Study guide
10.3 Learning Curve Theory (Wright Cumulative vs. Crawford Unit Models)
AACE Certified Cost Professional CCP
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: Introduction & Total Cost Management (TCM) Framework
1.1 AACE CCP Exam Structure, Eligibility & the In-Exam Memo 1.2 The Total Cost Management (TCM) Framework & PDCA Cycle 1.3 Strategic Asset Management vs. Project Control Processes
2 Chapter 2: Cost Estimating Fundamentals & Classification
2.1 AACE Recommended Practice 18R-97 Estimate Classification System 2.2 Direct vs. Indirect Costs and Cost Element Breakdown 2.3 Quantity Takeoff, Standard Cost Databases & Pricing Sources
3 Chapter 3: Conceptual & Parametric Estimating Techniques
3.1 Capacity Factoring & the Six-Tenths Rule 3.2 Parametric Estimating, Lang Factors & Hand Factors 3.3 Cost Indices, Location Adjustments & Escalation Modeling
4 Chapter 4: Cost Engineering Economics & Time Value of Money
4.1 Simple & Compound Interest, Equivalence & Cash Flow Diagrams 4.2 Discrete Compounding Formulas & Uniform Series Factors 4.3 Nominal vs. Effective Interest Rates & Continuous Compounding
5 Chapter 5: Project Profitability & Capital Investment Decision-Making
5.1 Net Present Value (NPV) & Present Worth Analysis 5.2 Internal Rate of Return (IRR) & Incremental IRR Analysis 5.3 Payback Period, Benefit-Cost Ratio (BCR) & Breakeven Analysis
6 Chapter 6: Depreciation Methods, Taxes & Asset Valuation
6.1 Classical Depreciation Methods (Straight-Line, DDB, Sum-of-Years-Digits) 6.2 Modified Accelerated Cost Recovery System (MACRS) & Tax Shield 6.3 After-Tax Cash Flow Analysis & Asset Replacement Economics
7 Chapter 7: Project Scope, Work Breakdown Structure (WBS) & Baselines
7.1 Work Breakdown Structure (WBS) Principles & the 100% Rule 7.2 Organizational Breakdown Structure (OBS) & Control Accounts 7.3 Establishing the Performance Measurement Baseline (PMB)
8 Chapter 8: Critical Path Method (CPM) Planning & Scheduling Logic
8.1 Precedence Diagramming Method (PDM) & Dependency Relationships 8.2 Forward Pass, Backward Pass & Float Calculations (Total vs. Free Float) 8.3 Schedule Baseline Development, Calendar Constraints & Milestones
9 Chapter 9: Schedule Optimization, Compression & Probabilistic PERT
9.1 Schedule Crashing & Least-Cost Slope Optimization 9.2 Fast-Tracking Risks, Overlap Analysis & Execution Strategies 9.3 PERT Three-Point Estimating & Probabilistic Duration Variance
10 Chapter 10: Resource Management, Productivity & Learning Curves
10.1 Resource Loading, Leveling vs. Resource Smoothing 10.2 Labor Productivity Factors, Work Sampling & Efficiency Loss 10.3 Learning Curve Theory (Wright Cumulative vs. Crawford Unit Models)
11 Chapter 11: Earned Value Management (EVM) Core Principles & Metrics
11.1 ANSI/EIA-748 EVM Framework & Core Metrics (PV, EV, AC) 11.2 EVM Variances & Performance Indices (CV, SV, CPI, SPI) 11.3 Work Progress Measurement Techniques (0/100, 50/50, Milestone Weighting)
12 Chapter 12: EVM Forecasting, Indices & Baseline Change Control
12.1 Estimate at Completion (EAC) & Estimate to Complete (ETC) Formulas 12.2 To-Complete Performance Index (TCPI) & Variance at Completion (VAC) 12.3 Integrated Change Control, Baseline Revisions & Trend Management
13 Chapter 13: Risk Management Planning & Qualitative Risk Analysis
13.1 Risk Management Framework & Risk Breakdown Structure (RBS) 13.2 Qualitative Risk Assessment & Probability-Impact (P-I) Matrix 13.3 Risk Response Strategies & Threat/Opportunity Management
14 Chapter 14: Quantitative Risk Analysis, Monte Carlo & Contingency Estimating
14.1 Expected Monetary Value (EMV) & Decision Tree Analysis 14.2 Monte Carlo Simulation & Probability Distributions in Cost/Schedule 14.3 Contingency Estimating & AACE Recommended Practice 40R-08
15 Chapter 15: Contract Types, Procurement Strategy & Commercial Terms
