Buy a tube laser or outsource tube cutting: how to calculate the decision point ART-190

Set the baseline unit of comparison

A kilogram or machine-hour is not universal. It is more useful to calculate the cost of an accepted set of parts or assembly delivered by the required date. The set includes different profiles, nesting, marking, sorting, inspection and the risk that one missing item stops assembly.

Create a landed accepted set cost for each family: material, processing, logistics, packaging, inspection, rework, scrap, administration and inventory cost. For the internal route add capital, fixed and variable cost, and cost of capacity. For the external route add price, minimum lot, expediting, delivery, incoming inspection and deviation cost.

Clean the historical data

Subcontractor invoices may include material or only processing; delivery and urgency can be separate. Internal estimates may omit engineering time, setup or depreciation. Put data on the same boundary before analysis.

Collect 12–24 months of part number, profile, material, wall, bar length, quantity, due date, order date, promised date, actual receipt, price, delivery, rejects, rework and emergency premium. Separate prototypes, series work and one-off projects. Do not model a total purchasing sum without its mix.

The subcontractor model

The external route has explicit and hidden costs. Explicit costs are quotation, material markup, cutting, programming, setup, packaging, delivery and minimum lot. Hidden costs are RFQ time, revision alignment, file control, queue waiting, incoming inspection, claims, repeat delivery and inventory needed because of lead time.

Benefits matter too: no asset to buy, specialists to retain, system to repair or idle capacity to fund. A subcontractor may have a wider machine envelope, a 3D head or automation you rarely need. Do not ignore that merely because the unit price looks high.

The in-house machine model

Capital cost is more than machine price. Add delivery, customs and financing under actual terms, foundation or floor, electrical supply, gases, compressor, extraction, cooling, network, rigging, installation, training, measuring equipment, a consumable-spares reserve and commissioning. Consider loader, unloading, software modules, licences and integrated mechanical options separately.

Fixed cost includes depreciation or lease, insurance, space, the base team, software support, preventive maintenance and a share of infrastructure. Variable cost includes material, gas, electricity, consumables, direct labour, scrap, maintenance by use and internal logistics. Do not confuse cash flow with accounting cost; show both.

Utilisation is a distribution, not one number

Annual volume is uneven. Build a weekly or monthly demand profile: median, peak, low season, urgent-order share and forecast error. A machine that looks loaded on an annual total may be idle for weeks and unable to meet peaks.

Calculate available productive hours separately: calendar time minus shifts, planned maintenance, changeovers, meetings, training, breakdown allowance and unprepared jobs. Theoretical cutting speed is not productive output. Programming, loading, piercing, support movement, sorting and short stops are part of the cycle.

Material and nesting

BLM GROUP highlights material’s role in work-order cost and different nesting configurations. The model needs real bars, lengths, clamp zones, cut gaps, tails and remnant-reuse rules. Two routes with the same cutting rate can differ through material yield.

A subcontractor may combine orders from clients and use the profile better, or add markup and not return a remnant. An in-house shop can optimise the mix but holds more SKUs and working capital. Model both without assuming in-house nesting wins automatically.

Lead time as an economic quantity

A shorter internal cycle can reduce WIP, inventory, tied-up capital and losses from drawing changes. It can also make a replacement or prototype sooner, but only when jobs are prepared, material is in stock, and machine and operator are available.

Estimate the cost of waiting: assembly stoppage over one item, delayed shipment, premium freight, overtime, penalty or lost order. Since not every delay has a direct penalty, use a range and probability. That makes it a scenario rather than invented precision.

Quality and cost of poor quality

For both routes calculate accepted-first-pass percentage, rework, scrap, investigation time and impact on the set. An internal machine gives faster feedback but can have a launch learning curve. An experienced subcontractor can be steadier, but claims and logistics take longer.

Run a comparison series of identical parts externally and on the proposed system. Measure not only hole and length, but frame assembly, marking, dross, cleaning and repeatability. Hidden manual finishing often changes the make-or-buy conclusion.

People and organisational readiness

An in-house tube laser needs a process owner, programmer or process engineer, operators, maintenance, quality and material handling. Roles may be combined, but ownership must be explicit. Add recruitment, training, shift cover, holidays, retention and time to stable productivity.

A subcontractor also needs a capable interface: correct models, revision control, purchase terms, acceptance criteria and supplier development. Outsourcing moves part of execution outside; it does not remove technical responsibility.

Downtime risk and a backup route

For an in-house system, model planned and unplanned downtime, service response, critical spares and an alternative supplier. If all geometry is consolidated on one machine, an outage can block several former operations at once.

Subcontractor risks differ: queue, price change, refusal of a small lot, geographic logistics, confidentiality and dependence on its equipment. A hybrid may cost more in an ideal month but less once risk tails are included.

Three demand scenarios

Build at least LOW, BASE and HIGH. In LOW reduce volume and price pressure; in BASE use confirmed backlog and cautious forecast; in HIGH add genuinely supported growth and test capacity. For each calculate cash, P&L, utilisation, headcount, inventory and outsourced overflow.

Add stress scenarios: delayed launch, weaker material yield, higher service cost, loss of the largest product, an interest-rate change and a breakdown in a peak week. If the investment works only in an optimistic forecast without margin, the decision point is unconfirmed.

A break-even formula without self-deception

At a basic level, annual break-even volume is fixed incremental cost divided by the difference between landed outsourced unit cost and internal variable accepted-unit cost. The denominator must be positive and the unit stable. For varied mix, use weighted contribution by family rather than an average kilogram.

