Maintenance budgets should describe the future, not just the past

This piece builds on production consumes asset life. If production drives future maintenance demand, the maintenance budget is where that demand becomes a financial forecast.

Maintenance budgets are often built from history.

Last year’s expenditure is reviewed.

Known major work is added.

Labour rates, parts prices and inflation are adjusted.

A contingency may be included.

The result becomes next year’s maintenance budget.

That approach is understandable.

Historical expenditure is measurable. It provides a useful baseline. It gives finance a familiar structure and allows year-on-year comparison.

But maintenance demand does not arise from accounting history.

It arises from what the assets are expected to do next.

That suggests a useful proposition:

A maintenance budget should not only explain what was spent before. It should describe what the asset base is expected to require in the future.

Historical expenditure is useful, but incomplete

Suppose maintenance expenditure for the past three years has been:

FY24    $8.1m
FY25    $8.5m
FY26    $8.9m

A conventional budget might start with:

FY26 actual                    $8.9m
Inflation / escalation          $0.4m
Known additional work           $0.3m
-------------------------------------
FY27 budget                     $9.6m

That may be entirely reasonable.

But what if, during FY27:

production is expected to rise by 15%;
eight major components are approaching intervention;
a large shutdown is planned;
several ageing assets are being retained another year;
workshop labour is already close to capacity;
critical spares have longer procurement lead times;
condition monitoring has identified emerging risks?

The historical trend may not tell us enough.

The future asset position may be materially different from the past.

Maintenance expenditure is created by asset behaviour

Maintenance cost is not generated directly by the budget.

It is generated by work.

Work is generated by things such as:

Usage
Condition
Failure
Inspection
Component life
Production demand
Compliance
Shutdown scope
Asset age
Operating environment

Those drivers create:

Maintenance Requirement
        ↓
Labour
        ↓
Parts
        ↓
Contractors
        ↓
Tools / Plant
        ↓
Downtime
        ↓
Cost

The budget is therefore the financial expression of expected asset behaviour.

If the expected asset behaviour changes, the budget should change with it.

Budget and forecast are not the same thing

This distinction is important.

A budget is usually an approved financial position.

A forecast is an evolving view of what is now expected to happen.

At the beginning of a year they may be similar.

As actual operating evidence accumulates, they should diverge when circumstances change.

A useful structure might distinguish:

Budget
Plan
Committed
Actual
Forecast Remaining
Expected Outturn

For example:

Maintenance Budget               $10.0m

Actual to Date                     $5.8m
Committed                          $2.6m
Forecast Remaining                 $2.4m
                                  ------
Expected Outturn                  $10.8m

A simple budget-versus-actual report might show:

Budget              $10.0m
Actual                $5.8m

Remaining             $4.2m

That looks comfortable.

Operationally, however, the organisation may already be heading toward an $800,000 overrun.

The difference is the future.

Commitments matter

Actual expenditure alone is a lagging measure.

Maintenance operations create commitments before invoices appear.

These may include:

purchase orders;
component rebuilds;
contractor packages;
scheduled shutdown work;
labour commitments;
exchange components;
long-lead materials.

A useful operational budget view should therefore include committed work.

For example:

Actual                         $5.8m
Open Purchase Orders           $1.4m
Contractor Commitments         $0.7m
Component Rebuilds             $0.5m
                              ------
Committed Position             $8.4m

The system can then ask:

What additional work is still expected but not yet committed?

That produces a much more useful forecast.

Future work should also be visible

Some future maintenance demand is already known even though no purchase order or work order has been created.

Examples include:

scheduled services;
major component replacements;
statutory inspections;
shutdown scope;
tyre replacement;
planned rehabilitation;
structural intervention;
seasonal maintenance programs.

These are future liabilities in an operational sense.

They may not yet be financial commitments.

But they are real enough to influence planning.

The maintenance position might therefore look like:

Actual
   +
Committed
   +
Known Planned Work
   +
Forecast Work
   +
Risk Allowance
   =
Expected Outturn

That is much closer to the way operations need to think.

Production forecasts should influence the budget

If production consumes asset life, then changing production changes expected maintenance demand.

Suppose the original budget assumes:

10 Mt production
68,000 truck hours
6 engine interventions
4 transmission interventions

The production plan is later revised to:

11.5 Mt production
78,000 truck hours
9 engine interventions
6 transmission interventions

The maintenance budget should not remain static simply because it has already been approved.

The operational forecast should change.

