Cost Estimation

What Drives Manufacturing Cost Per Unit and How to Build a Defensible Quote

A money-first breakdown of what actually moves cost per unit and how to quote so the standard survives contact with the shop floor.

Cost per unit is a stack, and knowing the proportions tells you where to sharpen the pencil. In general fabrication and machining, material commonly runs 40% to 60% of total cost, direct labor 15% to 30%, and manufacturing overhead 20% to 35%. Flip those weights for the wrong industry and the quote breaks: a labor-light CNC part where material is 55% of cost needs tight purchasing, while a manual assembly at 40% labor needs an accurate time standard. Estimate the biggest bucket first. A 5% miss on a material line that is half the cost hurts twice as much as a 5% miss on labor.

Material cost is quantity times price, and both drift. Quote from the frozen standard price, not a spot quote that expires in 30 days, and gross the quantity up for expected yield loss. If a part needs 96 good units of stock but the process routinely scraps 4%, you must buy and cost 100 units, not 96. That single yield adjustment moves material cost up 4.2% before you touch price. Lock in raw prices with the supplier for the quote horizon, and if you cannot, add a stated escalation clause rather than absorbing a 6% to 10% commodity swing into a margin you already promised.

Labor cost is standard hours times a fully loaded wage, and the loaded rate is where estimates leak. A $22.00 base wage becomes $30.00 to $34.00 once you add payroll tax, benefits, and paid time off at a 35% to 55% burden. Then apply the time standard honestly: 11.0 standard hours on a job, not the 10.0 base hours before the setup and handling allowance. Estimators who quote base hours and a bare wage understate labor by 20% to 40% and never understand why the job loses money. Setup is the quiet killer on short runs, so a 3-hour changeover spread over 100 units adds $1.00 per unit at a $33 loaded rate.

Machine time and overhead ride on the burden rate you built from the plant's indirect pool. If a CNC cell recovers $65.00 per machine hour and the part needs 0.20 machine hours, that is $13.00 of overhead per unit before labor. The trap is using a single plant-wide rate. A capital-heavy cell and a labor-heavy cell cost very differently to run, so a blended rate overcharges the simple part and undercharges the complex one, distorting the quote by 10% to 30%. Quote from department-level rates and absorb at practical capacity, because absorbing at 100% theoretical capacity understates the rate and quietly under-recovers fixed cost every time volume dips.

Scrap and rework are cost, not a rounding error, and they compound. A scrapped unit carries its accumulated cost at the point of loss, material plus labor plus overhead absorbed so far, not just raw material. Scrap a part after 60% conversion and you have thrown away roughly 60% of the full unit cost, not 10%. At a 3% scrap rate on a $50 part where losses average 60% through the process, that is 0.03 x 0.60 x $50 = $0.90 per good unit shipped that the quote must recover. Rework adds the extra labor, material, and overhead to bring defects back to spec, often $5 to $15 per touched unit.

A defensible quote shows its work so a customer, or your own controller, can pressure-test it. List material at grossed-up quantity and standard price, labor at loaded rate and full standard hours, overhead at the correct department burden rate, then scrap and rework as explicit line items rather than a vague 5% fudge. Add SG&A and target profit last. When the job runs, feed actuals back through Actual vs Standard Cost and the variance tools so the next quote learns. A quote you cannot decompose into these buckets is a guess wearing a decimal point.

Estimates go wrong in predictable places, and every one has a dollar signature. The five most expensive: costing good units instead of gross units and eating the yield loss, quoting base labor hours without the setup allowance, using a plant-wide burden rate on a mixed cell, pricing scrap at raw-material value instead of accumulated cost, and holding a stale standard while material prices moved 8% to 12%. Any one of these can turn a quoted 25% gross margin into an actual 12% to 15%. Run the quote against the current standard cost before it goes out, and re-roll the standard the moment variance exceeds 5% of standard cost period after period.

The quote is only as good as the standard behind it, so treat standards as living numbers. Re-roll material standards when supplier prices move more than 5%, re-time labor standards when a process or fixture changes, and refresh burden rates annually or whenever the absorption base shifts by more than 10%. A plant that quotes off standards updated once a year against a market moving quarterly will systematically win the jobs it underpriced and lose the ones it overpriced, a selection bias that erodes margin without any single quote looking obviously wrong. Tie every quote to a dated standard, and the money stops leaking.

Published 2026-07-01.