10.4 Performance Measurement: Productivity, KPIs, Wage Systems, and the Balanced Scorecard

10.4 Performance Measurement: Productivity, KPIs, Wage Systems, and the Balanced Scorecard

The NCEES FE Industrial and Systems specification lists performance measurement — key performance indicators, productivity, wage scales, balanced scorecard, and customer satisfaction — as a distinct sub-topic of Engineering Management. It is the discipline that converts operational activity into numbers an executive can act on. Industrial engineers own these measures because a badly chosen metric silently rewards the wrong behavior: reward machine utilization and you will get overproduction; reward pieces per hour with no quality gate and you will get scrap.

---

1. Productivity Measurement

Productivity is a ratio of output produced to input consumed over a defined period:

$\text{Productivity} = \frac{\text{Output}}{\text{Input}}$

The measure is only meaningful when the period, the boundary, and the units are stated. Three families are distinguished by how many inputs appear in the denominator.

Partial-Factor Productivity

A single resource sits in the denominator, so output and input may be in physical units:

$\text{Labor Productivity} = \frac{\text{Units produced}}{\text{Direct labor hours}} \qquad \text{Machine Productivity} = \frac{\text{Units produced}}{\text{Machine hours}}$

$\text{Material Yield} = \frac{\text{Good output weight}}{\text{Raw material weight}} \qquad \text{Energy Productivity} = \frac{\text{Units produced}}{\text{kWh consumed}}$

Partial measures are easy to collect and easy to game. Labor productivity rises when a plant automates, even if the total cost per unit rises, because the substituted capital never appears in the denominator.

Multifactor and Total-Factor Productivity

When several inputs are combined, they must be converted to a common currency:

$\text{MFP} = \frac{\text{Value of output}}{\text{Labor \$} + \text{Material \$} + \text{Energy \$} + \text{Capital \$}}$

Total-factor productivity (TFP) extends the denominator to every purchased input. A dimensionless MFP or TFP above 1.0 means the operation generates more output value than it consumes in input value.

Productivity Index and Growth

Comparisons across periods use an index against a base period:

$\text{Productivity Index} = \frac{\text{Productivity}_{\text{current}}}{\text{Productivity}_{\text{base}}} \times 100 \qquad \%\,\text{Growth} = \frac{P_{\text{current}} - P_{\text{base}}}{P_{\text{base}}} \times 100\%$

Exam Watchout — Deflate Before Comparing: If output is measured in dollars, a 5% selling-price increase inflates measured productivity even though nothing physical improved. Multi-period productivity comparisons must use constant (real) dollars for both output and inputs, applying the same base-year price deflator used in the inflation analysis of engineering economics.

---

2. Utilization, Efficiency, Effectiveness, and OEE

These three words are not interchangeable, and the FE exam tests the distinction directly.

A press can run 95% of the shift (high utilization) at 60% of standard speed (poor efficiency) producing parts that fail final gauge (zero effectiveness).

Overall Equipment Effectiveness (OEE)

OEE is the standard composite equipment metric in lean and TPM programs. It multiplies three independent loss categories:

$\text{OEE} = A \times P \times Q$

$A = \frac{\text{Operating Time}}{\text{Planned Production Time}} \qquad P = \frac{\text{Ideal Cycle Time} \times \text{Total Pieces}}{\text{Operating Time}} \qquad Q = \frac{\text{Good Pieces}}{\text{Total Pieces}}$

• Availability ($A$) captures downtime losses: breakdowns, changeovers, starvation. Planned production time already excludes scheduled non-production such as breaks and planned maintenance.
• Performance ($P$) captures speed losses: minor stops, idling, running below rated rate. $P$ can never legitimately exceed 1.0; a value above 1.0 means the "ideal" cycle time is wrong.
• Quality ($Q$) captures defect losses: scrap and rework, including startup rejects.

A useful algebraic shortcut collapses all three:

$\text{OEE} = \frac{\text{Good Pieces} \times \text{Ideal Cycle Time}}{\text{Planned Production Time}}$

World-class discrete manufacturing OEE is commonly benchmarked near 85% ($A \approx 90%$, $P \approx 95%$, $Q \approx 99%$); typical plants land between 45% and 65%.

Two related composites appear in maintenance-heavy plants: TEEP (Total Effective Equipment Performance) multiplies OEE by the schedule-loading fraction to measure against all 168 calendar hours, while OOE uses total operating time rather than planned production time in the availability term.

3. Designing Key Performance Indicators (KPIs)

A key performance indicator is a quantified measure tied to a specific objective and used to trigger action. A metric that no one will act on is reporting overhead, not a KPI.