Apply the capacity limit and timing afterward. Volume above break-even does not help if it arrives in peaks the machine cannot process. Also test NPV or equivalent annual cost over your horizon and discount policy. Payback is not a guarantee; it is the result of assumptions.

The hybrid model

Keep frequent families, engineering changes, prototypes and lead-time-critical jobs inside when appropriate. Send peak volume, rare sections, extra-long or heavy tubes, special 3D capability and downtime backup outside. The portfolio, not the prestige of ownership, defines the boundary.

Qualify the subcontractor in advance and keep data packages current. An emergency-only external route may fail when needed. The backup should pass at least periodic orders and quality review.

Data to request from the supplier

Request cycle-time calculation on your files, configuration list, consumables assumptions, energy and gas boundary, maintenance plan, availability basis, software licences, staffing concept, loader/unloader behaviour and acceptance test. TRUMPF describes software tools for processing-time and cost calculation, but a vendor estimate needs comparison with a live cutting trial.

Obtain raw assumptions rather than only a final number. Is loading included? Is the tail included? What bar length is used? Is sorting included? What operator action and changeover are assumed? The answers let you transfer the calculation to your model.

A staged investment decision

First normalise data and segment part families. Then conduct a market test with two subcontractors and a technology test on a candidate machine. Next build a scenario model, sensitivity analysis and risk register. Only then form an investment gate.

An intermediate step may be a framework agreement with a subcontractor, part design for tube-laser processing and gradual accumulation of stable volume. This gives actual data and lowers the risk of buying for an unready product mix. Another option is investment in software or design competence before equipment.

The decision-point criterion

Buying is justified when the confirmed portfolio gives sufficient contribution after all costs, capacity fits demand’s time profile, the organisation is ready for launch, and the downside remains acceptable. Lead-time, IP or product-development value must be described and tested, not just called strategic.

Outsourcing is justified when volume is unstable, broad capability is needed, the external market is competitive, and internal fixed cost and risk exceed the controllable contract price. A hybrid is justified when portfolio segments have different economics. The decision point is a refreshable transparent set of conditions, not a magic tonnage.

Segment the portfolio by reason for demand

Separate repeat series, engineering changes, prototypes, spare parts, peak overflow and new-business forecast. Every segment has a different value of speed and uncertainty. Repeat series supports a fixed asset; a one-off project should not automatically underpin a ten-year investment.

Add customer concentration. If most load depends on one contract, model its end or reduction. Assess separately whether free capacity can serve external orders: sales pipeline, quoting, quality system and payment terms must exist, not be optimistic revenue.

Account for ramp-up

The first months rarely equal steady production. Site preparation, commissioning, training, program transfer, process qualification and material-library build-up are required. Separate pre-production, ramp-up and steady state in the cash-flow model, with utilisation, scrap and external fallback for each.

Do not remove the subcontractor immediately at SAT. Maintain overlap until defined families pass repeatable acceptance and the team can recover from standard alarms. Overlap costs money, but reduces the risk of stopping core production.

Check working capital

An in-house machine can reduce finished tube-part stock through shorter replenishment, but needs raw-profile, gas, consumable and critical-spare stock. A subcontractor may hold material or require a full lot. Compare inventory days, payment terms and cash-conversion cycle.

Do not credit complete WIP reduction and full utilisation at once without proof: high utilisation often creates a queue. Define an operating policy for inventory and service level, then calculate working-capital effect. This matters especially for expensive special profiles.

The price of a design change

With outsourcing, a revision passes through RFQ, confirmation, queue and logistics. Internally it may be faster, but needs CAD/CAM competence, approval and controlled write-off of obsolete WIP. Calculate historical change frequency and parts lost to late revision.

If rapid iteration creates a product advantage, that is a real value of internal process. Assess it through a shorter prototype cycle, less obsolete stock or earlier revenue; do not replace measurement with the word “flexibility”.

A two-sided price check

Request at least two comparable external quotations for one data package and one internal estimate with raw assumptions for the candidate machine. Align material, lead time, quantity, quality, packaging and delivery. Where only one quotation includes special handling, the difference is not a clean process price.

After a trial order, replace quoted values with actual amount paid, quantity received, lead time, incoming rejects and administrative hours. For the internal test, use measured cycle time and material balance. The decision point should move gradually from estimate to evidence.

Financial-model governance

Assign an owner, version and date to every key input. Preserve its source: invoice, quotation, production log, maintenance plan or measurement. Separate confirmed, estimated and sensitivity assumptions. Protect formulas from accidental change and check units.

Update after every material fact: new price, mix change, test, service condition or financing rate. The investment committee should see not only base IRR or payback, but drivers, downside and the backup decision. The model becomes a managed tool, not a one-off spreadsheet to justify purchase.

Before final gate, appoint an independent reviewer to check formulas, units, duplicate savings and cost completeness. Labour reduction is often counted twice, or high utilisation and no queue are assumed together. Record observations and the model version used for the decision.

Limits of application

Sources support only stated functions of specific vendor ecosystems and the role of material and route. They do not establish price, utilisation, savings or payback for a future L-SEL investment.

Safety limits

This is a decision framework, not a financial, tax or investment guarantee. Input prices, rates, taxes, exchange rate, financing, technical package and risk assumptions must be checked at the decision date.

Need service advice?