7. Productivity Measurement and Performance Systems

Introduction and Context

Productivity is one of the central analytical concepts in production management. It expresses the relationship between the resources consumed by an operation and the goods or services produced from those resources. A productivity measure is therefore more than a ratio: it is a way of interpreting how effectively an organization converts labour, materials, capital, energy, time, technology, and managerial capability into value.

For a manufacturing organization, productivity information supports decisions about capacity, investment, staffing, process improvement, sourcing, automation, and cost control. However, productivity measurement can also mislead. A department may increase output per employee by building excessive inventory, extending working hours, reducing maintenance, or transferring work to another department. A high utilization rate may coexist with poor customer service, excessive defects, or unsafe work practices. Consequently, productivity must be measured as part of a broader performance system rather than treated as a single target.

This chapter develops the logic and practice of productivity measurement. It distinguishes partial, multifactor, and total-factor productivity; explains the related ideas of efficiency, effectiveness, and utilization; examines normalization and benchmarking; and presents a framework for balanced key performance indicators and productivity dashboards. The discussion is particularly relevant to production managers who must convert operational data into economically meaningful and behaviourally sound decisions.

Learning Objectives

After studying this chapter, students should be able to:

1. Define productivity and distinguish it from efficiency, effectiveness, and utilization.

2. Calculate and interpret partial, multifactor, and total-factor productivity.

3. Identify the effects of price, product mix, quality, capacity, and boundary decisions on productivity measures.

4. Normalize productivity data for meaningful comparison across periods, products, plants, and processes.

5. Design balanced productivity KPIs, benchmarks, and management dashboards.

6. Diagnose misleading productivity results and recommend corrective action.

7. Evaluate digital and analytical tools used in productivity management.

Productivity as a Conversion Relationship

At its simplest, productivity is expressed as:

Productivity = Output / Input

Output may be measured in physical units, equivalent units, standard hours, revenue, contribution margin, or customer value. Inputs may include labour hours, machine hours, material quantity, energy, capital employed, floor space, or a combination of resources. The selected units determine what the measure actually means.

For example, if a line produces 4,800 acceptable units in 1,200 direct labour hours, labour productivity is 4 units per labour hour. If the same line produces the output using 1,000 hours in the next month, labour productivity becomes 4.8 units per hour, a 20 percent improvement. This conclusion is valid only if product complexity, quality requirements, overtime, rework, and the definition of labour hours remain comparable.

Productivity is different from profitability. Profitability compares financial returns with financial costs, whereas productivity examines the physical or value relationship between output and input. Productivity improvement may increase profitability, but not necessarily. Producing more units per hour can reduce profitability if the additional units are unwanted, defective, or sold at an inadequate margin. Conversely, a product redesign may reduce physical productivity while increasing customer value and profit.

Important Terminology

Output is the quantity or value generated by a process. Input is a resource consumed by that process. Productivity growth is the percentage change in productivity between two periods. Productivity level is productivity measured at a particular time or under specified conditions. Driver is a factor that influences the performance result, such as changeover time, absenteeism, or material yield. Productivity loss is the gap between potential or standard performance and actual performance.

The measurement boundary must be established before calculation. A boundary may cover one machine, a production cell, a department, a plant, or an entire supply chain. If a plant excludes maintenance labour but another plant includes it, their labour productivity figures are not directly comparable. Good measurement therefore begins with a clear statement of the process, period, output definition, input definition, data source, and exclusions.

Four Related Ideas: Efficiency, Effectiveness, Utilization, and Productivity

These terms are often used interchangeably in organizations, although they describe different questions.

Efficiency concerns the relationship between actual resource consumption and a standard, plan, or technically feasible level. A process is efficient when it produces the required output with relatively little waste. Labour efficiency, for example, may compare standard hours earned with actual hours worked.

Effectiveness concerns whether the intended result has been achieved. A production schedule can be completed efficiently but be ineffective if it produces the wrong product, misses the customer delivery date, or fails quality requirements.

Utilization concerns the extent to which available capacity or resources are being used. A machine operating for 70 hours out of 100 available hours has 70 percent time utilization. High utilization does not necessarily mean high productivity. The machine may be producing slowly, making defective units, or manufacturing products that are not needed.