The Anatomy of a Defined KPI

Every KPI on a plant scorecard must specify seven attributes:

1. Name and objective it supports
2. Formula, written explicitly with numerator and denominator
3. Unit of measure
4. Data source and collection method (MES tag, ERP transaction, manual count)
5. Reporting frequency (per shift, daily, weekly, monthly)
6. Owner — a single named accountable person
7. Target and threshold for action

Leading vs. Lagging Indicators

• Lagging indicators report a completed outcome: recordable injury rate, scrap dollars, on-time delivery percentage, gross margin. They are accurate but arrive too late to change the result.
• Leading indicators predict the outcome and can still be influenced: near-miss reports filed, preventive maintenance compliance, first-pass yield at the upstream operation, training hours completed, supplier PPM at receiving.

A balanced operations review pairs each lagging outcome with the leading measures believed to drive it.

The SQDCP Families

Most manufacturing scorecards organize KPIs into five families, deliberately kept in this order so safety is discussed first:

Inventory turns, a recurring exam quantity, is defined as:

$\text{Inventory Turnover} = \frac{\text{Annual Cost of Goods Sold}}{\text{Average Inventory Value}} \qquad \text{Days of Supply} = \frac{365}{\text{Turns}}$

Common KPI Pathologies

• Vanity metrics: Impressive numbers with no decision attached (total pieces produced since 1998).
• Metric conflict: Rewarding departmental machine utilization while also demanding low WIP; the two cannot both be maximized.
• Denominator drift: Changing the population or period mid-year so the trend line becomes meaningless.
• Goodhart's law: "When a measure becomes a target, it ceases to be a good measure." Measuring only pieces per hour reliably produces pieces per hour — and scrap. Pair every rate metric with a quality gate.

4. Wage Scales, Job Evaluation, and Incentive Plans

Performance measurement and compensation are directly coupled: the wage plan determines whose behavior the measurement actually changes.

Establishing the Wage Scale: Job Evaluation

A wage scale is the structure of pay grades assigned to jobs. Four classical job-evaluation methods set it:

Point totals are plotted against market wage data to fit a wage curve; jobs above the line are red-circled (pay frozen until the grade catches up) and jobs below are green-circled (scheduled for increase).

Wage Payment Plans

Let $R$ = hourly base rate, $H$ = hours actually worked, $S$ = standard hours earned (pieces produced multiplied by standard time per piece), and time saved $= S - H$.

Two structural notes the exam probes:

• The guaranteed base rate. Modern piece-rate and standard-hour plans guarantee $H \times R$ regardless of output, so the incentive only ever adds to pay. The Taylor differential plan, which penalizes below-standard output, is largely of historical interest for exactly this reason.
• Rowan's self-limiting bonus. Because the bonus fraction is $(S - H)/S$, an operator who halves the standard time earns at most a 100% bonus and the marginal reward shrinks as efficiency climbs. Halsey's fixed 50% share does not self-limit, which is why loose standards are far more expensive under Halsey than under Rowan.

5. Customer Satisfaction Measurement

Internal efficiency metrics are meaningless if the customer is unhappy, so the specification pairs productivity and wage measures with customer satisfaction.

Perceptual Measures

• CSAT (Customer Satisfaction Score): Percentage of respondents selecting the top boxes on a satisfaction scale. Report the top-two-box percentage rather than a mean, because the underlying scale is ordinal.
• Net Promoter Score (NPS): From the 0-10 "how likely are you to recommend" question, classify 9-10 as Promoters, 7-8 as Passives, and 0-6 as Detractors: $\text{NPS} = \%\,\text{Promoters} - \%\,\text{Detractors}$ NPS ranges from $-100$ to $+100$ and is reported as an integer, not a percentage. Passives are counted in the denominator but never in the numerator.
• Customer Effort Score (CES): How much work the customer had to do to get resolution; a strong predictor of repurchase in service settings.
• SERVQUAL gap model: Measures the difference between customer expectation and perception across reliability, assurance, tangibles, empathy, and responsiveness.

Objective (Behavioral) Measures

Perceptual scores lag; objective service metrics can be measured every day:

• On-Time In-Full (OTIF): Fraction of orders delivered complete and on the promised date.
• Perfect Order Index: The product of the on-time rate, complete rate, damage-free rate, and correct-documentation rate. Because it multiplies, four separate 95% performances yield only $0.95^4 = 81.5%$ perfect orders — the same multiplicative erosion seen in series reliability.
• Warranty claim rate, return rate, and complaint rate per million units shipped.
• First-contact resolution rate and mean time to resolve.

