Validating a CAPEX budget: from the balance sheet to the factory

CAPEX (capital expenditure, capital expenditures) in industry: stop buying machines “by reflex”, decide with simulation

Work by the Organisation for Economic Co-operation and Development (OECD) shows that, with comparable equipment, productivity gaps between factories can exceed 2 to 3x.

In other words: it's not the machine that drives performance, it's how the flow is managed.

This finding forces a distinction between physical capacity and flow performance: an investment does not mechanically create throughput. Too often, the first response to a drop in rate is to buy, when the constraint is elsewhere.

Key takeaway: a profitable investment survives a flow stress test, not a budget meeting.

I. Defining CAPEX: what finance measures… and what the shop floor absorbs

CAPEX (capital expenditure, capital expenditures): an operational definition you can copy-paste

Capital expenditures correspond to amounts committed to acquire, create, or durably improve a fixed asset, with an expected benefit over several years. In accounting, they feed into fixed assets and are then spread through depreciation. On the shop floor, they translate into new capacity, a different constraint, or added complexity. To be defensible, each CAPEX must tie to a measurable industrial KPI (key performance indicator).

Capital expenditure vs operating expenditure: where the line is, in practice

Operating expenditures (OPEX, operating expenditures) cover day-to-day operations and are expensed in the period. Capital expenditures are capitalized when useful life and materiality justify it. The decision is based on length of use, asset identifiability, and the capitalization policy. In practice, the distinction is detailed in the notes to the financial statements and the company's internal policy set by the CFO function (Chief Financial Officer).

Maintenance investment vs growth investment: two logics, two KPIs

A maintenance CAPEX targets continuity and is evaluated through availability and reliability. A growth CAPEX targets deliverable throughput and is evaluated through OEE (overall equipment effectiveness), output capacity, and lead time. Mixing the two creates incoherent decisions. Each project must clearly state its target KPI.

Three industrial examples

Example 1: a machining center purchased to gain precision and rate on a family of parts.

Example 2: a full assembly line including conveyors, stations, and checks to stabilize throughput.

Example 3: a building and utilities (industrial utilities) such as compressed air or industrial cooling to support a load increase. In each case, the asset is useless if flows don't follow.

II. Where CAPEX shows up and how to calculate it: two verifiable methods

“Cash” view: find CAPEX in the cash flow statement

The most direct view identifies cash outflows in the investing cash flows section. This line generally includes acquisitions of tangible and intangible fixed assets, net of disposals. It speaks quickly to the CFO function because it shows what left the bank account. This approach still requires an operational analysis of the expected gain.

“Balance sheet” view: link fixed assets, depreciation, and net change

The balance sheet view starts from net fixed assets after depreciation. Reconstructing CAPEX is done by taking the change in net fixed assets and adding depreciation expense and adjustments. Disposals or reclassifications require documented corrections. This method helps when the cash flow statement lacks granularity.

Formula: change in net fixed assets + depreciation expense ± adjustments

The usual formula estimates gross acquisitions from the balance sheet. It requires correctly identifying disposals, reclassifications, and investment subsidies to avoid sign errors. A common trap: a major disposal can hide a high CAPEX if you don't add it back properly. This method helps finance and the shop floor align, provided assumptions are documented and reconciled with the cash flow statement.

Mini numerical example

Net fixed assets end of N‑1: €50M. Net fixed assets end of N: €52M. Depreciation expense in N: €6M. With no disposal, estimated investment is 52 − 50 + 6 = €8M. If a net disposal of €2M occurred, estimated gross investment becomes €10M.

This financial view is necessary.

But it hides a core point: CAPEX is justified in the accounts… but validated in the real flow.

III. The “technology band-aid” trap: when a purchase moves the bottleneck… and inflates WIP

Bottleneck: the resource that dictates throughput

The bottleneck is the resource that limits overall throughput for a given product mix. It can be a machine, a manual station, an inspection step, or internal logistics. A machine datasheet is not enough to justify CAPEX. Measuring constrained throughput must come before any purchase decision.

WIP (Work In Progress, work-in-progress): the enemy of lead time

WIP represents parts in process between two steps. High WIP hides problems and increases lead time as well as working capital requirement (WCR).

