal training, plus an estimated 340 additional diagnostic hours per year from technicians working outside their primary competency zone at $75/hr = $25,500.

Component 3 — Tooling and Diagnostic Equipment Duplication

OEM-specific diagnostic software, special service tools, and calibration equipment are not interchangeable between OEM brands. A workshop supporting 4 OEM brands requires 4 diagnostic software licences ($3,200–$6,400 per licence per year), 4 sets of OEM special service tools ($8,000–$22,000 per set), and in many cases 4 relationships with different OEM field service engineers. Mixed-fleet tooling cost premium vs standardised fleet: estimated $24,000–$48,000 per year on a 30-unit fleet.

Component 4 — Planning Complexity and Schedule Inefficiency

A maintenance planner scheduling a 30-unit standardised fleet works from one PM schedule structure, one set of failure mode parameters, and one parts catalogue. A mixed-fleet planner works from 4 different PM schedule structures, 4 different failure mode profiles, and 4 parts catalogues. The planning overhead of this complexity — measured in planner hours required per unit per month — is approximately 35–45% higher in a mixed fleet than in a standardised fleet. On a 30-unit fleet with a $95,000/year planner salary: excess planning cost from diversity: $33,000–$42,000 per year. Or: the planner serving 30 mixed-OEM units is performing the planning function of a 20-unit standardised fleet — effectively requiring 50% more planning resource for the same fleet size.

Component 5 — Telematics Fragmentation

Each OEM telematics platform captures fleet data in a proprietary format with a proprietary API. OEM telematics (Komatsu), OEM fleet monitoring systems (Caterpillar), mine optimisation platforms (Sandvik), and CareTrack (Volvo) do not natively integrate with each other or with most CMMS systems. A mixed fleet produces fragmented telematics data — requiring either manual cross-platform data compilation (estimated 6–10 hours per month of analyst time) or a third-party telematics aggregation platform ($18,000–$36,000 per year). The standardised fleet produces consolidated telematics data from one platform — enabling the cross-fleet analytics that are impossible across fragmented systems.

The Standardisation Strategy

Fleet standardisation is not achieved through a single procurement decision — it is achieved through a replacement strategy that progressively consolidates the fleet toward 1–2 primary OEM brands over 3–5 fleet replacement cycles. The strategy requires: (1) defining the target fleet configuration (which OEM brands, which models, maximum number of primary brands); (2) defining the replacement priority sequence (which non-standard units are replaced first); and (3) incorporating standardisation as a formal criterion in every future procurement decision — not just acquisition cost.

The financial case for standardisation: on a 30-unit mixed fleet with 4 OEM brands, the annual diversity tax (sum of 5 components above) is estimated at $180,000–$340,000. Over a 5-year fleet replacement cycle, this represents $900,000–$1.7M in avoidable cost — before accounting for the analytical benefits of consolidated telematics data that a standardised fleet makes possible.

Why Technically Correct Training Does Not Change Behaviour

The standard mining maintenance training programme has three characteristics that guarantee it will not change technician behaviour. First: it is classroom-based — the training occurs in a room, not on the machine. Second: it is generic — the content covers the equipment model category, not the specific machines at this site. Third: it is a one-time event — the technician attends, receives a certificate, and returns to the workshop, where the operating environment immediately reasserts the previous behaviour pattern.

None of these criticisms address the technical content. The OEM course on hydraulic system maintenance for a Sandvik LH514E is technically accurate, well-produced, and relevant to the technician's work. The problem is not the content. The problem is that the training delivery model does not create the conditions in which new behaviour replaces old behaviour. Knowledge transfer and behaviour change are not the same process — and most training programmes are designed for the former while expecting the latter.

Three Training Design Failures

Failure 1 — Transfer Without Application

Classroom training delivers knowledge. Knowledge is retained in the working environment only when it is immediately applied — ideally within 48–72 hours of the training session. Most mining training programmes deliver classroom content on one day and return technicians to the workshop the next day, where the specific machine and the specific task from the training are not guaranteed to be waiting for them. The knowledge that was demonstrated in the training room is not reinforced by application in the work environment — and begins to decay within days. Design response: All training content is scheduled to precede a specific planned PM task on the exact machine type covered in the training, within 48 hours of the classroom session. The supervisor confirms that the trained technician performed the specific task within the reinforcement window. This converts classroom knowledge into applied work behaviour.