Productivity combines an output perspective with an input perspective. It asks how much acceptable and relevant output is obtained from the resources consumed. A useful performance interpretation often requires all four concepts. For instance, a plant may have high machine utilization but low effectiveness because it produces late orders; acceptable effectiveness but poor efficiency because it uses excessive labour; or rising labour productivity but falling total-factor productivity because automation has greatly increased capital and energy costs.

Levels and Methods of Productivity Measurement

Partial Productivity

Partial productivity relates one input to output. Common measures include:

Labour productivity = Output / Labour hours or labour cost

Material productivity = Output / Material quantity or material cost

Machine productivity = Output / Machine hours

Energy productivity = Output / Energy consumed

Floor-space productivity = Output / Production area

Partial measures are simple, transparent, and useful for operational control. A supervisor can act on labour hours, scrap, downtime, or energy consumption. Their weakness is that they may shift responsibility from one input to another. Labour productivity may rise because more machines, subcontracting, overtime, or material costs have been introduced. It is therefore dangerous to interpret a partial measure as an overall improvement without examining complementary inputs.

Suppose a plant produces 10,000 acceptable units using 2,000 labour hours. Labour productivity is 5 units per hour. After automation, output rises to 12,000 units and labour hours fall to 1,600, giving 7.5 units per hour. This appears excellent. Yet if machine costs, maintenance, energy, and software support increase substantially, the overall economic result may be less impressive. The labour measure remains useful, but it does not provide the complete answer.

Multifactor Productivity

Multifactor productivity combines two or more input categories while excluding some others. A common form is:

Multifactor productivity = Output / (Labour input + Material input + Energy input)

Inputs must be expressed in comparable units, usually monetary values. For example, if monthly output valued at standard prices is $500,000 and labour, materials, and energy cost $100,000, $240,000, and $40,000 respectively, multifactor productivity is:

500,000 / (100,000 + 240,000 + 40,000) = 1.316

The result means that each monetary unit of the selected inputs corresponds to 1.316 monetary units of standardized output. The number itself has meaning mainly in comparison with another period or benchmark. If output is measured at actual selling prices while inputs are measured at accounting cost, changes in prices or product mix may be incorrectly interpreted as operational productivity change.

Multifactor productivity is particularly helpful when management is evaluating a process improvement involving several resources. It can reveal whether labour savings have been purchased through excessive material, energy, or service consumption.

Total-Factor Productivity

Total-factor productivity attempts to relate output to all major productive inputs, commonly labour, capital, materials, energy, purchased services, and sometimes land or technology. A general expression is:

TFP = Real output index / Composite input index

The input index is constructed by aggregating different inputs using suitable weights, often their cost shares. Because capital services and intangible capabilities are difficult to measure, total-factor productivity requires careful design and is more appropriate for strategic, plant-level, industry, or research analysis than for minute-by-minute control.

TFP growth is often interpreted as improvement in the efficiency of the combined production system after accounting for measured input growth. It may reflect technology, learning, better organization, improved quality of management, scale effects, or measurement error. It should not automatically be attributed to employee effort alone.

Measurement Choices That Shape the Result

Physical and Value-Based Output

Physical output is attractive when products are homogeneous. In a mixed-product plant, however, 1,000 small components and 1,000 large assemblies are not equivalent. Organizations may use standard hours, equivalent units, weighted output points, or value-added measures. Value-based output is flexible but sensitive to inflation, exchange rates, price changes, and product mix.

A robust system separates operational volume from price effects. Real output may be calculated using constant or standard prices, while nominal revenue is used for financial analysis. The choice should be documented and applied consistently.

Quality-Adjusted Output

Output should normally mean good output, not gross production. A useful measure is:

Good output = Total output − Scrap − Rework − Rejected units

A process producing 1,000 units with a 10 percent defect rate has only 900 acceptable units. If productivity is calculated using 1,000 units, the measure rewards poor quality. Some organizations use first-pass yield, rolled throughput yield, or good units per total input to make quality visible.

Time and Capacity

The denominator may be scheduled hours, paid hours, attendance hours, productive hours, or machine-running hours. Each answers a different question. A plant manager may use paid-hour productivity for workforce planning, while a maintenance team may use running-hour productivity for equipment analysis.

Capacity utilization can be expressed as:

Capacity utilization = Actual output / Practical capacity × 100

Practical capacity allows for planned maintenance, breaks, changeovers, and normal operating conditions. Theoretical capacity assumes uninterrupted operation and is usually unsuitable for performance evaluation. A comparison based on theoretical capacity can make a well-managed plant appear inefficient.

