Tooling

Managing Setup Cost: Batch Sizes and the Real Tradeoffs

Setup cost quietly sets your minimum profitable batch. How to amortize it correctly, cut it with SMED, and keep quotes priced from live floor data.

Setup cost decides money because it silently sets your minimum profitable batch. A changeover that takes 3 hours of a $45 per hour setup tech plus 3 hours of a machine worth $120 per hour costs $495 before the first good part. Spread over a 1,000 piece run that is $0.50 a unit; over a 50 piece run it is $9.90 a unit, which quietly turns a quoted 22 percent margin job into a loss. Plants that do not amortize setup into unit economics underprice small runs and overprice large ones, and their sales team cannot tell which orders feed the plant and which eat it.

The mechanics: setup cost equals setup hours times the sum of labor rate and machine rate, divided by batch quantity. Worked example: 2.5 hours, $38 labor, $95 machine burden, batch of 400. Cost is 2.5 times 133, or $332.50, which is $0.83 per unit. Run the same die on an 80 piece order and it is $4.16 per unit. The Setup Cost calculator does the amortization; your job is the inputs. Use loaded labor rates, and on a bottleneck machine use contribution margin per hour instead of accounting burden, because an hour of setup on a sold-out asset costs whatever that hour would have earned, often 2 to 3 times the burden rate.

Benchmarks: setup as a share of available machine time runs 5 to 15 percent in typical job shops and under 5 percent in well-run cells; above 20 percent, setup is your real product. Per-event times: CNC machining changeovers commonly run 30 minutes to 2 hours, stamping die changes 1 to 4 hours untreated, injection mold swaps 1 to 3 hours. On unit economics, setup above 10 percent of a part's total cost signals a batch size problem or a setup time problem. The classic tension: doubling batch size halves per-unit setup cost but grows average inventory, which carries 18 to 25 percent of its value annually in holding cost.

Three levers, in order of return. First, cut setup time itself with SMED discipline; converting internal work to external, staging tools and programs before the machine stops, routinely cuts 30 to 50 percent with near-zero capital. Second, sequence the schedule: running similar jobs back to back, light to dark colors, same tooling families, cuts effective changeover time 20 to 40 percent without touching the setup method. Third, right-size batches with real math instead of habit; when setup drops from 3 hours to 1, the economic batch shrinks by roughly 40 percent, and inventory falls with it. Plants that cut setup time but keep old batch sizes capture half the prize.

Failure modes: first, amortizing setup over the quoted quantity but running short; if the customer takes 250 of a quoted 400, your realized setup cost per unit just rose 60 percent, so track quoted versus actual run quantities monthly. Second, ignoring first-article and warm-up scrap; a setup that consumes 6 pieces of $28 material adds $168 that never hits the setup account. Third, treating all machine hours alike; setup on an idle machine costs labor, setup on the constraint costs throughput, and pricing both the same misallocates the whole schedule. Fourth, hiding setup inside cycle time standards, which makes every batch size decision invisible to the people making it.

Cadence: daily, record actual setup duration per changeover on the machine board next to the standard; anything 25 percent over standard gets a one-line cause note. Weekly, review total setup hours by machine and the five worst overruns; feed repeat causes, missing tools, waiting on programs, crane conflicts, into the SMED list. Monthly, recompute setup cost per unit for the top 20 SKUs at actual batch sizes and hand the numbers to whoever quotes, because stale setup assumptions in the quote model are how job shops lose money politely. Quarterly, re-run batch size economics on any machine where setup time moved more than 20 percent.

World-class setup management looks like this: standard setup times documented per part and machine pair and hit within 10 percent, setup under 5 percent of machine hours, single-digit-minute changeovers on the machines that change over most, and quotes that price setup from live data. The compounding effect is the point: a shop that cuts average setup from 2 hours to 45 minutes can profitably accept runs one third the size, which usually opens 10 to 20 percent more addressable orders. Setup cost per unit, reviewed monthly by SKU, is the number that keeps scheduling, sales, and the floor arguing from the same sheet.

Published 2026-07-02.