Failure 2 — Assessment Without Consequence

End-of-course assessments in most mining training programmes are pass/fail written tests. Passing the test earns the certificate. The certificate has no consequence for work assignment — technicians who failed the practical aspect of the work are returned to the same tasks as technicians who excelled. Without performance-linked work assignment, the assessment creates no incentive to master the skill beyond the written test standard. Design response: Competency-based work assignment — technicians are assigned to specific task categories based on their demonstrated practical competency level, not their training certificate status. A technician at Level 2 hydraulic competency works on Level 2 hydraulic tasks under Level 3 supervision. Advancement to Level 3 is earned by demonstrated performance on Level 2 tasks, assessed by the supervisor quarterly, not by course attendance.

Failure 3 — Training Without Reinforcement Structure

The maintenance operating environment is the most powerful training force available — more powerful than any classroom programme. If the operating environment rewards speed over quality (because breakdown metrics dominate the supervisor's performance incentive), technicians learn speed. If the operating environment rewards WO volume over WO accuracy (because WO count is the productivity metric), technicians learn volume. Training programmes that exist in isolation from the operating environment's incentive structure will always lose the behaviour battle to the environment. Design response: The training programme is designed as part of the operating system — not alongside it. Supervisor performance metrics include: technician competency score improvement (from quarterly RECI assessment), WO documentation quality (from the weekly audit), and post-repair callback rate (from 30-day followup). When supervisor performance is measured in part by workforce capability development, the supervisor becomes the training programme's primary delivery mechanism.

The Reliability Academy Architecture

MitWin's Reliability Academy model integrates classroom training, on-machine application, competency-based assessment, and work assignment into a single learning architecture designed to close the knowledge-to-behaviour gap. The architecture has five layers: (1) Technical knowledge — OEM content adapted for site-specific equipment and conditions; (2) Practical application — within 48 hours, on the specific task, with supervisor observation; (3) Competency assessment — quarterly, hands-on, skills-matrix scored, not written test; (4) Work assignment calibration — task responsibility matched to demonstrated competency level; (5) Reinforcement environment — supervisor performance metrics that create organisational incentive for capability development.

At a 44-unit underground mining operation in West Africa where MitWin implemented the Reliability Academy model alongside S3, RECI score improved from 18/50 to 34/50 over 12 months — compared to a prior year where $280,000 in OEM training spend produced RECI improvement of 3 points. The architecture change produced a substantial improvement in training ROI without changing the OEM training content.

The Multi-Site Governance Failure Modes

In MitWin's multi-site reliability governance engagements, two failure modes appear with near-equal frequency:

Failure Mode A — Group Uniformity. The group mining executive imposes a single maintenance standard, a single PM interval library, and a single CMMS configuration across all sites. Sites in tropical laterite environments apply the same air filter interval as sites in temperate hard-rock. Sites with ARD underground environments use the same seal specifications as sites with neutral-pH groundwater. The governance is consistent. The failure rates are predictably inconsistent — in the direction the physics requires.

Failure Mode B — Site Autonomy Without Comparability. Each site manages its own maintenance strategy independently. The group CFO receives a maintenance cost per hour figure from each site but cannot determine whether Site C's $184/hr CPH reflects a worse operating environment, a worse maintenance programme, or a different cost accounting methodology. Group-level reliability intelligence is absent because the site-level data is not comparable. Capital allocation decisions between sites are made without the analytical basis that comparable reliability data would provide.

The Three-Layer Architecture

Layer 1 — Group Intelligence Framework (Standardised)

The group intelligence framework is the set of analytical standards, metrics definitions, and reporting formats that every site uses identically — regardless of operating environment, equipment fleet, or commodity. This layer covers: MRMM scoring methodology (same 10 domains, same 80-point scale, same domain definitions at every site), CREM calculation methodology (same 6 components, same production value calculation approach, same financial attribution rules), FSM classification (same 5-zone maturity scale, same scoring thresholds), and KPI definitions (MTBF calculated the same way at every site — same failure event definition, same scheduled hours denominator, same interval period). The group intelligence framework produces comparable data across sites — the prerequisite for group-level capital allocation and performance benchmarking.

