Cost per unit produced is the total cost of making one good unit, direct material, direct labor, and applied overhead (machine time, energy, tooling, and facility cost) divided by the count of sellable units. Scrap and equipment losses raise it by shrinking the denominator while the costs stay put.
Most plants carry a “standard” unit cost set once a year and quoted to customers. The number that runs the floor is different: it is the actual cost of the units that shipped this week, and it moves every shift with yield, downtime, and speed. This guide builds a true unit cost from the four buckets that feed it, shows how OEE losses inflate the figure invisibly, and works a hand-checkable example. If you want the loss side quantified first, the free OEE calculator gives you the denominator this whole calculation depends on.
What is cost per unit produced?
Cost per unit produced is total production cost for a period divided by the good units that came out of that period. The formula is simple; the honesty is in the two halves. On top you sum every cost consumed to run, material, labor, and the overhead that keeps the line moving. On the bottom you count only units you can sell. A part that gets scrapped still consumed its material and its share of labor and machine time, so its cost does not disappear; it reloads onto the good units that survived.
That is the trap in the metric. Two lines can buy identical material at the identical price and pay operators the identical wage, and still post unit costs that differ by 20%, entirely because one line converts more of what it starts into sellable product. Unit cost is a yield-and-effectiveness number wearing an accounting costume.
What goes into a true unit cost?
A true unit cost has four buckets, and the fourth is the one standard costing usually buries. Keep them separate so you can see which one is moving:
| Bucket | What it covers | Moves with |
|---|---|---|
| Direct material | Raw material, components, packaging that end up in the product | Purchase price, usage, material yield |
| Direct labor | Operator hours paid to run and change over the line | Wage, crew size, run rate, downtime |
| Machine time & overhead | Energy, depreciation, tooling, maintenance, facility cost applied per run-hour | Uptime, speed, throughput |
| Scrap & rework | Material and time lost to defects, plus the labor to rework recoverable units | First-pass quality, defect rate |
The first bucket is mostly a purchasing story. The other three are effectiveness stories: they are fixed or semi-fixed per hour, so what determines their cost per unit is how many good units that hour produced. That is why unit cost and OEE are the same conversation from two directions, and why a plant chasing unit cost with only a purchasing lever is fighting with one hand.
How do OEE losses inflate cost per unit?
OEE losses inflate unit cost by shrinking the denominator while the fixed costs above the line barely move. A crew is paid the same for an eight-hour shift whether the line runs at 85% OEE or 55%. The building draws the same power for lighting and HVAC. Depreciation does not pause during a breakdown. So when downtime slow cycles, and scrap take a third of your output, roughly a third more fixed cost piles onto every unit that makes it out the door.
Work an example. Say fixed labor and overhead for a shift total \$4,000. At the line’s nameplate rate the shift would yield 10,000 good units, so fixed cost per unit is \$0.40. Run the same shift at 60% OEE and you get 6,000 good units, the same \$4,000 now spreads to about \$0.67 per unit. The material cost per unit did not change. The wage did not change. Unit cost rose 67% on the fixed portion, and nothing on a purchase order explains it. It hides inside the six big losses.
How do you calculate cost per unit produced, step by step?
Pick a period short enough to act on, a shift or a run, not a quarter, and work these steps in order:
- Fix the period and pull the good-unit count. Count only first-pass good units. Reworked units cost more, not less, so do not quietly credit them as if they came out clean.
- Total direct material. Material issued to the run at standard or actual price, including the material embedded in scrapped parts. Do not net out scrap here; you want to see it.
- Total direct labor. Crewed hours × loaded wage for the period, including changeover and cleanup time, not just clean run time.
- Apply machine time and overhead. Use a per-run-hour rate that carries energy, depreciation, tooling, and maintenance. Multiply by the hours the line was staffed, because that is what you actually paid for.
- Add the scrap and rework load. Material and labor consumed by defects that produced nothing sellable, plus the extra labor to rework recoverable units.
- Divide and decompose. Sum the four buckets, divide by good units, then split the result into per-unit material, labor, overhead, and scrap so you know which bucket to attack. Recompute every period and watch the trend, not the single number.
What does a worked example look like?
These numbers are hypothetical chosen so you can check the arithmetic by hand. One shift on a packaging line:
| Input | Value |
|---|---|
| Good units produced (first pass) | 6,000 |
| Direct material (incl. scrapped material) | $3,000 |
| Direct labor (crew × hours × wage) | $1,600 |
| Machine time + applied overhead | $2,400 |
| Scrap + rework load | $500 |
| Total cost | $7,500 |
Cost per unit produced = \$7,500 ÷ 6,000 = \$1.25. Decomposed, that is \$0.50 material, about \$0.27 labor, \$0.40 overhead, and about \$0.08 scrap. Now the lesson: if this line had run at nameplate and produced 8,000 good units on the same staffed shift, labor and overhead (\$4,000 combined) would have spread over 8,000 units instead of 6,000. Fixed cost per unit falls from about \$0.67 to \$0.50, and total unit cost drops toward \$1.08, a 14% cut with no new supplier, no wage change, and no capital. The lever was effectiveness, tracked through throughput.
Why does standard cost hide the real number?
Standard cost is a budget: a per-unit figure set from expected material prices and an assumed run rate, used to quote jobs and value inventory. It is useful for planning and useless for improvement, because it bakes in an assumed OEE and then stops moving. When actual OEE drops below the assumption, the extra cost shows up later as an unfavorable variance on a monthly report, long after the shift where you could have done something about it.
The gap between standard and actual is the cost of poor effectiveness and it is real money. A plant that reports only standard cost is flying on a number that was true last January. Measuring actual cost per unit each shift, or better, computing it live from the same machine signals that feed OEE, turns unit cost from a backward-looking variance into a scoreboard the crew can move today. That is the same argument behind measuring at the source rather than reconstructing it from memory at month-end (see the platform).
How do you drive cost per unit down?
Attack the bucket that is actually moving, in this order of usual payoff:
- Effectiveness first. Every point of OEE you recover spreads fixed labor and overhead over more units. Chase downtime and speed losses before you reopen a supplier contract, the internal lever is faster and free. Cutting cycle time lifts the same denominator.
- Yield next. Scrap is triple waste: lost material, lost machine time, and lost capacity. Driving first-pass yield up shrinks the fourth bucket and grows the denominator at once.
- Changeover. Setup time is unstaffed capacity you paid for. Quick-changeover methods convert it back into sellable units, especially on high-mix lines running in small batches.
- Material and price last. Real, but slower and often already tight. It rarely moves unit cost as fast as recovering a line that runs at 55%.
Track the components against a baseline the way a plant tracks its other manufacturing KPIs and pair the number with the loss detail behind it, the six big losses tell you which bucket is bleeding. See how one plant tied unit economics to floor effectiveness in the CLS case study.
Data & sources
The link between cost and effectiveness is not a Harmony idea; it is how national statistics agencies define productivity and cost.
- The U.S. Bureau of Labor Statistics defines unit labor cost as hourly compensation divided by output per hour, algebraically, higher productivity lowers unit cost even when wages hold. That is the fixed-cost-spread effect at the level of a whole economy.
- BLS publishes labor productivity and cost series for the manufacturing sector, the same output-per-hour logic that governs a single line.
- For macro context on how far real plants run below capacity, the Federal Reserve’s G.17 release tracks U.S. manufacturing capacity utilization, which has held in the mid-70s percent range in recent years, a reminder that the denominator in any unit-cost calculation is usually well short of its ceiling.