Concretely, a 20% increase in WIP can:

• increase working capital requirement by several million euros on a site

• degrade customer lead time by several weeks

With no throughput gain.

It also increases quality risk through handling and prolonged storage. Reducing WIP lightens the system and makes problems visible.

Why adding capacity can make things worse

Increasing capacity on a non-constrained operation produces surplus that accumulates as intermediate inventory. The bottleneck remains unchanged, so overall flow does not improve. Congestion and variability increase, according to principles from queueing theory, explained by Britannica. Buying without a model turns a symptom into a cost.

IV. Flow simulation: uncover hidden performance before committing capital

What simulation settles

In practice, a CAPEX budget is often arbitrated based on:

• implicit assumptions

• theoretical averages

• or expert intuition

Simulation changes the rules: it replaces assumptions with tested scenarios, integrating breakdowns, variability, and scheduling rules. It's no longer a debate. It's a demonstration.

It distinguishes a capacity bottleneck from an organizational bottleneck, for example a release rule or poor synchronization. It also shows the side effects of a local improvement. Quantification reduces debate and speeds up decisions.

Buffers (buffers, buffer stocks) and priority rules

A well-sized buffer protects the constraint without creating an ocean of WIP. A clear priority rule reduces unnecessary changeovers and stabilizes queues. These levers require little investment but a solid understanding of variability. Simulation provides numbers and curves, not intuition.

Typical case: +10% productivity through line balancing

What: an assembly line is short on rate; an automated station is in the budget. How: simulation tests three release rules, a short buffer before inspection, and grouping product families. Impact: +10% throughput with the same headcount and reduced WIP, because the constraint stays better fed. The model incorporates breakdowns, setup times, and product mix to stay robust.

Minimum data to model

Cycle times, breakdown distributions, scrap rate, product mix, and calendars are often enough for a first useful model. You also need to formalize scheduling rules, because they generate queues and priorities. The team then improves model fidelity iteratively. Short data collection followed by shop-floor validation reduces error risk.

V. Stress-testing an investment project: prove assumptions wrong before they get expensive

Product mix

Product mix often changes faster than depreciation schedules. Simulation tests profitability when volumes shift toward longer or more demanding variants. It identifies thresholds beyond which the asset becomes useful. This approach avoids depreciating equipment that only serves for one year.

Infrastructure

A machine depends on conveyors, energy, and layout. The model forces checking electrical distribution, compressed air, and industrial cooling for the target rate. Without this, the asset risks being limited by its environment. Simulation exposes these hidden constraints.

Ramp-up (ramp-up, production ramp-up)

Ramp-up includes operator learning, setups, and non-conformities. Simulation integrates progressive maturity to avoid an overly clean business plan. It allows testing the introduction of new equipment alongside existing lines. The financial timeline becomes realistic.

VI. The ROI of avoidance: the best margin is the one you don't spend

Rates and financing

When rates rise, every euro not spent avoids financing cost and immediately improves the CFO function's view. Reducing CAPEX frees up debt capacity for differentiating projects. Avoiding investment also reduces future maintenance commitments and operating complexity.

TCO (total cost of ownership, total cost of ownership)

TCO adds purchase price and all life-cycle costs: maintenance, energy, spare parts, skills, and obsolescence. Cheaper equipment can cost more over ten years if consumption or breakdowns are high. The buying criterion becomes life-cycle spending, not the initial quote. Simulation feeds the TCO calculation with realistic scenarios.

Technological frugality

Fewer assets mean fewer breakdowns, fewer spare parts to manage, and reduced variability. A frugal factory puts every asset in service of a measurable KPI. Operational control becomes more readable for teams. Simulation puts a price on useless complexity.

In many industrial cases, flow optimizations enable:

• +10% to +30% throughput

• without heavy investment

In other words: the best CAPEX is clearly the one you avoid.

VII. Investment budget and project portfolio: a factory-oriented arbitration grid

Scoring criteria

Ranking projects by measurable impact requires evaluating capacity on the constraint, OEE (overall equipment effectiveness) at the bottleneck, quality in avoided scrap, energy in specific consumption, and time-to-commissioning in weeks of disruption. These criteria must be expressed in comparable units. The final score weighs operational stakes and projected financial benefits. Simulation provides the system-level impact estimate needed for ranking.