The chain is:

Production Forecast
        ↓
Asset Utilisation
        ↓
Life Consumption
        ↓
Maintenance Demand
        ↓
Resource Demand
        ↓
Forecast Cost

Finance may continue to retain the approved budget.

Operations should retain the updated forecast.

Both are useful.

They answer different questions.

Component life can create predictable expenditure

Major components make this particularly visible.

Suppose a fleet contains engines with estimated remaining lives of:

Truck 12      1,200 h
Truck 18      1,450 h
Truck 27      1,900 h
Truck 31      2,100 h
Truck 36      2,250 h

The production forecast suggests most will reach intervention during the next financial year.

If the estimated cost per intervention is known, the future maintenance requirement is already becoming visible.

This is different from historical budgeting.

The system is effectively saying:

Based on current asset position and expected usage, these costs are approaching.

That is useful information even if the exact intervention dates remain uncertain.

Condition monitoring should modify the forecast

Life forecasts are not fixed.

Condition monitoring may bring work forward or allow it to move later.

Suppose the budget assumes a crusher bearing replacement in twelve months.

New vibration data suggests deterioration is accelerating.

The expected intervention moves into the current financial year.

That should change:

Maintenance Forecast
Parts Requirement
Shutdown Scope
Labour Demand
Cash Requirement

The approved budget does not need to be rewritten every time new evidence appears.

But the expected operational outturn should.

Otherwise the organisation may maintain a financially stable budget while its actual asset requirements move elsewhere.

Shutdowns create their own budget dynamics

Shutdown budgets are another clear example.

At the beginning of planning:

Shutdown Budget       $4.0m

As scope matures:

Baseline Scope         $3.5m
Approved Contingency   $0.5m

Then readiness develops:

Committed              $3.3m
Forecast Remaining     $0.9m
Expected Outturn       $4.2m

During execution, break-in work appears:

Break-in Work          $0.4m
Revised Outturn        $4.6m

That progression is more meaningful than simply:

Budget      $4.0m
Actual      $2.7m

The organisation needs to know where the shutdown is heading, not just how much has already been invoiced.

Resource constraints can also change cost

Maintenance demand does not translate into cost in a fixed way.

The same work may cost differently depending on resource capacity.

Suppose forecast maintenance demand is:

Required fitter hours       35,000 h
Internal capacity           30,000 h

The deficit may require:

5,000 contractor hours

At a higher hourly rate.

So the cost forecast should reflect not merely:

Work × Standard Rate

but also the likely method of delivery.

The same applies to:

outsourced rebuilds;
overtime;
additional shifts;
crane hire;
specialist contractors;
temporary workshop facilities.

The resource forecast and budget forecast are therefore linked.

Parts pricing and lead times also affect the position

A maintenance forecast may identify a component requirement eighteen months ahead.

The financial consequence depends on:

current price;
expected escalation;
exchange versus new;
freight;
currency;
lead time;
supplier availability.

Long-lead items introduce another important distinction.

The maintenance intervention might occur in FY28.

The purchase commitment may need to occur in FY27.

So:

Maintenance Timing
≠
Procurement Timing
≠
Cash Timing

That is important for both maintenance and finance.

A forward-looking asset model can make those timing differences visible.

Operational budgets should work in quantities as well as dollars

One of the weaknesses of purely financial budgeting is that a dollar figure can hide the physical assumptions behind it.

Suppose the annual tyre budget is:

$2.4m

That number becomes much more useful when connected to:

Expected tyres required       96
Average forecast cost         $25,000

Likewise:

Engine rebuild budget        $1.8m
Expected rebuilds                 6

or:

Road maintenance budget      $3.2m
Expected grading            4,800 km
Expected gravel            18,000 t

This allows the system to distinguish:

Did the budget vary because we performed more work?

from:

Did the same work cost more?

Those are very different management questions.

Variance should therefore have causes

A useful budget variance model should be able to explain:

Volume Variance
Price Variance
Timing Variance
Scope Variance
Productivity Variance
Failure Variance
Production Variance

Suppose maintenance is $1 million over budget.

The explanation might be:

Additional production demand       +$300k
Earlier component failures         +$250k
Parts price escalation             +$180k
Contractor labour deficit          +$220k
Deferred planned work              -$150k
                                  -------
Net variance                       +$800k

That is much more useful than:

Maintenance is 8% over budget.

The variance becomes operational evidence.

This can improve future budgets

Actual performance should feed back into the next planning cycle.

If a standard engine rebuild was budgeted at:

420 labour hours

but recent actuals consistently show:

510 labour hours

future forecasts should change.