Cost and Price Effects

Productivity can appear to improve because output prices rise, input prices fall, or currency values change. To isolate operational improvement, analysts may use constant prices, deflated values, standard costs, purchasing-power adjustments, or physical measures. Standard costs must themselves be reviewed because outdated standards can create artificial variances.

Normalization and Comparability

Raw productivity figures are rarely comparable without adjustment. Normalization is the process of transforming data so that differences caused by scale, product, time, environment, or measurement conventions do not obscure the underlying performance relationship.

Common normalization approaches include:

• Index numbers: Set a base period to 100 and express later performance relative to it. An index of 112 indicates a 12 percent increase from the base.
• Constant-price valuation: Value different-period output using a common price basis.
• Standard-hour conversion: Convert products of different complexity into the labour or machine hours normally required.
• Equivalent-unit conversion: Express partially completed or varied products in a common output unit.
• Rate normalization: Use output per labour hour, machine hour, or unit of capacity rather than total output.
• Mix adjustment: Recalculate output as if the product or service mix had remained constant.
• Seasonal adjustment: Separate predictable seasonal effects from unusual performance.
• Risk and quality adjustment: Account for defects, warranty exposure, safety incidents, or process instability.

For example, a plant makes Product A, requiring two standard hours per unit, and Product B, requiring five standard hours. In a month it produces 1,000 A units and 400 B units. Equivalent output is 2,000 + 2,000 = 4,000 standard hours. If 1,000 direct hours were used, standardized labour productivity is 4 standard output hours per actual labour hour. This is more meaningful than simply reporting 1,400 units.

Normalization must not become an exercise in manipulating results. Every adjustment should be theoretically justified, transparent, reproducible, and reviewed by relevant stakeholders. Excessive adjustment can hide genuine operational problems.

Benchmarks and Targets

A benchmark is a reference point against which performance is assessed. It may be historical, planned, internal, competitive, industry-based, or based on a technically designed standard. Each type has advantages and risks.

Historical benchmarks show improvement over time but may preserve past inefficiencies. Budget benchmarks support planning but can become politically negotiated. Internal benchmarks encourage learning between comparable lines, although differences in product mix and equipment must be controlled. External benchmarks can reveal strategic gaps but may use different accounting boundaries or data definitions. Theoretical best performance is useful for diagnosing losses but may be unrealistic as an immediate target.

A sound target has a defined metric, time horizon, owner, baseline, data source, and improvement logic. Targets should combine results and drivers. For example, a target of reducing labour hours per unit should be accompanied by limits for defects, overtime, safety incidents, and customer lateness. Otherwise, employees may achieve the target by sacrificing important outcomes.

Designing a Balanced Productivity Performance System

A productivity system should translate organizational priorities into a connected set of measures. The system begins with strategic questions: What value does the customer expect? Which resources constrain performance? What risks must not be traded away? Which processes create or destroy value?

A practical structure includes four layers.

Outcome indicators measure results such as cost per acceptable unit, on-time delivery, contribution per constraint hour, and total-factor productivity.

Process indicators monitor the mechanisms producing the results, such as cycle time, changeover time, first-pass yield, downtime, schedule adherence, and material yield.

Resource indicators track labour hours, energy intensity, inventory, machine capacity, absenteeism, and maintenance expenditure.

Guardrail indicators prevent harmful optimization. They include safety incidents, ergonomic exposure, quality escapes, customer complaints, overtime, environmental emissions, and employee turnover.

A balanced scorecard for a manufacturing plant may therefore combine financial productivity, customer service, internal process capability, and learning or innovation. The exact framework can vary, but no single dimension should dominate all management attention.

KPI Design Principles

A useful KPI is clearly defined, decision-relevant, controllable or influenceable by its owner, timely, stable enough for comparison, and difficult to game. The KPI documentation should state its formula, numerator, denominator, data source, frequency, responsible owner, target, tolerance, and escalation rule.

Leading indicators signal future performance. Examples include preventive-maintenance compliance, training hours, changeover preparation, and schedule stability. Lagging indicators record achieved outcomes, such as cost per unit, customer returns, and monthly productivity. Effective systems use both.