Layer 2 — Site Maintenance Strategy (Calibrated)

The site maintenance strategy is where environmental calibration occurs. Each site develops its own PM task library, CBM programme, interval specifications, and equipment-specific standards — calibrated to the site's specific operating environment (dust load, humidity, grade profile, groundwater pH, ambient temperature). The site maintenance strategy is not standardised across the group — it is calibrated. Two sites with the same Komatsu HD785 fleet will have different air filter intervals if one operates in the operation site laterite and the other in Atacama desert. Both intervals are correct for their environments. Neither is the OEM default. The calibration methodology is standardised (same 8-factor environmental assessment process) even though the outputs differ.

Layer 3 — Group Governance Cadence (Structured)

The group governance cadence is the meeting structure and reporting rhythm that converts site-level intelligence into group-level decisions. It has three components: Monthly Site Performance Pack (4-page standardised report from each site, produced by the site RE and reviewed by the site Mine Director — covers MRMM score, FSM score, CREM, and top-3 reliability risks); Quarterly Group Reliability Review (90-minute meeting with group mining executive, all site Mine Directors, and group asset management lead — benchmarks site performance, identifies cross-site best practice, allocates group reliability improvement budget); and Annual Group MRMM Assessment (independent MRMM assessment of each site by MitWin, producing comparable maturity scores that inform the group capital allocation process).

The Five Group-Level Reliability Metrics

The Cross-Site Learning Dividend

A multi-site operation that implements the three-layer architecture has one capability that a single-site operation can never have: the ability to observe the same failure mode managed differently across multiple sites and compare outcomes. When Site A and Site C both experience hydraulic pump contamination failures in a laterite environment, but Site A has a 28% recurrence rate and Site C has a 4% recurrence rate for the same failure mode — the group governance architecture allows the group mining executive to identify that Site C's sealed breather kit installation programme is the differentiating variable, and to mandate its implementation at Site A. This is the cross-site learning dividend — and it is only available to groups that have comparable reliability data across their site portfolio.

The Problem a $220M Fleet Should Not Have

The operation was not under-resourced. The maintenance team comprised 84 technicians, 6 planners, 2 supervisors per shift, and a Reliability Engineer with 9 years of mining experience. The CMMS was fully implemented. PM schedule compliance averaged 91%. The Weekly Reliability Review had been running for 18 months. And yet: the large mining shovel fleet was experiencing swing bearing replacement at an average of 8,200 hours — against an expected life of 14,000 hours. Three swing bearing replacements in 14 months at $840,000 per replacement = $2.52M. Two additional undercarriage assembly replacements on the CAT 6060 class: $680,000. Total premature wear failure cost: $3.2M in 14 months.

The RE had identified the pattern. The swing bearings were failing from lubricant starvation — not from contamination, not from overloading, not from operating conditions. The grease was not reaching the bearing. The PM schedule showed greasing completed on time, every time. The investigation revealed why these two statements were simultaneously true.

What the Investigation Found

MitWin S1 field investigation (Days 5–8): physical inspection of grease nipple access on the CAT 6060 swing bearing circuit. Finding: 4 of 12 grease nipples on the swing bearing distribution circuit were blocked — grease nipple heads filled with hardened grease/ore fines mixture that had set to concrete consistency. Grease applied at the fitting head was returning pressure to the gun — indicating "complete" to the technician — but was not reaching the bearing. The PM record: "swing bearing lubrication — 100 pump strokes applied — complete." The bearing: receiving approximately 20% of the required grease volume on each lubrication cycle. The failure: developing over 6 months of technically-compliant PM that was delivering inadequate lubrication to the most expensive component on the machine.

The second finding: 2 of 6 automatic lubrication system (ALS) distribution blocks on the Komatsu PC8000 had failed — one with a blocked distribution line, one with a failed check valve. The ALS fault light had been active for 12 weeks. No WO had been generated. The Reliability Engineer had never been notified. The bearing circuit was on manual lubrication bypass for 12 weeks — a temporary measure that had become permanent through administrative invisibility.