Likewise:

Expected component life     12,000 h
Actual median life           9,800 h

should affect future component budgets.

The loop becomes:

Budget Assumption
      ↓
Actual Work
      ↓
Variance
      ↓
Learning
      ↓
Revised Assumption
      ↓
Future Forecast

The budget process begins to learn from operations.

The same model applies beyond mobile equipment

For roads:

Condition Forecast
      ↓
Treatment Quantities
      ↓
Plant + Material Demand
      ↓
Budget

For structures:

Inspection / Risk
      ↓
Expected Intervention
      ↓
Engineering + Construction
      ↓
Budget

For processing plants:

Production Duty
      ↓
Wear / Condition
      ↓
Shutdown Scope
      ↓
Labour + Materials
      ↓
Budget

For facilities:

Service Demand
      +
Compliance Requirements
      ↓
Maintenance Program
      ↓
Budget

The financial structures can remain common.

The operational drivers are different.

This suggests an operational position rather than just a budget report

A useful CPM view might be:

MAINTENANCE POSITION

Budget                    $10.0m
Actual                     $5.8m
Committed                  $2.6m
Known Planned              $1.3m
Forecast Reactive          $1.1m
Risk Allowance             $0.4m

Expected Outturn          $11.2m

Variance                  +$1.2m

Then allow the user to drill into the causes.

For example:

Expected Overrun

Component interventions      $450k
Production increase          $320k
Contractor capacity          $210k
Parts escalation             $140k
Other                         $80k

That is closer to an operational control system than a simple accounting report.

It also points toward scenario budgeting

Once asset behaviour drives forecasts, management can compare alternatives.

Base production

Production              10.0 Mt
Maintenance             $9.8m
Availability               87%

Higher production

Production              11.5 Mt
Maintenance            $11.4m
Availability               84%

Higher production plus additional workshop capacity

Production              11.5 Mt
Maintenance            $11.0m
Availability               87%
Additional capacity      $0.6m

The maintenance budget is now part of a management decision rather than simply a spending limit.

Financial systems should remain financially authoritative

None of this means the maintenance system should become the general ledger.

The ERP or accounting system should remain authoritative for:

invoices;
financial postings;
accruals;
tax;
statutory reporting;
financial periods.

The operational system has a different role.

It should explain:

What maintenance demand is emerging?

What work is already committed?

What resources will be needed?

What is the likely financial consequence?

Why is the forecast changing?

That distinction is important.

The operational forecast can be richer without trying to replace finance.

What this could mean for CPM and DSLCore

CPM could treat budget, commitment, forecast and actual as related but distinct states.

For example:

Maintenance Requirement
        ↓
Estimate
        ↓
Budget / Plan
        ↓
Commitment
        ↓
Actual

Alongside:

Asset Behaviour
        ↓
Updated Forecast
        ↓
Expected Outturn

That allows the system to show both:

What was approved?

and:

What is now expected?

The same mechanism could apply to:

dollars;
labour hours;
contractor hours;
parts;
component interventions;
downtime;
production.

That creates a consistent operational planning model.

A working hypothesis

The proposition is not that historical budgeting is wrong.

History remains one of the most useful sources of evidence.

The question is whether history should remain the primary description of future maintenance demand.

A stronger approach may combine:

History
+
Current Asset Position
+
Condition
+
Production Forecast
+
Lifecycle Forecast
+
Known Work
+
Resource Constraints

to produce:

Expected Maintenance Outturn

The budget then becomes one important reference point within a continuously evolving operational forecast.

A useful distinction may be:

The budget tells us what we planned to spend.

The operational forecast should tell us what the assets are now expected to require.

Both matter.

Confusing them can hide risk.

Questions worth testing

How much of a maintenance budget should be derived from asset and component forecasts rather than historical expenditure?

Should production plans automatically create revised maintenance forecasts?

How should organisations represent future work that is probable but not yet committed?

Should operational forecasts include quantities and resource demand as well as dollars?

How should condition-monitoring evidence change the financial forecast?

Can budget variance be classified into operational causes rather than simply financial accounts?

Should maintenance forecasts be continuously recalculated as work, condition and production change?

And perhaps the broader question:

If the asset base is continuously changing, why should our best view of future maintenance remain fixed at the point when the annual budget was approved?

A useful maintenance budget should help govern expenditure.

A useful maintenance operating model should also explain where that expenditure is heading, why it is changing, and what the organisation can still do about it.