---

6. The Balanced Scorecard

Robert Kaplan and David Norton introduced the Balanced Scorecard in 1992 to correct the dominance of purely financial reporting, which measures only past results and encourages short-term decisions. The scorecard translates strategy into objectives, measures, targets, and initiatives across four linked perspectives.

                     The Balanced Scorecard Causal Chain

   ┌──────────────────────────┐
   │  LEARNING AND GROWTH     │  Skills, information systems, culture
   │  "Can we keep improving?" │  KPIs: cross-training coverage, training
   └────────────┬─────────────┘  hours, employee engagement, turnover
                │ enables
                ▼
   ┌──────────────────────────┐
   │ INTERNAL BUSINESS PROCESS│  The processes we must excel at
   │  "What must we excel at?" │  KPIs: OEE, first-pass yield, cycle time,
   └────────────┬─────────────┘  schedule attainment, cost of poor quality
                │ drives
                ▼
   ┌──────────────────────────┐
   │        CUSTOMER          │  How customers see us
   │  "How do customers see us?"│ KPIs: OTIF, NPS, warranty rate, share
   └────────────┬─────────────┘
                │ produces
                ▼
   ┌──────────────────────────┐
   │        FINANCIAL         │  How we look to shareholders
   │  "How do we look to owners?"│ KPIs: ROI, EVA, revenue growth, margin
   └──────────────────────────┘

Key properties tested on the exam:

• The four perspectives are Financial, Customer, Internal Business Process, and Learning and Growth. "Competitor" and "Regulatory" are not scorecard perspectives.
• The perspectives are causally linked bottom-up, and this chain is drawn explicitly on a strategy map. Learning and Growth is the foundation; Financial is the outcome.
• The scorecard deliberately balances lagging financial outcomes against leading operational and human-capital drivers, and balances external (financial, customer) against internal (process, learning) views.
• Each objective carries a measure, a target, and a funded initiative; a scorecard without initiatives is a report, not a management system.

Benchmarking

Targets are set credibly by benchmarking against a reference: internal (best shift or sister plant), competitive (direct rivals), functional (best-in-class at the same function in another industry — the classic case being a hospital studying an aircraft pit crew for patient handoffs), or generic (universal processes such as order entry).

7. Step-by-Step Worked Engineering Calculations

Worked Example 10.4.1: Multifactor Productivity and Productivity Growth

Problem: A machining plant reports the following weekly figures. Output is valued at $50.00 per unit and all dollar figures are already expressed in constant base-year dollars.

1. Compute baseline labor productivity and baseline multifactor productivity.
2. Compute both measures after the Kaizen event.
3. Compute the percentage growth in each, and explain the divergence.

Solution:

Step 1: Baseline

• Labor cost: $1{,}600 \times $28.00 = $44{,}800$
• Output value: $12{,}000 \times $50.00 = $600{,}000$
• Total input value: $$44{,}800 + $310{,}000 + $18{,}000 + $52{,}000 = $424{,}800$ $\text{Labor Productivity}_{\text{base}} = \frac{12{,}000\text{ units}}{1{,}600\text{ hr}} = 7.50\text{ units/labor hour}$ $\text{MFP}_{\text{base}} = \frac{\$600{,}000}{\$424{,}800} = 1.4124$

Step 2: After Kaizen

• Output value: $13{,}200 \times $50.00 = $660{,}000$
• Total input value: $$44{,}800 + $335{,}000 + $18{,}500 + $52{,}000 = $450{,}300$ $\text{Labor Productivity}_{\text{new}} = \frac{13{,}200}{1{,}600} = 8.25\text{ units/labor hour}$ $\text{MFP}_{\text{new}} = \frac{\$660{,}000}{\$450{,}300} = 1.4657$

Step 3: Growth $\%\,\Delta\text{Labor Productivity} = \frac{8.25 - 7.50}{7.50} \times 100\% = 10.0\%$ $\%\,\Delta\text{MFP} = \frac{1.4657 - 1.4124}{1.4124} \times 100\% = 3.77\%$

Engineering interpretation: Labor productivity jumped 10% because the same 1,600 hours produced 1,200 more units. Multifactor productivity rose only 3.8% because material consumption grew nearly in proportion to output ($+8.1%$), so most of the extra output was bought, not earned. Reporting only the labor number would overstate the improvement by a factor of roughly 2.7. This is exactly why capital-substitution projects must be judged on MFP.