15.1 Contract Types & Risk Allocation (Firm Fixed Price, Cost-Plus, T&M) 15.2 Project Delivery Systems & Procurement Lifecycle (DBB, DB, EPC, CMAR) 15.3 Commercial Terms, Bonds, Retainage & Claims Management
16 Chapter 16: Quality Management, Value Engineering & Technical Communication
16.1 Quality Management & Cost of Quality (CoQ) 16.2 Value Engineering, Function Analysis System Technique (FAST) & Life Cycle Costing 16.3 The CCP Memo, Cost Report Writing, Data Presentation & AACE Ethics 16.4 AACE CCP Capstone Review & High-Yield Exam Strategies
17 Chapter 17: Cost Accounting, Budget Administration & Cost System Governance
17.1 Historical Cost Databases, Data Organization & Benchmarking 17.2 Committed Costs, Accruals, Work-in-Place & Invoice Validation 17.3 Budget Creation, Operating Budgets & Funding Requirement Forecasting 17.4 Cost System Governance, the Code of Accounts & Preserving Project Controls Data
18 Chapter 18: Estimate Preparation, Review & Cost Driver Analysis
18.1 ROM, Summary-Level and Detailed Estimates: Preparing and Reviewing 18.2 Verifying the Basis of Estimate & Analyzing Estimates 18.3 Cost Drivers, Change Order Impact & Resource Utilization Analysis
19 Chapter 19: Interfacing with Other Disciplines & Stakeholders
19.1 Coordinating with Internal Disciplines & Supporting Project Leadership 19.2 Communicating Cost Estimates and Cost/Schedule Key Performance Indicators 19.3 Supporting Tenders, Procurement, Audits, Field Verification & Recovery
20 Chapter 20: Cost Reporting, Dashboards & Profitability Analysis
20.1 Cost Reports, Performance Dashboards & the Reporting Cycle 20.2 Variance Reports & Explaining Project Financial Health 20.3 Profit Margin Analysis & Profitability Enhancement
21 Chapter 21: Cost Support to Scheduling & the CCP Exam Memo
21.1 Cost and Resource Loading Summaries from the Project Schedule 21.2 Translating Schedule Change into Cost Impact 21.3 Writing the CCP Exam Memo: A Worked Domain 7 Walkthrough
10.3 Learning Curve Theory (Wright Cumulative vs. Crawford Unit Models)
Key Takeaways Explain This Topic AI
Learning curve theory establishes that each time the cumulative production quantity of repetitive units doubles, the labor hours (or cost) required per unit decreases by a constant percentage (the Learning Rate, LR).
T.P. Wright's Cumulative Average Model states that the cumulative average labor hours per unit across the first x units follows Y_bar(x) = k * x^s, where s = ln(LR) / ln(2) is the learning slope.
J.R. Crawford's Unit Model states that the labor hours required to produce the individual x-th unit follows y(x) = k * x^b, where b = ln(LR) / ln(2).
For an identical nominal Learning Rate (e.g., 80%), Wright's Cumulative Average Model forecasts faster total labor hour reduction and lower total program hours than Crawford's Unit Model.
Typical industry learning rates are 70% for aerospace/complex assembly, 80%–85% for manufacturing and modular fabrication, and 90% for repetitive construction operations.
Last updated: August 2026
10.3 Learning Curve Theory (Wright Cumulative vs. Crawford Unit Models)
In repetitive manufacturing, modular prefabrication, and serial construction operations (such as multi-story buildings, tract housing, pipeline spreads, and solar farm installations), the labor hours required to perform a task decrease as workers repeat the identical operation. This phenomenon is modeled in cost engineering through Learning Curve Theory (also known as the Experience Curve, Progress Curve, or Improvement Curve).
First quantified by Theodore Paul Wright in 1936 during his studies of aircraft manufacturing costs at the Curtiss-Wright Corporation, learning curve analysis is a vital tool for the Certified Cost Professional (CCP). Cost engineers use it to prepare competitive bids, negotiate change orders, forecast cash flows, and establish production targets.
