13.1 Separating variable and fixed costs
Variable costs arise with every unit: material, direct labor, energy, consumable tooling. Fixed costs occur whether production runs or not: depreciation, interest, space costs, salaries. The machine hour rate contains both — which is why it has to be split for contribution margin accounting.
| Component | €/year | Character | Reasoning |
|---|---|---|---|
| Imputed depreciation | 51,462 | fixed | time depreciation, runs independent of use |
| Imputed interest | 18,398 | fixed | capital commitment |
| Energy | 30,435 | variable | arises only during machining |
| Maintenance | 22,000 | 50 / 50 | maintenance intervals partly time-, partly runtime-dependent |
| Consumable tooling | 18,500 | variable | wear per cut |
| Space costs | 16,530 | fixed | floor area independent of utilization |
| Σ variable | 59,935 | = €17.66/MH | |
| Σ fixed | 97,390 | = €28.70/MH | |
| Σ MHR | 157,325 | = €46.37/MH | |
13.2 CM I, CM II and result
The reference quantity is cost center 210 at full operation: 3,393 productive machine hours, loaded with orders of the titanium-housing type (0.34 h occupancy per unit). That gives about 9,980 units per year — a model calculation that exhausts the capacity exactly.
| Item | €/unit | % revenue |
|---|---|---|
| Net revenue (cash price) | 182.77 | 100.0 % |
| − MEK Ti6Al4V | 85.00 | 46.5 % |
| − MGK (9.0 %, material-related) | 7.65 | 4.2 % |
| − DL production | 9.60 | 5.3 % |
| − DL CAM | 1.15 | 0.6 % |
| − MHR variable (€17.66/MH × 0.34 h) | 6.01 | 3.3 % |
| − SEKF special mill | 4.20 | 2.3 % |
| − SEKF fixture | 4.80 | 2.6 % |
| − SEKV free delivery | 3.50 | 1.9 % |
| = Σ variable costs | 121.91 | 66.7 % |
| = Contribution margin I | 60.86 | 33.3 % |
Fixed cost per unit: RGK 3.77 + MHR fixed 9.76 + VwGK 15.57 + VtGK 7.92 = €37.02. Variable costs 121.91 + fixed costs 37.02 = €158.93 — exactly the total cost from Chapter 11. And 182.77 − 158.93 = €23.84 profit. Both calculations describe the same part.
| Item | €/year | % revenue |
|---|---|---|
| Net revenue (9,980 units × €182.77) | 1,824,045 | 100.0 % |
| − Variable costs (9,980 × €121.91) | 1,216,662 | 66.7 % |
| = Total CM I | 607,383 | 33.3 % |
| − Fixed costs of the cost center itself (fixed MHR 97,390 + RGK 91,218) | 188,608 | 10.3 % |
| = CM II | 418,775 | 23.0 % |
| − Allocated administration and sales (9,980 × €23.49) | 234,430 | 12.9 % |
| = Result | 184,345 | 10.1 % |
Full costing reports €23.84 profit per unit, i.e. €237,923 at 9,980 units. The stepwise calculation arrives at €184,345. The difference of about €53,600 is an under-absorption: through its rates, costing charges only about 72 % of the fixed costs that actually occur in the cost center (RGK €91,218 plus fixed MHR share €97,390 = €188,608).
The reason lies in the part's structure. The RGK surcharge of 35.1 % is calibrated to an annual volume of €260,000 direct labor. But the titanium housing ties up only €10.75 of labor per unit — at 9,980 units that is €107,285, i.e. 41 % of the calculated labor volume. The machine side, by contrast, is covered at 100 %: the 3,393 machine hours are fully loaded.
From this follows a practical rule: parts with high material and machine share but low labor share under-absorb their cost center's residual overhead. Whoever fills their plant exclusively with such orders covers their machines but not their staff organization. Full costing would not have shown this — it would only have reported that every unit brings €23.84 profit.
Many ERP systems show a “contribution margin” or “margin %” on the sales document. By default it is usually calculated as revenue minus unit cost or standard price — that is, against full cost of production, not against variable costs. A warning “margin negative” there means: below cost of production, not: below the price floor. Check which cost value sits behind it in your system before you use it to decide on additional orders.
