APS Cost

APS Cost Estimation: What Scheduling Decisions Actually Cost Per Unit

A money-first look at advanced planning and scheduling: the cost drivers hidden in changeovers, idle capacity, expedites, and missed adherence, and how to build an APS ROI case that survives scrutiny.

Scheduling rarely appears as a line on a quote, yet it moves cost per unit by 5 to 20% through capacity utilization and changeover burden. Start with the loaded machine rate, typically 90 to 180 dollars per hour for CNC and 250 to 600 dollars per hour for large presses, including depreciation, energy, and floor overhead. A part needing 6 minutes of run time carries 9 to 18 dollars of machine cost at those rates. Idle time you scheduled but did not fill still accrues that rate, so a work center at 70% utilization spreads the same fixed overhead across fewer units and lifts per-unit cost by roughly 43% versus 100%.

Changeover is the cost driver estimators most often bury. If a line loses 45 minutes per changeover and runs 8 changeovers per shift, that is 6 hours of lost production daily. On a line producing 400 units per hour at a 12 dollar contribution margin, that is 28,800 dollars of forgone margin per shift, or 7.5 million dollars annualized across 260 days. Spreading setup cost correctly means dividing total changeover minutes by batch size: a 45 minute setup over a 500 unit batch adds 5.4 minutes per 100 units, but over a 50 unit batch it adds 54 minutes, a tenfold swing in per-unit setup cost.

Expedites and premium freight are the silent tax of a schedule that slips. A single expedited air shipment to recover a late order runs 3 to 8 times ground cost, often 2,000 to 6,000 dollars per event. Overtime to recover a bottleneck adds 50% to direct labor, so a 28 dollar loaded rate becomes 42 dollars for those hours. Plants running below 90% schedule adherence typically spend 2 to 4% of cost of goods sold on unplanned recovery. The Schedule Adherence Cost Impact calculator converts your adherence percentage and volume into an annual dollar figure so the number stops hiding in variance accounts.

Build a defensible quote by separating the four true unit costs: material, direct labor, machine time, and scrap, then layer overhead by activity, not a flat percentage. For a machined bracket: 4.20 dollars material, 6 minutes at a 120 dollar machine rate equals 12 dollars, 3 minutes direct labor at 32 dollars loaded equals 1.60 dollars, plus a 3% scrap allowance adding about 0.54 dollars. That is 18.34 dollars before overhead. A flat 35% overhead pushes it to 24.76 dollars. Where estimates go wrong is applying that flat rate to a job that actually consumes triple the changeover time of the average.

Scrap and rework tie back to sequencing more than most estimators admit. A poorly sequenced schedule that forces frequent material or tooling changes raises first-piece scrap: every changeover on a stamping line commonly burns 5 to 30 scrap parts on setup. At 500 changeovers per year and 15 parts each, that is 7,500 scrapped parts. If each carries 2.80 dollars of material and value added, scrap alone is 21,000 dollars a year traceable to sequence, not to the process. The Changeover Sequence Savings calculator quantifies the recoverable portion so you can credit it against an improvement project.

The APS investment case rests on payback, and the honest inputs are software license, implementation, and data cleanup versus recovered margin. A mid-market APS runs 80,000 to 300,000 dollars in year one all in. Recoverable value comes from three buckets: 2 to 5 points of utilization, 20 to 40% less changeover time, and 1 to 3 points of adherence. On a 40 million dollar cost base, 3 points of utilization alone can free 1.2 million dollars of capacity. The APS ROI Payback calculator combines these so you present a payback in months, typically 8 to 18, rather than a vague productivity claim.

Quote risk lives in the gap between planned and actual capacity. If you quote at 95% utilization but the shop runs at 78%, your per-unit overhead absorption is off by more than 20% and every job is underpriced. Estimators should quote against realistic finite capacity, not nameplate. A press rated at 60 strokes per minute that averages 44 net after jams and setups delivers 73% of theoretical output, so a quote built on 60 loses money on volume. The Finite Capacity Load calculator gives the realistic denominator to price against.

Close the estimate by pricing the constraint, not the average station. Cost per unit is set at the bottleneck because that is where an hour is irreplaceable. An hour lost at a non-constraint costs nothing incremental, but an hour lost at the constraint costs a full unit of throughput times contribution margin. If the constraint yields 92 parts per hour at 14 dollars margin, every bottleneck hour is worth 1,288 dollars. Load that figure into the quote as the real opportunity cost of scheduling, and use the Bottleneck Schedule Impact calculator to confirm which resource actually sets your price floor.

Published 2026-07-01.