1. Core Principles & Sources of Learning
Learning curve theory is governed by a fundamental axiom:
Each time the cumulative production quantity of a repetitive product doubles, the time (or cost) required per unit decreases at a constant percentage rate. +-----------------------------------------------------------------------------+
- | SOURCES OF PRODUCTIVITY GAINS |
- | |
- | 1. OPERATOR PROFICIENCY: Muscle memory, dexterity, familiarity with |
- | tools, drawings, and physical assemblies. |
- | 2. TOOLING & JIGS: Development of specialized fixtures, custom |
- | templates, and ergonomic holding rigs. |
- | 3. METHOD ENGINEERING: Optimization of work sequence, waste elimination|
- | (Lean principles), and reduced motion fatigue. |
- | 4. MANAGEMENT LOGISTICS: Improved material staging, standardized parts, |
- | and reduced supervisory instruction time. |
- | 5. QUALITY REDUCTION: Fewer fabrication defects, less non-destructive |
- | testing (NDT) rejection, and reduced rework. |
- +-----------------------------------------------------------------------------+
Learning Rate (LR) & Rate of Learning (RoL):
Learning Rate (LR or Progress Ratio PR): The percentage of labor hours (or cost) remaining after a doubling of production quantity (e.g., 80% or 0.80).
Rate of Learning (RoL / Percentage Reduction): The percentage reduction achieved per doubling: RoL = 1 - LR (e.g., for an 80% curve, RoL = 1.0 - 0.80 = 20% reduction).
Zero Learning (100% Curve): Completely machine-paced operations where human repetition yields 0% time reduction (LR = 1.00).
2. The Two Primary Mathematical Models: Wright vs. Crawford
In cost engineering, two distinct mathematical formulations exist. Distinguishing between them is essential on the CCP exam: +-----------------------------------------------------------------------------+
- | WRIGHT VS. CRAWFORD MATHEMATICAL MATRIX |
- | |
- | MODEL: T.P. WRIGHT (1936) J.R. CRAWFORD (1944) |
- | "Cumulative Average Model" "Individual Unit Model" |
- | ------------------ ------------------------- -------------------------- |
- | Applies LR to: CUMULATIVE AVERAGE of INDIVIDUAL x-th UNIT |
- | all units produced so far direct unit time |
- | |
- | Primary Formula: Y_bar(x) = k * x^s y(x) = k * x^b |
- | |
- | Slope Parameter: s = ln(LR) / ln(2) b = ln(LR) / ln(2) |
- | |
- | Total Hours (Tx): Tx = x * Y_bar(x) Tx = Sum_{i=1}^x y(i) |
- | Tx = k * x^(1 + s) Tx ~ [k / (1+b)] * x^(1+b) |
- | |
- | Marginal Unit x: y(x) = Tx - T(x-1) y(x) = k * x^b |
- | |
- | Historical Domain: Aerospace, defense, Discrete manufacturing, |
- | long-cycle programs modular fabrication, constr.|
- +-----------------------------------------------------------------------------+
- 3. T.P. Wright's Cumulative Average Model
Under Wright's model, the cumulative average time [Y_bar(x)] across the first x units decreases by the factor LR whenever cumulative output doubles: Y_bar(x) = k * x^s
ln[Y_bar(x)] = ln(k) + s * ln(x)
Where:
- Y_bar(x) = Cumulative average labor-hours (or cost) per unit for the first x units
- k = Labor-hours (or cost) required to produce the first unit (Unit 1)
- x = Cumulative production quantity
- s = ln(LR) / ln(2) = Learning curve slope index (s < 0)
- Total Program Hours (Tx) and Individual Unit Hours [y(x)]: Tx = x * Y_bar(x) = x * [k * x^s] = k * x^(1 + s)