Risk criteria

You must score potential production stoppage, supplier dependency, maintainability, and internal skill level. Each risk becomes a penalty in the final score. A high-impact, high-risk project can still be prioritized if phasing reduces risk. The decision must remain owned and traceable.

Choosing between two competing projects and phasing

Comparing contribution to constrained throughput and then penalizing by a risk coefficient simplifies arbitration. If impacts are close, prioritizing shorter commissioning time reduces exposure to temporary losses. Phasing actions stabilizes existing operations first, modernizes next, and increases capacity after proof. This limits costly big-bang moves and losses during transition.

VIII. Mini industrial cases

Case 1: capacity increase on the real bottleneck

• What: a rail site targets +20% annual volume.

• How: simulation compares a second test bench versus one bench plus a new release rule and a short buffer.

• Impact: the optimized scenario reaches +18% throughput at 35% lower cost than the “two benches” scenario. The critical assumption is reducing micro-stops.

Case 2: modernization through retrofit (retrofit)

• What: a machining workshop suffers from quality drift.

• How: a control system retrofit, in-line measurement, and standardized setups validated by simulation.

• Impact: bottleneck OEE +6 points and scrap rate −20% without heavy flow changes. The project includes maintenance training to sustain the effect.

Case 3: digital investment (digital twin)

• What: a decision between two layouts and three ramp-up scenarios.

• How: a digital twin compares throughput, WIP, lead time, and sensitivity to product mix.

• Impact: decision made in weeks instead of months and avoidance of oversizing ancillary equipment. Initial data quality remains the critical assumption.

IX. Traps to avoid before budget approval

Underestimating commissioning costs

Commissioning consumes time, skills, and lost production. Hidden costs include rework, tuning, and IT or utility interfaces. A credible budget quantifies these losses and builds schedule buffers. Otherwise, an investment that looks profitable in Excel becomes an operational drift.

Forgetting the ramp-up (production ramp-up)

Ramp-up requires a stabilization period with throughput below nominal for some time. Omitting this phase pushes teams to compensate with WIP and overtime. The financial result may look good, but durability erodes. Simulation avoids this illusion by integrating learning effects.

Optimizing CAPEX at the expense of OPEX and TCO

A cheaper purchase can increase OPEX (operating expenditures) through higher energy consumption or frequent breakdowns. It can also create supplier dependency and rare consumables. Ten-year cost matters more than purchase price. TCO must frame the decision.

Neglecting flow integration and obsolescence

Adding a machine changes layout, travel paths, and intermediate inventories. If flow integration is weak, the bottleneck shifts to internal logistics. Obsolescence and maintainability complicate the asset's real service life. Accounting for these elements avoids costly surprises.

FAQ: CAPEX, calculation, and industrial arbitration

• What exactly is CAPEX?

CAPEX groups spending committed to acquire or improve a durable asset, capitalized as a fixed asset and then depreciated. In a factory, it must translate into a measurable gain in capacity, quality, energy, safety, or lead time. Without a KPI, it remains an accounting line. Simulation links the investment to expected gains.

• Where do you see CAPEX in the financial statements?

You find it in the cash flow statement, investing section, and on the balance sheet via fixed assets. It appears in the income statement as depreciation. The cash view is the most direct for the CFO function. The balance-sheet reconstruction complements the analysis if needed.

• How do you calculate CAPEX?

Method 1: read the cash flow statement, acquisitions of fixed assets. Method 2: reconstruct from the balance sheet: change in net fixed assets + depreciation expense ± adjustments. Both methods must reconcile with documented disposals and reclassifications. Financial consistency supports industrial arbitration.

• Can you reduce CAPEX without slowing growth?

Yes, if growth comes from better throughput at the constraint, lower WIP, and better-controlled variability. Simulation reveals gains via line balancing, priority rules, and right-sized buffers. The factory improves efficiency before buying. The approach preserves investment capacity for projects that truly differentiate.

Dillygence helps industrial companies compare scenarios, phase decisions, and measure the real impact of investments using a flow-oriented digital twin and shop-floor performance.