13.3 Make-or-buy: the classic fallacy
Many companies compare the buy price with the full costs of in-house production. That is wrong: on buying, only the variable costs fall away. The fixed costs remain — as long as the capacity is not actually reduced.
| Item | Make | Buy | Difference |
|---|---|---|---|
| Wrong: full-cost comparison | |||
| MEK + variable costs | €48.50 | — | — |
| Fixed costs (depr., interest, OH) | €61.50 | — | — |
| Full cost in-house | €110.00 | €95.00 | −€15.00 |
| → Conclusion "buy is cheaper" — and it is wrong | |||
| Right: partial-cost comparison | |||
| Variable costs in-house | €48.50 | — | — |
| Buy price (decision-relevant) | — | €95.00 | +€46.50 |
| → Make is €46.50 per unit cheaper — the fixed costs occur on buy anyway | |||
The calculation flips as soon as the capacity is actually reduced — machine sold, staff reassigned — or when the freed capacity can be filled with better-paid orders. Then the comparison figure is not the fixed-cost block but the forgone contribution margin of the alternative. Add to that the non-calculable criteria: know-how loss, supplier dependency, quality risk, responsiveness to changes.
13.4 Additional order below full cost: when yes, when no?
A customer offers €140/unit for the titanium housing from Chapter 11 — the calculated total cost is €158.93/unit. Decline? The answer depends first on capacity — and then on four conditions.
| Item (CC 210) | without order | with order (€140, lot 500) | Δ |
|---|---|---|---|
| Utilization | 70 % (2,375 h) | 75 % (2,545 h) | +170 h |
| Revenue / year | €1,276,648 | €1,346,648 | +€70,000 |
| Variable costs (€121.91/unit) | €851,541 | €912,496 | +€60,955 |
| CM I | €425,107 | €434,152 | +€9,045 |
| Fixed costs: cost center + allocation (unchanged) | €423,038 | €423,038 | €0 |
| Result CC 210 | €2,069 | €11,114 | +€9,045 |
At €140/unit each part covers its variable costs of €121.91 and yields €18.09 of contribution margin. Since the fixed costs occur anyway, every unit above the variable-cost threshold improves the result — even though the price is below full cost. Conversely at full utilization: then the additional order displaces regular orders with a higher CM — here the relative contribution margin per bottleneck hour applies (13.6), and the order should be declined.
The additional order only changes the slope from the actual point onwards: €18.09 CM per unit instead of €60.86. The base volume stays at the regular price. A revenue line at €140 starting from the origin describes a different case — price erosion for the entire volume (fig. 13.3). Valued at the additional spending of €98.71, the same order brings +€20,645 instead of +€9,045.
1. Capacity is really free in the delivery period — not on annual average, but in the
weeks in which production takes place.
2. No step-fixed costs arise — no overtime, no extra shift, no temporary staff.
3. No other order is displaced — otherwise 13.6 applies.
4. The price does not spill over — a discount for a regular customer is the new price at
the next annual negotiation. Additional orders below full cost belong to cases that can be ring-fenced:
a different part, a different market, a one-off quantity.
Whoever accepts purely on “contribution margin positive” fills the shop and still does not cover the fixed costs (13.5).
The €121.91 treats direct labor (€10.75), material overhead (€7.65) and the fixture (€4.80) as variable. For costing that is correct. In the individual case it is too strict: at 70 % utilization the workforce is there anyway, so is purchasing, and the fixture already exists. Additional cash then goes out only for material €85.00, variable machine costs €6.01, special cutters €4.20 and freight €3.50 — together €98.71. That is the absolute floor for the one-off case. €121.91 applies as soon as hours are bought in or fixtures are newly built for the order.
13.5 Break-even
Step 1 — the cost center carries itself:
= €188,608 / €60.86 = 3,100 units/year = 1,054 machine hours = 31 % of capacity
Step 2 — it also carries its share of administration and sales:
= (€188,608 + €234,430) / €60.86 = 6,951 units/year
equals 6,951 × 0.34 h = 2,363 machine hours
of 3,393 available hours = 70 % of capacity
Only from the 6,952nd unit does the cost center earn a profit. Below 31 % utilization it does not even carry its own fixed costs, below 70 % not its share of the company — regardless of how well each individual order is calculated. The 70 % utilization from 13.4 sits exactly at this threshold: that is why the result there is only €2,069.
Price erosion: when the exception price becomes the regular price
| Price | CM I per unit | Break-even | Utilization | Result at 9,980 units |
|---|---|---|---|---|
| €182.77 (regular price) | €60.86 | 6,951 units | 70 % | +€184,345 |
| €164.30 (minimum price) | €42.39 | 9,980 units | 100 % | ±€0 |
| €160.00 | €38.09 | 11,106 units | 111 % | −€42,900 |
| €140.00 | €18.09 | 23,385 units | 234 % | −€242,500 |
Every price above €121.91 yields a positive contribution margin. Whether it covers the fixed costs is decided by capacity: below €164.30 the break-even lies beyond 9,980 units — the shop is full, the result negative. The variable floor applies to the single order, this one to the entire volume.