The Intervention

Action 1 (Week 1): All 6 large mining shovels removed from service for a 4-hour lubrication circuit inspection. All grease nipples cleaned, tested with pressure gauge, and confirmed clear. Blocked nipples replaced. ALS distribution blocks inspected and repaired. 14 specific lube points found inadequate — all corrected before return to service. This single action cost 24 workshop hours and restored lubrication to standards that had not existed for an estimated 8–14 months.

Action 2 (Week 2–3): Lubrication verification protocol designed by the RE: each large shovel has a laminated lube circuit diagram in the cab. Supervisor physically confirms clear grease purge from 6 primary lube points on each shovel every 500 hours — not every PM cycle, but every 500 hours specifically. Purge confirmation: visible fresh grease emerging from the purge point adjacent to the bearing. If purge not visible: circuit blocked — immediate investigation WO raised. This protocol takes 18 minutes per shovel per 500-hour cycle.

Action 3 (Week 3–4): ALS fault alert integrated into CMMS: any ALS fault code generates a WO automatically (same day) with "do not defer" flag. ALS fault logs reviewed weekly by RE. Three ALS faults identified and resolved in the first month of the new protocol — none of which would previously have generated a WO before the failure of the component being lubricated.

Results at 12 Months

Operating at 4,200 Metres — the Maintenance Environment Nobody Planned For

The operation's fleet was specified for standard altitude operating conditions. At 4,200m above sea level, atmospheric pressure is 60% of sea-level pressure — reducing available oxygen for combustion, reducing engine cooling airflow, and reducing turbocharger efficiency. The practical consequences: engine power output at 4,200m is 15–22% below sea-level rating, fuel consumption is 18–25% higher per unit of work produced, and engine thermal loading per operating hour is significantly higher due to reduced cooling air density. None of the PM intervals in the CMMS had been adjusted for altitude operating conditions. The engine oil change interval (OEM: 500 hours at sea level) had not been adjusted for the increased thermal loading at altitude. The air filter interval (OEM: 500 hours) had not been adjusted for the reduced filter efficiency at low atmospheric pressure combined with high-altitude dust characteristics (fine volcanic tuff).

The S1 Audit Findings

The CREM: production loss from 69.4% availability (target 88%): $3.8M. Recurring failure cost: $2.6M. Reactive maintenance premium: $1.2M. Emergency procurement premium (380km from supply, high international parts costs): $0.8M. Total: the total annual maintenance cost.

Six of the eight recurring failure modes were directly attributable to altitude interval mismatch: engine overheating (oil change interval too long at altitude), air filter bypass (filter interval too long at altitude), turbocharger wastegate failure (thermal stress from altitude derating), hydraulic oil viscosity failure (standard viscosity grade insufficient at 4,200m temperature differential between night −4°C and day 18°C), fuel injection fouling (incomplete combustion at altitude creates injector deposit faster than sea-level interval accounts for), and cooling system scaling (water quality at altitude combined with incorrect flush interval).

The Intervention — 14 Months

S2 — Altitude-Calibrated Strategy (Months 1–4): All PM intervals recalibrated for 4,200m operating conditions using MitWin's altitude adjustment methodology: engine oil 500 hrs → 320 hrs. Air filter 500 hrs → 220 hrs. Turbocharger inspection 1,000 hrs → 600 hrs. Hydraulic oil: upgraded to multi-grade specification covering −4°C to 18°C operating range. Fuel injector testing 2,000 hrs → 1,200 hrs. Cooling system flush 2,000 hrs → 1,100 hrs. Interval adjustments implemented in CMMS Week 4. First full altitude-calibrated service cycle completed Month 2.

S3 — Transformation (Months 3–14): Planner activated — removed from reactive coordination, 4-week rolling schedule built by Month 4. Planning resource challenge: 28 units across surface and underground with one planner at 380km from supply. Solution: 8-week forward parts order horizon (vs standard 4-week) to account for supply chain lead time. All planned PM parts ordered 8 weeks in advance. C1 critical spares held at site: 18 items identified and pre-positioned. RE coaching: existing RE (process plant background, 2 years mining) coached on altitude-specific failure mode analysis, oil analysis threshold adjustment for altitude viscosity conditions, and FMEA methodology applied to underground high-altitude equipment. First RCA Month 3. First independent RCA Month 5.

MRMM tracked monthly: 31 → 38 → 44 → 49 → 54 → 58 at Month 14.