Measures should be reviewed as a portfolio. If a KPI produces no decision, it may be unnecessary. If several KPIs measure the same phenomenon, the system may become bureaucratic. If a target conflicts with another target, management must resolve the conflict rather than leave operators to choose.

Productivity Dashboards and Data Architecture

A dashboard is not merely a collection of charts. It is an information system designed to support a recurring management conversation. A good dashboard presents the current result, trend, target, variance, significance, and likely driver. It should allow users to move from plant-level performance to line, product, shift, and loss category without losing definitions.

At the operational level, visual management may show hourly output, downtime, defects, and schedule adherence. At the tactical level, a weekly dashboard may show labour productivity, material yield, overtime, changeover performance, and maintenance losses. At the strategic level, management may examine total-factor productivity, cost competitiveness, capital productivity, customer value, and improvement-project benefits.

Digital manufacturing execution systems, enterprise resource planning systems, machine sensors, barcode systems, and energy meters can automate data collection. Analytics can identify correlations between downtime and conditions such as product family, shift, temperature, or maintenance status. Machine learning may support anomaly detection and predictive maintenance. Nevertheless, automation does not correct poor definitions or inaccurate master data. A highly sophisticated dashboard built on inconsistent product codes and unreliable downtime categories can increase confidence without increasing truth.

Data governance should address ownership, access rights, validation, version control, audit trails, cybersecurity, and privacy. Where individual employee data are used, organizations must distinguish process improvement from surveillance and comply with applicable labour and data-protection requirements.

Illustrative Case: Improving Productivity at Meridian Components

Background

Meridian Components, a fictional manufacturer of precision housings, operated three machining cells. Management reported that labour productivity had improved by 14 percent over six months. At the same time, customer complaints, overtime, and work-in-process inventory were increasing.

Situation and Problem

The reported measure was total pieces produced divided by direct labour hours. It counted rework as output and excluded material handlers, maintenance technicians, and temporary labour. Product mix also changed: the plant was producing more of a small, simple housing and fewer complex housings. Supervisors were rewarded for hourly output, so operators sometimes ran large batches before changeovers and moved unfinished work to the next process.

Analysis

A cross-functional team reconstructed the measurement boundary. It included all labour assigned to the cells, counted only accepted first-pass units, and converted products into standard machining hours. It also examined machine availability, first-pass yield, changeover time, overtime, and work-in-process.

The revised results showed that reported labour productivity had risen 14 percent, but standardized good-output productivity had risen only 3 percent. First-pass yield had fallen from 96 to 92 percent, and overtime had increased by 11 percent. The apparent improvement was largely explained by product mix, rework counting, and labour exclusion. Material productivity had also declined because of increased scrap.

Approach and Solution

Meridian replaced the single output-per-worker KPI with a balanced cell dashboard containing:

• good standard hours per total labour hour;
• first-pass yield;
• changeover time and schedule adherence;
• material yield;
• machine availability and performance rate;
• overtime and safety incidents;
• work-in-process days.

The company introduced common definitions, daily data validation, and weekly review by production, quality, maintenance, and planning managers. It applied a product-mix adjustment and separated planned changeover time from unplanned downtime. A small improvement team redesigned changeover preparation and investigated the main sources of rework.

Outcome

After three months, the plant reduced changeover time and rework while maintaining customer delivery performance. The immediate labour productivity increase was less dramatic than the original report, but the revised measure was more credible and supported better decisions. Management also recognized that productivity improvement required coordination across production, quality, maintenance, and planning rather than pressure on operators alone.

Learning Points

The case demonstrates that a productivity number is only as reliable as its boundary, definitions, and incentives. Good-output measures are superior to gross-output measures when quality varies. Product mix must be normalized before comparison. Excluding supporting resources creates distorted partial productivity. Finally, a balanced system prevents managers from improving one ratio by damaging the wider operation.

Challenges, Risks, and Common Mistakes

The most common error is confusing activity with productivity. More production hours, machine utilization, or employee effort do not automatically create more value. Other frequent mistakes include changing the denominator without documenting it, using inconsistent cost valuations, ignoring outsourced work, counting rework as output, comparing incomparable products, and setting targets without considering capacity constraints.