Unit costs: the full-cost death spiral
| Utilization | units/year | fixed cost per unit | full cost | vs. price €182.77 |
|---|---|---|---|---|
| 50 % | 4,990 | €84.78 | €206.69 | −€23.92 |
| 70 % | 6,985 | €60.56 | €182.47 | +€0.30 |
| 85 % | 8,483 | €49.87 | €171.78 | +€10.99 |
| 100 % | 9,980 | €42.39 | €164.30 | +€18.47 |
Whoever recalculates at full cost as volume falls raises the quoted price — and loses more volume. At 50 % utilization the titanium housing "costs" €206.69, at full utilization €164.30. The part is the same. Reconciliation with Chapter 11: the costing shows €158.93, the curve ends at €164.30. The difference of €5.37 per unit is the under-absorption of residual overheads from 13.2 (€53,600 / 9,980 units).
Step-fixed costs: extra volume beyond capacity
Three-shift capacity (ch. 7.2): 5,088 h / 0.34 h = 14,965 units
Extra volume at which the shift pays for itself: €60,000 / €60.86 = 986 units → from 10,966 units
Additional order of 500 units with extra shift: €121.91 + €60,000 / 500 = €241.91/unit
The €60,000 is a model assumption, not a figure from the reference company. The logic holds for any amount: an additional shift only pays if the extra volume covers the step at least once — at €60.86 CM per unit, with 986 units. For a single order beyond capacity, the entire step is spread over its volume.
13.6 Contribution margin per bottleneck hour
CM I per unit answers the question "is this order worthwhile?". Once the machine is fully loaded, the question changes fundamentally: no longer "is A worthwhile?", but "A or B — which order gets the scarce spindle hour?". Here the per-unit contribution margin misleads.
| Order | CM I per unit | bottleneck time per unit | CM per bottleneck hour |
|---|---|---|---|
| A | €180 | 0.90 h | €200.00/h |
| B | €260 | 1.60 h | €162.50/h |
Order B, with €260, has the higher per-unit contribution margin and looks more attractive on the quotation sheet. Related to the scarce spindle hour, it clearly loses: A brings €200/h, B only €162.50/h. For every 100 bottleneck hours that run on A instead of B, €3,750 more contribution margin arises. Whoever prioritizes by per-unit CM when the machine is full gives money away.
| Ranking | hours A | hours B | CM I / year |
|---|---|---|---|
| CM per bottleneck hour (A first) | 1,500 h | 1,893 h | €607,613 |
| CM per unit (B first) | 993 h | 2,400 h | €588,600 |
| Difference: 507 h × €37.50/h | €19,013 |
When the machine is fully loaded, the short-term price floor rises above the variable costs: every accepted order displaces another. The displaced contribution margin per bottleneck hour must be added as opportunity cost. For the titanium housing the CM per machine hour is €60.86 / 0.34 h = €179.00/h — this value is also the right valuation figure for unplanned downtime, as used by the article on maintenance strategy.
13.7 Three price floors — one function of utilization
Short term, bottleneck, long term: these are not three rules but three sections of the same curve. Which one applies is decided by utilization in the delivery period.
| Situation | Recognisable by | Curve | Price floor | Rule | Sect. |
|---|---|---|---|---|---|
| Idle, one-off | staff and fixture in place, nothing bought in | result kink, fig. 13.1 | €98.71 | price > additional spending | 13.4 |
| Idle, standard case | free hours in the delivery period | result kink, fig. 13.1 | €121.91 | CM > 0 and four conditions | 13.4 |
| Price for the entire volume | price round, framework contract, market price | revenue fan, fig. 13.3 | €164.30 | break-even within capacity | 13.5 |
| Permanent price | series part, annual negotiation | unit costs, fig. 13.4 | €158.93 / €164.30 | cover full cost at planned activity | 11, 13.5 |
| Extra volume beyond capacity | extra shift, temporary staff needed | result step, fig. 13.5 | €241.91 (500 units) | extra volume ≥ step / CM = 986 units | 13.5 |
| Bottleneck | machine full, orders compete | CM per bottleneck hour, fig. 13.6 | €182.77 | rank by CM per bottleneck hour | 13.6 |
An additional order on a full machine can displace another order (€182.77) or trigger an extra shift (€241.91 for 500 units). The price floor is the cheaper of the two options — here displacement. If it displaces orders with a lower CM per bottleneck hour, the floor falls accordingly.