---

Worked Example 10.4.2: Overall Equipment Effectiveness

Problem: An injection molding press is scheduled for a 480-minute shift. Contractual breaks and a planned mold-preventive-maintenance window total 45 minutes and are excluded from planned production time. During the shift the press suffered 55 minutes of unplanned downtime (a hydraulic fault and a material starvation event). It produced 620 total shots at an ideal cycle time of 0.50 minutes per shot; 595 shots passed final inspection.

Compute Availability, Performance, Quality, and OEE, and identify the dominant loss.

Solution:

Step 1: Availability $\text{Planned Production Time} = 480 - 45 = 435\text{ minutes}$ $\text{Operating Time} = 435 - 55 = 380\text{ minutes}$ $A = \frac{380}{435} = 0.8736 \quad (87.36\%)$

Step 2: Performance $P = \frac{\text{Ideal Cycle Time} \times \text{Total Pieces}}{\text{Operating Time}} = \frac{0.50 \times 620}{380} = \frac{310}{380} = 0.8158 \quad (81.58\%)$

Step 3: Quality $Q = \frac{595}{620} = 0.9597 \quad (95.97\%)$

Step 4: OEE $\text{OEE} = A \times P \times Q = 0.8736 \times 0.8158 \times 0.9597 = 0.6839 \quad (68.4\%)$ Verification using the shortcut: $\text{OEE} = \frac{595 \times 0.50}{435} = \frac{297.5}{435} = 0.6839 \quad \checkmark$

Step 5: Loss diagnosis: The largest single loss is Performance at 81.6% — the press ran 380 minutes but delivered only 310 minutes of ideal-cycle output, meaning 70 minutes were consumed by minor stops and slow cycles. Availability losses cost 55 minutes and quality losses cost 25 shots. Because minor stops are invisible on a downtime log, the corrective action is short-interval cycle logging at the press, not another maintenance program.

---

Worked Example 10.4.3: Comparing Wage Incentive Plans

Problem: The standard time for a bracket weldment is 12.0 minutes per piece, so the standard output is 5 pieces per hour. The guaranteed base rate is $R = $24.00$ per hour. In an 8-hour shift ($H = 8$), an operator completes 46 acceptable pieces; the shift standard is 40 pieces.

Compute the operator's shift earnings and the employer's direct labor cost per piece under (a) day work, (b) straight piece rate, (c) the standard hour plan, (d) Halsey 50-50, and (e) the Rowan plan.

Solution:

Step 0: Establish standard hours earned $S = 46\text{ pieces} \times \frac{12.0\text{ min}}{60\text{ min/hr}} = 46 \times 0.200 = 9.20\text{ standard hours}$ $\text{Time saved} = S - H = 9.20 - 8.00 = 1.20\text{ hours} \qquad \text{Efficiency} = \frac{9.20}{8.00} = 115\%$

(a) Day Work $E = H \times R = 8 \times \$24.00 = \$192.00 \qquad \text{Labor cost/piece} = \frac{192.00}{46} = \$4.174$

(b) Straight Piece Rate Piece price $= R \times$ standard time per piece $= $24.00 \times 0.200 = $4.80$ per piece. $E = 46 \times \$4.80 = \$220.80 \qquad \text{Labor cost/piece} = \$4.80$

(c) Standard Hour Plan $E = S \times R = 9.20 \times \$24.00 = \$220.80 \qquad \text{Labor cost/piece} = \$4.80$ Identical to the straight piece rate — the two plans are algebraically the same whenever the piece price is derived from the base rate and the standard time.

(d) Halsey 50-50 $E = H R + 0.5 (S - H) R = 192.00 + 0.5(1.20)(24.00) = 192.00 + 14.40 = \$206.40$ $\text{Labor cost/piece} = \frac{206.40}{46} = \$4.487$

(e) Rowan $E = H R + \left(\frac{S - H}{S}\right) H R = 192.00 + \left(\frac{1.20}{9.20}\right)(192.00) = 192.00 + 25.04 = \$217.04$ $\text{Labor cost/piece} = \frac{217.04}{46} = \$4.718$

Comparison and interpretation:

At 115% efficiency the plans rank Halsey < Rowan < piece rate in operator earnings, and the ranking of employer unit labor cost is the mirror image. Day work shows the lowest unit labor cost here only because the operator produced 15% above standard without additional pay — a condition that does not persist, which is the entire rationale for incentive plans. Note also that at very high efficiency the Rowan bonus fraction $(S-H)/S$ approaches 1.0 and flattens, capping the employer's exposure to a loose standard, whereas Halsey's fixed 50% share and the straight piece rate do not self-limit.