Productivity measures can also create dysfunctional behaviour. Employees may rush work, hide downtime, defer maintenance, build inventory, or avoid difficult products. Individual rankings may discourage cooperation and lead to under-reporting of problems. Measures based on estimated standards can become contentious when standards are unrealistic or controlled by a department distant from the process.

Managers should interpret variation statistically rather than reacting to every daily fluctuation. Control charts, moving averages, confidence intervals, and stratification by product and shift can distinguish common-cause variation from a special event. A productivity decline should lead to questions about demand, mix, quality, staffing, equipment condition, materials, and measurement changes before blame is assigned.

Ethical and Professional Considerations

Productivity management has a human dimension. Targets should not encourage unsafe speeds, excessive fatigue, discriminatory evaluation, or concealment of defects. Employee participation improves measurement quality because operators understand practical losses that may not appear in databases. Performance information should be used for learning and improvement before punitive action, except where misconduct or deliberate falsification is established.

Professional reporting requires transparency. Managers should disclose assumptions, exclusions, data limitations, and changes in definitions. Productivity gains arising from layoffs, outsourcing, or reduced quality should not be represented as unqualified operational excellence. Environmental impacts and resource depletion should increasingly be included in productivity discussions, particularly where energy, water, emissions, or hazardous materials are significant.

Current Developments and Best Practices

Manufacturing organizations are increasingly combining productivity measurement with real-time production data, digital twins, energy monitoring, advanced scheduling, and process-mining techniques. Artificial intelligence can help classify downtime, forecast bottlenecks, identify abnormal consumption, and estimate the likely effect of improvement actions. Human review remains essential because models may reproduce biased historical decisions or mistake correlation for causation.

Best practice is to begin with a small number of well-defined measures, validate them at the process, and expand only when the data are trusted. Organizations should:

1. Define the measurement boundary and purpose before selecting a formula.

2. Use acceptable, quality-adjusted output wherever possible.

3. Normalize for product mix, complexity, price, season, and capacity.

4. Combine partial productivity with multifactor or total-factor perspectives.

5. Pair outcome KPIs with process drivers and guardrails.

6. Establish common definitions and auditable data ownership.

7. Review incentives for unintended consequences.

8. Use benchmarks as learning references rather than simplistic rankings.

9. Involve operators and support functions in interpretation.

10. Retire measures that no longer support a decision.

Chapter Summary

Productivity measures how effectively resources are converted into relevant output. Partial productivity examines one input, multifactor productivity combines selected inputs, and total-factor productivity considers the broader input system. Each level serves a different managerial purpose, and none should be interpreted without a clear boundary and reliable definitions.

Efficiency concerns resource use relative to a standard; effectiveness concerns achievement of intended outcomes; utilization concerns the use of available capacity. Productivity analysis must account for quality, product mix, capacity, price, labour boundaries, and rework. Normalization methods such as index numbers, constant prices, standard hours, equivalent units, and mix adjustment make comparisons more meaningful.

A strong performance system uses balanced KPIs that combine outcome, process, resource, and guardrail measures. Dashboards should support decisions, not merely display data. Digital tools can improve timeliness and diagnosis, but technology cannot compensate for weak definitions or poor governance. Ethical productivity management protects safety, quality, fairness, and long-term organizational capability.

Key Terms

Productivity: The relationship between relevant output and resources consumed.

Partial productivity: Output divided by one input category.

Multifactor productivity: Output divided by a combination of selected inputs.

Total-factor productivity: Output relative to a composite of major productive inputs.

Efficiency: Resource performance compared with a standard or feasible level.

Effectiveness: The extent to which intended objectives are achieved.

Utilization: The proportion of available capacity or resource time being used.

Good output: Output that satisfies defined quality and acceptance requirements.

Product mix: The relative combination of products produced during a period.

Normalization: Adjustment of data to improve comparability.

Standard hour: An equivalent output unit based on predetermined standard time.

Capacity utilization: Actual output relative to practical or available capacity.

Benchmark: A reference value used for performance comparison.

Key performance indicator: A defined measure used to monitor an important result or driver.

Leading indicator: A measure that signals conditions likely to influence future results.

Lagging indicator: A measure of performance already achieved.

Guardrail metric: A measure that prevents improvement in one area from harming another.

Productivity dashboard: A visual information system that presents performance, trends, targets, and drivers.

Data governance: Policies and controls ensuring data quality, ownership, security, and appropriate use.