Results at Month 14

The Zero-Base Starting Point

MitWin's engagement approach for operations transitioning from artisanal or semi-mechanised to structured mechanised production. These operations typically have: no maintenance records (failure history is oral — "that machine broke twice last rainy season"), no maintenance budget line (repairs are paid from operational cash when they occur), no qualified maintenance personnel above Workshop Foreman level (mechanics with trade experience but no mining maintenance training), and no reliability awareness at the management level. The Managing Director of this operation, when asked for his fleet availability figure, responded: "I don't know what that means — but I know we lose machines for weeks at a time and it costs us."

The engagement was scoped for this context: no assumption of CMMS, no assumption of existing data, no assumption of management familiarity with reliability engineering concepts. The output: a working maintenance system that a 4-person team could operate with minimal external support after 90 days.

The 90-Day Programme — Simplified for Scale

Week 1–2 — Fleet Assessment and Priority Classification: Physical inspection of all 14 units. Fleet age, condition, and critical defect identification. Two machines found with active structural issues requiring immediate attention before further operation (one dozer with cab riser cracks, one dump truck with brake hydraulic leak). Both removed from service and repaired before the programme formally began. This immediate action — before any governance system was in place — prevented two incidents that would have been far more expensive than the repairs.

Week 3–4 — Simple PM Schedule Design: OEM manuals translated into a simple weekly/monthly PM table for each machine — in English and the local language (French). No CMMS required: a laminated poster for each machine in the workshop with service intervals, service items, and a checklist for the mechanic to initial on completion. The poster was the PM system. It was free to print, visible to every team member, and required no computer access or training to use.

Week 5–8 — Workshop Organisation and Parts Minimum Stock: Workshop organised by machine class — tools, parts, and service materials sorted and labelled. Top-10 failure-mode spare parts identified from OEM failure guides and from Workshop Foreman's oral history. Minimum stock of 1 unit each established for these 10 items. Supplier relationship established with the nearest town parts distributor (180km away) for emergency procurement. Parts total investment: $12,400.

Week 6–10 — Shift Handover and Defect Reporting: One-page shift handover form introduced — defects, fluid levels, warning lights. Collected by the Workshop Foreman at each shift end. One machine defect identified per shift on average. Three developing failures detected in the first 4 weeks from handover observations alone — all resolved before breakdown. Estimated breakdown cost prevented in first month: $84,000.

Week 8–12 — Weekly Review Meeting: 30-minute Monday morning meeting — Workshop Foreman, Operations Manager, Managing Director. Agenda: what broke last week, what is scheduled this week, what parts are needed. The MD attended every week. By Week 10, the MD was asking "why did it break?" rather than just "what broke?" — the beginning of reliability thinking at the management level.

Results at Month 12

The Same Machine, Three Different Failure Rates

Motor Grader GR01 (Caterpillar 16M3) on Project 1 (flat coastal alluvial, 24km road, sea-level altitude): 4 breakdown events in 8 months. Motor Grader GR01 on Project 3 (rolling laterite terrain, 38km road, 1,200m altitude): 6 breakdown events in 7 months. Motor Grader GR02 (same model, same age) on Project 5 (volcanic highland, 28km mountain road, 2,400m altitude, 18% grades, basalt subgrade): 14 breakdown events in 5 months.

The project management response to Project 5: "It's a hard project — the conditions are extreme." The maintenance management response: "GR02 isn't well-suited to this type of terrain." MitWin's analysis: the failure rate difference between Project 1 and Project 5 was entirely attributable to five specific maintenance strategy mismatches that were predictable from the project specification and preventable with a 4-hour pre-mobilisation review. None of the 5 strategy mismatches had been identified before GR02 was mobilised to Project 5. All 5 were eliminated within 6 weeks of MitWin engagement — and Project 5's failure rate fell from 14 events per 5 months to 4 events per 5 months in the period after intervention.

The Five Project 5 Strategy Mismatches

Mismatch 1 — Grade-adjusted engine oil interval. Project 5's 18% grades produce sustained uphill engine load at 88–94% of rated power. OEM oil change interval (500 hours) designed for mixed-duty at standard grade. Altitude + grade combined thermal load: oil degradation rate 1.8× standard. Required interval: 270 hours. Actual interval applied: 500 hours. Engine oil degradation contributed to 3 of 14 breakdown events.