All curves with your own figures: Calculator: cost curves and price floors.
13.8 Every cost has an expiry date
Every costing system freezes a time horizon. The question is not which system is right — but whether its horizon fits your decision.
Fixed and variable are not properties of a cost type but a question of commitment period. Production labor is committed until the next shift plan, administrative and sales staff until the end of the notice period, the hall until the lease ends, the machine until the replacement decision. For a decision, exactly those costs count whose commitment ends or begins within the decision horizon. The longer the horizon, the more costs become controllable — and the higher the price floor.
| Decision horizon | commitment additionally ends for | + €/unit | price floor |
|---|---|---|---|
| single order, weeks | material, variable machine costs, special cutter, freight | 98.71 | €98.71 |
| quarter | production labor and CAM, material overhead, fixture | 23.20 | €121.91 |
| 1–3 years | residual overhead, administration, sales — staff and contracts | 27.26 | €149.17 |
| asset life, ≈ 13 years | imputed depreciation, interest, space, time-dependent maintenance | 9.76 | €158.93 |
The last step is exactly the full cost from Chapter 11. Whoever calculates with partial costs in the long run adds all fixed costs back step by step and arrives at the same figure. In the long run partial costing delivers nothing different — it only makes the path visible. Two subtleties remain: what counts is the future payment, not the past one; that is why depreciation is based on replacement value (6.4). And full costing is right in total, but per product only as good as the allocation — hence the €164.30 from 13.5.
| System | frozen horizon | treated as fixed | blind spot |
|---|---|---|---|
| Full costing | long — plant and staff permanently | nothing, everything allocated per unit | wrong for the single decision |
| Contribution margin | one period, usually the plan year | everything not reducible within the year | hard cut: what could be terminated in three months still counts as fixed |
| Flexible and marginal standard costing | the plan year | the same cut, built into planning | capacity steps only appear in the next plan |
| Activity-based costing | medium to long term | little — overheads controllable via activities | assumes back-office capacity is actually reduced |
| Target costing | life cycle incl. development | nothing | no tool for short-term decisions |
| Relative direct costing (Riebel) | none — several, staggered | ordered by commitment period and notice period | makes commitment explicit; laborious, rare in practice |
Sunk costs: if the material is already in stock, the price paid is irrelevant to the decision. What
counts is the replacement cost or what the material would earn elsewhere.
Cost stickiness: fixed costs build up faster than they come down. A shift is set up quickly but only
reduced through notice periods and contracts. Additional orders that tie up capacity permanently are therefore more
dangerous than one-off ones.
13.9 Volume risk: who carries the fixed costs when call-offs fall?
A framework agreement covers 9,980 titanium housings a year at €182.77. The customer reduces inventory and calls off only 6,985 units, or 70% of the planned volume. For the customer this is working capital management. For the contract manufacturer the cost center's result falls from €184,345 to €2,069. Cost accounting makes the damage visible; only the contract can prevent it.
Two degression effects, but only one depends on call-offs
| Batch-size degression | Utilization degression | |
|---|---|---|
| Costs | fixed per order: setup, CAM, first article | fixed per period: machine, building, staff, administration |
| Driver | quantity per batch | quantity per year |
| In this book | 10.4 and 17.3: setup share below 1–2% from batch 50 | 13.5: full cost €164.30 at full and €182.47 at 70% utilization |
| Lever | cut setup time (SMED), combine batches | allocate volume risk by contract, fill capacity elsewhere |
| When call-offs fall | the customer pays via the batch-size price scale | without a clause, the contract manufacturer pays |
Shorter setup times make small batches cheaper. They do nothing against a falling annual call-off, because the €423,038 of fixed costs stay the same whether the part comes in four batches or forty. In a market downturn both effects come together, and if a shift is cut, the machine hour rate and the surcharge rates rise as well. The calculation is in 7.6.