Review Questions

Short Answer Questions

1. Define partial productivity and give two manufacturing examples.

2. Why is utilization not the same as productivity?

3. What is meant by good output?

4. State two reasons why productivity data may require normalization.

5. Distinguish a leading indicator from a lagging indicator.

Long Answer Questions

1. Explain partial, multifactor, and total-factor productivity, including the managerial purpose and limitations of each.

2. Discuss how quality, product mix, price changes, and capacity assumptions influence productivity measurement.

3. Explain the relationship between efficiency, effectiveness, utilization, and productivity with suitable examples.

4. Design a balanced productivity measurement system for a machining department.

5. Discuss the role of dashboards, digital manufacturing systems, and data governance in productivity management.

Analytical / Application Questions

1. A plant produces 6,000 accepted units using 1,500 labour hours. After an improvement project, it produces 6,600 total units, of which 5 percent require rework, using 1,200 labour hours. Calculate labour productivity on an accepted-output basis for both periods and interpret the result.

2. A factory reports higher labour productivity after outsourcing material handling and maintenance. Analyze why the result may be misleading and identify the additional inputs or boundary changes that should be examined.

3. Develop five KPIs for a production cell that has high machine utilization but poor delivery performance and rising defects. Explain how each KPI would prevent dysfunctional optimization.

22. Multiple Choice Questions with Answers

Question 1

Which measure best represents output after accounting for quality losses?

A. Total units produced divided by scheduled hours

B. Accepted units divided by relevant input

C. Revenue divided by theoretical capacity

D. Machine hours divided by total units

Correct Answer: B. Accepted units divided by relevant input

Question 2

A plant's labour productivity rises after automation, but capital and energy inputs rise substantially. Which measure is most useful for examining the combined effect?

A. Partial labour productivity

B. Machine utilization

C. Multifactor productivity

D. Schedule adherence

Correct Answer: C. Multifactor productivity

Question 3

Which statement best distinguishes effectiveness from efficiency?

A. Effectiveness concerns achieving the intended result, while efficiency concerns resource use

B. Effectiveness measures capacity use, while efficiency measures customer satisfaction

C. Effectiveness always means low cost, while efficiency means high output

D. There is no practical distinction between the terms

Correct Answer: A. Effectiveness concerns achieving the intended result, while efficiency concerns resource use

Question 4

A plant produces two products requiring two and five standard hours respectively. Which method creates a more comparable output measure?

A. Counting both products as one physical unit each

B. Using only the selling price of the products

C. Converting output into standard-hour equivalents

D. Dividing output by theoretical machine capacity

Correct Answer: C. Converting output into standard-hour equivalents

Question 5

Why can high machine utilization coexist with poor productivity?

A. Utilization includes neither machine time nor output

B. The machine may run continuously while producing defects or unwanted inventory

C. High utilization automatically reduces all input costs

D. Productivity does not include output

Correct Answer: B. The machine may run continuously while producing defects or unwanted inventory

Question 6

Which is a guardrail KPI in a balanced productivity system?

A. Units per labour hour

B. Changeover preparation time

C. First-pass yield

D. Safety incidents

Correct Answer: D. Safety incidents

Question 7

A historical productivity benchmark is potentially misleading because it may:

A. Use any form of comparison

B. Preserve inefficiencies already present in the process

C. Require a data source

D. Be expressed as an index

Correct Answer: B. Preserve inefficiencies already present in the process

Question 8

What is the main purpose of constant-price valuation?

A. To increase reported revenue

B. To remove all quality defects from output

C. To reduce the effect of price changes on output comparison

D. To measure employee satisfaction

Correct Answer: C. To reduce the effect of price changes on output comparison

Question 9

Which dashboard design practice is most appropriate?

A. Display as many measures as possible without definitions

B. Show results, trends, targets, and likely drivers linked to decisions

C. Use only financial indicators

D. Rank individual employees using one productivity ratio

Correct Answer: B. Show results, trends, targets, and likely drivers linked to decisions

Question 10

In the Meridian Components case, the reported productivity improvement was overstated mainly because the measure:

A. Included accepted output only and all support labour

B. Adjusted perfectly for product mix

C. Counted rework, excluded supporting labour, and ignored product-mix effects

D. Used total-factor productivity

Correct Answer: C. Counted rework, excluded supporting labour, and ignored product-mix effects