Blog · Board and management reporting · Manufacturing
Production analytics connects accepted output, scrap, recorded labor, and production costs in a manufacturing operating review. This example shows why better yield and labor productivity can coexist with a higher cost per good unit.
Production analytics compares accepted output, scrap, recorded labor, and production costs to explain what changed in a manufacturing operation. Start with matching plant reports and approved cost schedules, then calculate measures on the same product and period boundary. A higher output total becomes useful when the review shows the hours and dollars behind it.
The review connects a small set of business questions. Did the plant produce more accepted units? Did it use more labor hours? Did fewer inputs become scrap? Did the cost of an accepted unit improve?
Use accepted output rather than every unit touched by a machine. A repaired part may appear on several operation records but should count once in the final output total. Keep rework hours in the labor boundary if the review includes the units those hours recovered.
The EPA's Six Sigma guidance describes establishing performance metrics and baseline data before investigating quality problems. It also distinguishes scrapped products from defective products that require additional resources to repair. That distinction matters when improving accepted output hides continued rework effort.
Use yield analysis for the detailed reconciliation of starts, accepted output, scrap, and unfinished work. Production analytics carries those approved quantities into a broader labor and cost discussion.
Bring a production-order export with plant, product, period, starts, accepted units, final scrap, and opening and closing work in process. Add shift only if all the other files can support that comparison without allocating an entire month's cost to whichever shift happens to appear first.
Operations approves quantities and recorded hours. Quality approves the acceptance and scrap definitions. The controller approves cost categories, accruals, and the treatment of unfinished production. Procurement helps explain documented material-price changes, while maintenance validates recorded equipment-loss reasons.
Labor files need a stated boundary: direct production hours, paid hours, or another approved measure. Do not put paid hours including leave in one month's denominator and direct hours in another. If contractors contribute to output, include their relevant hours consistently or disclose the omission.
Cost files need the same discipline. State whether material, direct labor, and overhead are actual, standard, or allocated amounts. A report that mixes actual wages with standard material cost can still answer a defined question, but its total is not an all-actual production cost.
An illustrative Ohio manufacturer reviews one unchanged bracket product. Both periods have no opening or closing work in process, and the approved schedules assign the listed costs to the accepted output. Costs include consumed material, recorded direct labor, and allocated production overhead; selling and corporate expenses are outside this boundary.
| Recorded input | Prior month | Current month |
|---|---|---|
| Units started | 10,000 | 12,000 |
| Accepted units | 9,600 | 11,760 |
| Final scrap units | 400 | 240 |
| Direct labor hours | 480 | 560 |
| Material cost (USD) | 48,000 | 60,000 |
| Direct labor cost (USD) | 14,400 | 17,920 |
| Allocated production overhead (USD) | 9,600 | 11,200 |
First prove the physical totals. Prior-month starts equal 9,600 accepted plus 400 scrap. Current-month starts equal 11,760 accepted plus 240 scrap. No unfinished quantity or rework activity count needs to be added to either identity.
Then calculate the review measures:
Accepted-output yield = Accepted units / Units started
Labor productivity = Accepted units / Recorded direct labor hours
Recorded production cost per good unit = (Material + Direct labor + Allocated production overhead) / Accepted units
| Calculated measure | Prior month calculation | Current month calculation |
|---|---|---|
| Accepted-output yield | 9,600 / 10,000 = 96.00% | 11,760 / 12,000 = 98.00% |
| Scrap rate | 400 / 10,000 = 4.00% | 240 / 12,000 = 2.00% |
| Good units per labor hour | 9,600 / 480 = 20.00 | 11,760 / 560 = 21.00 |
| Total production cost (USD) | 48,000 + 14,400 + 9,600 = 72,000 | 60,000 + 17,920 + 11,200 = 89,120 |
| Production cost per good unit (USD) | 72,000 / 9,600 = 7.50 | 89,120 / 11,760 = 7.58 |
| Recorded labor cost per hour (USD) | 14,400 / 480 = 30.00 | 17,920 / 560 = 32.00 |
Displayed rates and unit costs are rounded to two decimals. Calculations retain unrounded values.
Accepted output grew by (11,760 / 9,600 - 1) = 22.50%. Labor hours grew by (560 / 480 - 1) = 16.67%, so good units per hour improved by (21 / 20 - 1) = 5.00%.
The plant produced more accepted output per recorded hour, yet recorded cost per good unit increased. Total cost grew by (89,120 / 72,000 - 1) = 23.78%, slightly faster than accepted output. Using unrounded unit costs, the increase per good unit is 1.04%.
Break the cost into its approved components:
| Cost component per accepted unit (USD) | Prior month | Current month |
|---|---|---|
| Material | 48,000 / 9,600 = 5.00 | 60,000 / 11,760 = 5.10 |
| Direct labor | 14,400 / 9,600 = 1.50 | 17,920 / 11,760 = 1.52 |
| Production overhead | 9,600 / 9,600 = 1.00 | 11,200 / 11,760 = 0.95 |
The unrounded components sum to the unit costs above. Rounded components can differ from the rounded total. Material and labor per good unit rose while overhead per good unit fell. The higher recorded hourly labor cost is visible, but these rows cannot determine whether it came from pay changes, overtime, or staffing mix.
Ask procurement for material quantities and prices, and ask payroll for the labor-rate detail. Cost variance analysis can then separate a supplied standard from actual cost; manufacturing cost per unit owns the fuller cost-boundary discussion.
Tie accepted quantities to approved production reports, hours to the relevant labor records, and cost components to the controller's schedules. If multiple reports repeat a cost at order and operation level, remove the duplication before summing.
For spreadsheet summaries, Google Sheets SUMIFS adds rows meeting multiple criteria. Use the same plant, product, and period filters for each quantity and cost column, then divide the totals. Covirage is not affiliated with Google.
Bring a sample of your own production report, labor records, and approved cost schedule with matching plant and period identifiers. Covirage maps the columns, and its tools compute and check the figures; the external AI model explains the comparison with every figure traced to its rows. Create a free Covirage account, which takes about a minute, then set up your business and bring your files.
See cost variance analysis in Covirage for the supporting business application.
Start with production orders showing starts, accepted output, scrap, and unfinished quantities; recorded labor hours; and approved production cost schedules. Include plant, product, shift, and period identifiers so quantities, hours, and dollars refer to the same operating boundary.
Output measures the quantity produced. Labor productivity divides a consistent output quantity by recorded labor hours. Output can rise because the plant worked more hours, because productivity improved, or because the product mix became easier to produce.
A business review can compare recorded downtime and its reported reasons, but those records do not establish an engineering diagnosis. Maintenance and engineering owners investigate causes. Keep operating comparisons separate from equipment failure prediction and process control.
Material prices, labor rates, overhead allocation, and product mix can move independently of good output per labor hour. Compare each cost component per accepted unit using the same accounting boundary before attributing an increase to production efficiency.