The hidden factory is the work and capacity consumed by defects, rework, workarounds, queues, and other non-value-adding activity that ordinary reports do not join. Internal failure costs are the costs of defects found before the customer receives the product. Start with one event, map its quality and operating consequences, and calculate a first COPQ figure without pretending the worksheet is a universal accounting formula.
The hidden factory concept
The hidden factory is the unofficial operation that grows inside the official one. It includes the time, material, labour, machine capacity, inspection, troubleshooting, and coordination consumed when the process does not run as designed. The World Economic Forum describes hidden factories as workarounds that become institutionalized. An OEE explanation frames the same idea as untapped production potential.
Those descriptions are compatible, but they answer different parts of the problem. The hidden factory is not a second building. It is the unofficial work that makes the planned process possible, plus the capacity that work consumes. A plant can have acceptable final output and still spend a shift correcting defects, waiting for disposition, repeating inspections, or rescheduling work.
The term is useful when it leads to a recordable question: which event created this work, which records show it, and who can decide whether the condition should be accepted, corrected, or prevented?
Hidden factory examples in a plant
The hidden factory usually appears as ordinary work, not as a dramatic failure. Look for the repeated activities that do not appear together in a management report:
- A batch is held, inspected again, corrected, and returned to the line.
- A changeover runs long, creating extra downtime and startup scrap.
- Operators use a spreadsheet or informal instruction to work around a master-data or process problem.
- Maintenance defers a recurring issue because the line can still run, although at a slower speed or with more stops.
- Product waits for quality disposition while the next process step is rescheduled.
- A final-good number includes units that became acceptable only after another processing or inspection loop.
The World Economic Forum recommends looking for missing time, rescheduling, maintenance backlogs, pressure, and queues. These are clues, not a formula. The useful next step is to attach each clue to a specific event and source record.
Do not start by adding every possible loss to one total. Start with one product, line, shift, or event window. Record what happened, what should have happened, and where the evidence lives. This protects the calculation from becoming a new number that no team trusts.
Cost of quality: the four categories
The American Society for Quality defines cost of quality as a method for understanding resources used to prevent poor quality, appraise quality, and respond to internal and external failures. The four categories below are a classification aid. They are not a universal chart-of-accounts formula.
| Category | When it occurs | Manufacturing examples |
|---|---|---|
| Prevention | Before the defect occurs | Process design, training, preventive controls, supplier or process planning |
| Appraisal | When checking whether requirements are met | Incoming inspection, in-process checks, testing, audits, measurement equipment |
| Internal failure | After a defect occurs but before the customer receives the product | Scrap, rework, rectification, reinspection, failure analysis, approved internal recovery work |
| External failure | After the customer receives the product or service | Returns, warranty work, complaint handling, field repair, recall response |
The timing boundary matters. A quality check is appraisal cost when it is performed to determine conformance. The cost of correcting a defect found by that check is an internal failure cost. If the same defect reaches the customer, the response moves into external failure.
The categories also prevent a common mistake: treating every quality-related hour as poor-quality cost. Prevention and appraisal are planned activities that can reduce failure. Internal and external failure costs are the consequences of a failure that occurred. Finance should decide how local overhead, shared labour, recoverable scrap, and contribution margin are treated.
Internal failure costs
Internal failure costs are the costs of remedying defects discovered before delivery to the customer. ASQ lists waste, scrap, rework or rectification, and failure analysis. AccountingTools also lists scrapped product, rework, and throughput lost when a bottleneck operation’s margin is irrecoverably lost.
For a manufacturing event, the first internal-failure list should ask:
- What material was scrapped, downgraded, or consumed by the defect?
- What labour was spent on correction, sorting, retest, inspection, cleaning, or documentation?
- What machine time, utilities, maintenance, or testing was required because the defect occurred?
- Did the event consume a constrained production window, or did it use time that would otherwise have remained unused?
- Which costs are already included in another row and must not be counted twice?
The fourth question needs discipline. A four-hour interruption on a non-constrained line is not automatically four hours of lost contribution. A four-hour interruption on a constrained operation may displace feasible production, but the plant needs an approved finance basis and a documented event window. Treat capacity loss as an optional branch until those conditions are met.
This page uses COPQ as a practical umbrella for failure-related costs. It does not provide legal, financial, GAAP, or management-accounting advice. The correct local treatment depends on the plant’s costing rules and the purpose of the analysis.
Cost of poor quality calculation
A first COPQ figure should be small enough to check. Use one event ID, one time window, and one owner. Pull the values from records that already exist, then mark which rows are measured, estimated, or awaiting finance confirmation.
The worksheet below is illustrative. It is not plant data, a survey, an industry benchmark, or a universal formula.
| Worksheet row | Illustrative input | Calculation | Source record or owner |
|---|---|---|---|
| Scrap material | $2,400 | Approved material basis | Material or inventory record, finance |
| Rework labour | $756 | 18 hours x $42 | Labour or work-order record, production |
| Reinspection | $252 | 6 hours x $42 | Quality record, quality lead |
| Utilities and cleaning | $180 | Event estimate pending local rule | Utility or sanitation record, operations |
| Maintenance response | $255 | 3 hours x $85 | Maintenance work order, maintenance lead |
| Internal-failure subtotal | $3,843 | Sum of the rows above | Event owner validates duplicates |
| Optional constrained-capacity effect | $1,200 | 4 hours x approved $300 basis | Constraint record, finance |
| Illustrative event total | $5,043 | Subtotal plus approved capacity branch | Use only if non-duplicative |
The direct subtotal is:
$2,400 + $756 + $252 + $180 + $255 = $3,843
The optional capacity branch is:
$3,843 + $1,200 = $5,043
The arithmetic is easy. The boundary is the work. For each event, record:
- event ID, date, line, product, shift, and process step;
- defect or failure description and the first point it was detected;
- units or material affected, including scrap and rework disposition;
- first-pass result and final disposition, where those metrics exist;
- production, quality, maintenance, and finance record identifiers;
- capacity state, named constraint, and the approved basis for any opportunity-cost row;
- one accountable owner and the next decision date;
- rows excluded because they would duplicate another cost.
This is a scoping worksheet, not a promise that every row belongs in your official COPQ report. It gives the team enough structure to ask finance and operations the next precise question.
Why separate dashboards hide the causal chain
Separate dashboards hide the causal chain because each team measures a different consequence. The supplied CPAI interview gives a practical example: a slower changeover in one shift can create extra downtime, which can create startup scrap, which can create more rework labour. If the changeover owner, efficiency lead, quality lead, and labour report are read separately, no one report carries the full event.
The remedy is not automatically another dashboard. Start with a shared event ID and a short chain:
condition -> downtime or delay -> first-pass loss -> rework or scrap -> quality and maintenance effort -> capacity or customer consequence
Then ask whether the chain is supported by records. A clean chain is more useful than a large total with uncertain boundaries. If the plant has identified a recurring event, the Root Cause Analysis Workbook is a practical next step for assigning causes, evidence, actions, and owners.
For the wider measurement boundary, read how to measure the hidden cost of manufacturing waste and rework. For the first-pass distinction that often exposes a rework loop, see First Pass Yield vs Final Yield: Formula and Examples.
The operating rule is simple: define the defect boundary, join the records around one event, and let the approved finance basis decide which costs belong in the official total.