The damage in figures
| Plan | 70% call-off | Difference | |
|---|---|---|---|
| Volume | 9,980 | 6,985 | −2,995 |
| Revenue (€182.77) | 1,824,045 | 1,276,648 | −547,396 |
| Variable cost (€121.91) | 1,216,662 | 851,541 | −365,120 |
| Contribution margin | 607,383 | 425,107 | −182,276 |
| Fixed cost incl. allocation | 423,038 | 423,038 | 0 |
| Result | 184,345 | 2,069 | −182,276 |
The difference equals the lost contribution margin exactly: 2,995 units × €60.86. Fixed costs have not risen; they are just spread over fewer units.
Five contract models at a 70% call-off
| Model | Clause | Compensation | Result | of plan |
|---|---|---|---|---|
| A · Pure unit price | €182.77/pc, no compensation | 0 | 2,069 | 1% |
| B · Volume corridor ±15% | fixed-cost share per unit below 8,483 pcs | 63,498 | 65,567 | 36% |
| C · Fixed-cost recharge | €42.39 per missing unit at the cut-off date | 126,954 | 129,023 | 70% |
| D · Two-part price | €35,253/month + €140.38/pc | in base fee | 129,023 | 70% |
| E · Take-or-pay on CM | €60.86 per missing unit | 182,276 | 184,345 | 100% |
C: compensation = max(0; P − x) × f
D: base fee = F / 12 · unit price = v + planned profit / P
E: compensation = max(0; P − x) × (p − v)
Below plan, C and D are economically identical. Both cover the fixed costs in full and earn €18.47 per unit delivered. The gap to E is €55,322, which is the profit on the volume not called off. Anyone who negotiates only fixed-cost cover gives that up. That can be right, but it should be a deliberate decision.
The difference between C and D lies in cash flow and enforcement. The base fee arrives monthly and is undisputed. The recharge arrives only at the cut-off date, as a claim against a customer who is trying to save money. In return, D costs upside: at 120%, the contract manufacturer earns €221,214 under D and €305,821 under the pure unit price.
A clause such as “recharge based on full cost of the actual volume” raises the unit price to €182.47 full cost plus markup. The markup then also applies to fixed costs that were never incurred additionally. The customer refuses, and the new price becomes the baseline for next year. The correct approach compensates only the under-recovery: missing volume times fixed-cost share per unit of the planned volume.
Consignment stock: working capital only changes balance sheets
A consignment stock solves the customer's problem, not the supplier's. The contract manufacturer produces in economic batches, the customer withdraws as needed and pays only on withdrawal. The inventory disappears from the customer's balance sheet and reappears on the supplier's.
| Item | Basis | Value |
|---|---|---|
| Monthly call-off | 6,985 / 12 | 582 pcs |
| Batch size | 3 months of demand | 1,746 pcs |
| Safety stock | ½ month of demand | 291 pcs |
| Average inventory | batch / 2 + safety stock | 1,164 pcs |
| Capital tied up | × manufacturing cost €131.94 (ch. 11) | €153,600 |
| Imputed interest | × 5.7% (yield curve) | €8,755/yr · €1.25/pc |
| Residual stock at phase-out | up to batch + safety stock | 2,037 pcs · €268,800 |
The residual stock risk of €268,800 is 130 times the actual result of €2,069. Three points therefore belong in the contract: interest compensation or a storage fee from the day of storage, here €1.25 per unit; a maximum stock level, for example two months of demand; and an obligation to take over residual stock at phase-out or a call-off stop, valued at manufacturing cost. The free storage period from the article on toll machining follows the same logic.
Result over call-off · shaded area = corridor
| Model | Clause | Compensation | Result | of plan |
|---|
Consignment stock held by the contract manufacturer
13.10 Checklist
- 1Costs cleanly split into variable and fixed — including within the MHR?
- 2On make-or-buy, only the variable costs compared, not full cost?
- 3Checked whether the capacity can actually be reduced?
- 4When the machine is fully loaded, prioritized by CM per bottleneck hour?
- 5On additional orders: displacement accounted for? → 13.6
- 6Qualitative criteria documented, not just calculated?
- 7Exception price checked: does the break-even at this price still lie within capacity? → 13.5
- 8Price floor chosen to match utilization in the delivery period — idle, bottleneck or permanent price? → 13.7
- 9Decision horizon defined: which costs become controllable within it? → 13.8
- 10Volume risk covered: corridor, recharge or base fee — and a deliberate decision on whether only fixed costs or profit as well are secured? → 13.9
- 11For consignment: interest compensation, maximum stock and take-over obligation for residual stock agreed? → 13.9
Chapter 14 picks up the OEE from Chapter 7.2 again — this time as a management task: how do you measure downtime reliably, how do you prioritize improvements, and what does one percentage point actually bring?