Sign in

Blog · How-to guides · Manufacturing

Data analytics in manufacturing: questions and measures

Data analytics in manufacturing compares production, scrap, inventory and costs across plants and periods. Work through a checked plant example and assign the findings to the managers who can investigate them.

The short answerData analytics in manufacturing uses production, scrap, inventory and cost records to compare operating performance. Managers calculate measures such as accepted units per labor hour, scrap rate and production cost per accepted unit, then investigate differences by plant and month. The useful result is a reconciled comparison with consistent products, periods and cost definitions, rather than an unexplained plant ranking.

Data analytics in manufacturing compares production, scrap, inventory and costs to show where operating performance differs. Start with the monthly business exports managers already use, calculate measures on a common basis, and reconcile them before assigning an action. A plant comparison is useful only when its units, scope and reporting dates agree.

Start with a management question

Ask a question that a named manager can act on: why does one plant spend more per accepted unit, why is finished stock rising, or why did labor hours grow faster than output? Each needs a comparison and a check. A collection of unrelated measures makes the meeting longer without identifying the decision.

Keep this review focused on operating summaries. Customer coverage, distributor sales and parts opportunities belong in the separate sales KPIs for industrial manufacturers. Those questions need customer and installed-equipment records rather than production counts.

For a monthly plant review, request these exports:

Export Rows and fields to retain Question it answers
Production Plant, product, month, units started, accepted units, scrapped units How much usable output was completed?
Labor Plant, product, month, included labor hours How many hours supported that output?
Production cost Plant, product, month, materials, labor, production overhead What did producing it cost?
Finished inventory Opening units, receipts, shipments, closing units, adjustments Where did completed output go?

Set a common cutoff and keep each report at its own original detail until totals reconcile. Combining a monthly cost total with several daily production rows can repeat the same cost many times.

Choose measures with explicit denominators

Scrap rate = Scrapped units / Units started

Labor productivity = Accepted units / Included labor hours

Production cost per accepted unit = Matched production cost / Accepted units

These definitions fit the example below, where every started unit is either accepted or scrapped during the month. A business with unfinished production needs a separate work-in-process reconciliation; it cannot assume that started units equal completed dispositions.

State the cost policy before comparing plants. The IRS inventory guidance describes produced merchandise cost as the direct and indirect costs that must be capitalized under the applicable rules. That supports documenting the included cost categories; your controller must still determine the company's accounting treatment.

Keep raw materials, work in process and finished goods distinct. The Census Bureau's manufacturing definitions distinguish those inventory stages and define finished goods as completed products ready for shipment. An increase in unfinished production answers a different question from an increase in ready-to-ship stock.

A worked comparison of two US plants

An illustrative US manufacturer produces the same standard pump housing at its Ohio and Indiana plants. These are September 2026 results, in USD, under one cost policy. There is no opening or closing work in process, no rework, and no difference in product specification.

Plant Started units Accepted units Scrap units Labor hours Production cost (USD)
Ohio 10,500 10,000 500 2,000 500,000
Indiana 8,400 8,000 400 1,800 440,000
Total 18,900 18,000 900 3,800 940,000

Apply the three formulas to each row. Round scrap rates to one decimal place and productivity and unit costs to two; retain unrounded values for totals.

Plant Scrap calculation Scrap rate Accepted units / hour Cost / accepted unit (USD)
Ohio 500 / 10,500 4.8% 10,000 / 2,000 = 5.00 500,000 / 10,000 = 50.00
Indiana 400 / 8,400 4.8% 8,000 / 1,800 = 4.44 440,000 / 8,000 = 55.00
Total 900 / 18,900 4.8% 18,000 / 3,800 = 4.74 940,000 / 18,000 = 52.22

The first check is physical: 18,000 accepted units plus 900 scrap units equals 18,900 started units. The cost check uses unrounded rates: 10,000 × $50 plus 8,000 × $55 equals $940,000. Calculate company measures from summed numerators and denominators, rather than averaging the displayed plant rates.

Reconcile output to finished inventory

Use the same product, plants and month. Assume there are no transfers, customer returns or inventory adjustments in this example.

Plant Opening units Accepted production Shipped units Calculated closing units Recorded closing units
Ohio 3,000 10,000 9,500 3,500 3,500
Indiana 2,500 8,000 8,200 2,300 2,300
Total 5,500 18,000 17,700 5,800 5,800

Opening stock plus production minus shipments gives 5,500 + 18,000 − 17,700 = 5,800 units. Stock rose by 300 units overall: Ohio added 500 while Indiana used 200 from opening stock. Match recorded closing quantities to the inventory report before discussing excess production.

This quantity reconciliation does not establish inventory value. For the financial measure of how quickly inventory turns into sales, use inventory turnover with matching cost and inventory values.

Turn the differences into accountable questions

Indiana's unit cost is $5 higher. Applied to its 8,000 accepted units, that is a $40,000 difference from Ohio's rate, not a promised saving. Finance should split the difference into materials, labor and overhead using the dedicated manufacturing cost analysis method.

Both plants have the same scrap rate before rounding, so this table does not support blaming scrap for the cost gap. Indiana also produces fewer accepted units per labor hour. Its plant manager should examine staffing coverage, recorded downtime and work assignment summaries before drawing a productivity conclusion.

Ohio's manager and the inventory owner should investigate why production exceeded shipments. Confirm order timing and the agreed stock requirement before reducing output. This moves from observing a difference to explaining it, the job of diagnostic analytics.

Where it goes wrong

  • Different products in one unit count. A complex assembly and a simple housing cannot share an unqualified productivity comparison.
  • Hours with different scope. Including supervisors at one plant and only direct labor at another changes the denominator.
  • Unfinished work ignored. Costs incurred this month may relate to units completed later; reconcile work in process first.
  • Inventory movements omitted. Transfers and adjustments must appear in the stock bridge when present.
  • A rate treated as a cause. Higher unit cost identifies where to investigate; it does not establish which manager or expense caused it.

Compare your own manufacturing exports in Covirage

Bring a sample of your own production, inventory and cost exports to Covirage. Covirage maps the columns; its tools compute and check the comparisons, and the external AI model explains the results 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 file.

See Covirage for industrial manufacturers for the broader business context.

Questions people ask

What data should a manufacturing manager start with?

Start with monthly production summaries, scrap counts, labor hours, inventory movements and production cost reports. Keep plant, product family and reporting period on every row. These business exports can answer a focused operating question before the company considers any additional collection work.

Is the plant with the lowest unit cost always the best?

No. Product complexity, purchased content, overhead allocation and capacity usage can change unit costs. Compare like products under the same cost policy, then review delivery and quality alongside cost. A lower reported rate is a reason to investigate, not proof of better management.

Can production output and shipments be compared directly?

They can be reconciled, but they describe different movements. Accepted production adds finished units; shipments remove them. Opening stock plus accepted production minus shipments should equal closing stock when there are no transfers, returns or adjustments. Include those movements when they exist.

Who owns manufacturing analytics?

The plant manager owns operating follow-up, while finance owns cost definitions and ledger reconciliation. Production, quality and inventory owners confirm the underlying counts and cutoffs. Agreeing on these responsibilities makes the comparison usable when reports from different departments disagree.