Mismatch 2 — Blade edge specification for basalt subgrade. Standard blade edge (grader blade cutting edge) designed for laterite and sedimentary soil. Basalt subgrade wear rate on standard blade: 4.2× standard. Replaced twice on Project 1 per full project. Replaced 9 times on Project 5 (same project duration) at $2,800 per replacement = $25,200 in blade consumption on Project 5 vs $5,600 on Project 1. Carbide-insert blade edges were available and appropriate for basalt — not specified at mobilisation.

Mismatch 3 — Transmission oil and interval for sustained grade operation. Sustained grade-loading (18%) on the motor grader transmission produces thermal loading that the standard oil viscosity cannot sustain continuously. Transmission oil temperature gauge reading on Project 5: 108–114°C sustained. OEM operating limit: 110°C. Required action: upgraded to synthetic transmission oil rated to 140°C continuous. Applied at Week 6 after MitWin engagement. No transmission failures after Week 6. Before Week 6: 2 transmission failures at $38,000 combined.

Mismatch 4 — Tyre specification for volcanic basalt haul road. Standard tyre (bias ply, standard compound) on volcanic basalt haul road surface: sidewall puncture rate 4.8× laterite road. 11 tyre replacements on Project 5 in 5 months at $1,200 each = $13,200. Cut-resistant compound with reinforced sidewall specified at Week 7. Zero sidewall punctures in subsequent 8 weeks.

Mismatch 5 — Pre-mobilisation spare positioning. Project 5 was 340km from the nearest parts distributor. Emergency procurement lead time: 4–7 days for most items. 9 of 14 breakdown events involved parts waiting periods extending MTTR by 3–7 days. Required: pre-positioned critical spare kit at project site. Implemented at Week 5 (19 items, total value $28,400). Zero parts-wait MTTR extensions in subsequent 12 weeks.

The 12-Item Pre-Mobilisation Checklist

MitWin designed a 12-item pre-mobilisation reliability checklist for the contractor — applied to every new project before the first machine arrives on site. The checklist takes 4–6 hours to complete for a typical 15-machine project. It covers: grade profile and adjusted engine intervals, subgrade characterisation and blade/bucket/tyre specification adjustment, altitude operating parameter check, supply chain lead time and critical spare pre-positioning requirement, operator competency assessment for project-specific conditions, and fuel and fluid quality verification for the project location. The checklist is completed by the Project Maintenance Supervisor and reviewed by the Group Reliability Manager before mobilisation approval is granted.

Results — 2 Project Cycles Post-Implementation

Four Years of Subscribed, Unread Intelligence

The Maintenance Manager confirmed that OEM telematics was accessed for one purpose: generating the fleet hour report that triggered PM scheduling. He was unaware that OEM telematics contained fault code history, payload data, idle time ratio, and engine load factor trend — all of which had been generating continuously for 4 years. When MitWin opened the OEM telematics Machine Event History for HD325-11 (one of the haul trucks flagged by the S1 audit as a high-risk unit from physical inspection), the system showed: 34 J01CB fault code events (engine coolant temperature high) in the preceding 12 months, including 8 events in the preceding 4 weeks. The machine had been generating a recurring overheating signal for a year. No work order had been generated from any of these events.

The Six Developing Failures in the Data

The Analytics Programme Installed

MitWin designed a 4-metric weekly OEM telematics analytics protocol for the operation's Maintenance Planner: (1) Engine load factor deviation — flag any unit showing >8% above class average on same route; (2) Idle time ratio — flag any unit above 38%; (3) Fault code recurrence — flag any fault code appearing 3+ times in 30 days; (4) Payload histogram — review weekly for overload events above 105% of rated payload. Total weekly analytics time: 3.5 hours for 24 Komatsu units. No additional software required. OEM telematics standard subscription. Results reviewed at weekly reliability meeting.

By Month 6 post-implementation: 14 developing failures identified and prevented from OEM telematics data. Estimated combined prevented failure cost (based on failure mode and progression at time of detection): $2.4M. Monthly analytics cost: approximately $1,200 in planner time. Programme return at Month 6: compelling relative to programme investment.