- y(x) = Tx - T(x-1) = k * [x^(1+s) - (x-1)^(1+s)]
Wright's Doubling Table (k = 1,000 hours, LR = 80%, s = -0.32193):
Cumulative Units (x) Doubling Step Cumulative Average Hours [Y_bar(x)] Total Cumulative Hours (Tx) Marginal Unit Hours [y(x)]
- 1 Baseline 1,000.0 hrs 1,000.0 hrs 1,000.0 hrs
- 2 1st Doubling 1,000 * 0.80 = 800.0 hrs 2 * 800.0 = 1,600.0 hrs 1,600 - 1,000 = 600.0 hrs
- 4 2nd Doubling 800.0 * 0.80 = 640.0 hrs 4 * 640.0 = 2,560.0 hrs T4 - T3 = 453.7 hrs
- 8 3rd Doubling 640.0 * 0.80 = 512.0 hrs 8 * 512.0 = 4,096.0 hrs T8 - T7 = 354.3 hrs
- 16 4th Doubling 512.0 * 0.80 = 409.6 hrs 16 * 409.6 = 6,553.6 hrs T16 - T15 = 280.0 hrs
- 4. J.R. Crawford's Unit Model (Stanford Model)
Under Crawford's model, the direct labor-hours required to produce the individual x-th unit [y(x)] decreases by the factor LR whenever the unit number doubles: y(x) = k * x^b
ln[y(x)] = ln(k) + b * ln(x)
Where:
- y(x) = Direct labor-hours (or cost) required to manufacture the specific x-th unit
- k = Labor-hours (or cost) for the first unit (Unit 1)
- b = ln(LR) / ln(2) = Learning curve slope index (b < 0)
Total Program Hours (Tx) under Crawford: Tx = Sum_{i=1}^x y(i) = k * Sum_{i=1}^x [i^b] ~ [k / (1+b)] * x^(1+b) (continuous approximation)
Y_bar(x) = Tx / x ~ [k / (1+b)] * x^b
Crawford's Doubling Table (k = 1,000 hours, LR = 80%, b = -0.32193):
Unit Number (x) Doubling Step Individual Unit Hours [y(x)] Total Cumulative Hours [Tx = Sum y(i)] Cumulative Average [Y_bar(x) = Tx / x]
- 1 Baseline 1,000.0 hrs 1,000.0 hrs 1,000.0 hrs
- 2 1st Doubling 1,000 * 0.80 = 800.0 hrs 1,000 + 800 = 1,800.0 hrs 1,800 / 2 = 900.0 hrs
- 4 2nd Doubling 800.0 * 0.80 = 640.0 hrs Sum_{i=1}^4 y(i) = 3,142.1 hrs 3,142.1 / 4 = 785.5 hrs
- 8 3rd Doubling 640.0 * 0.80 = 512.0 hrs Sum_{i=1}^8 y(i) = 5,345.9 hrs 5,345.9 / 8 = 668.2 hrs
- 16 4th Doubling 512.0 * 0.80 = 409.6 hrs Sum_{i=1}^16 y(i) = 8,919.7 hrs 8,919.7 / 16 = 557.5 hrs
5. Critical Comparison: Wright vs. Crawford for the Same Learning Rate +-----------------------------------------------------------------------------+
- | COMPARISON OF TOTAL HOURS FOR 16 UNITS (k = 1,000 LH, LR = 80%)|
- | |
- | MODEL UNIT 16 TIME TOTAL 16 UNITS CUMULATIVE AVERAGE |
- | ----------------- ---------------- ---------------- ------------------ |
- | Wright Cumulative 280.0 Hours 6,553.6 Hours 409.6 Hours/Unit |
- | Crawford Unit 409.6 Hours 8,919.7 Hours 557.5 Hours/Unit |
- | |
- | KEY TAKEAWAY: Wright's model produces significantly faster total program |
- | labor reduction and lower total hours than Crawford's model for the same |
- | nominal learning rate percentage! |
- +-----------------------------------------------------------------------------+
- 6. Industry Typical Learning Rates
Learning rates vary across industries depending on the ratio of manual craft labor to automated machine control: +-----------------------------------------------------------------------------+
- | TYPICAL INDUSTRY LEARNING RATE BENCHMARKS |
- | |
- | LEARNING RATE (LR) INDUSTRY / OPERATION TYPE |
- | ------------------ -------------------------------------------------- |
- | 70% - 75% Aerospace, manual airframe assembly, complex piping |
- | (High manual labor content, ~75% labor / 25% machine)|
- | |
- | 80% - 85% Standard manufacturing, modular precast fabrication, |
- | electromechanical assembly (50% labor / 50% machine) |
- | |
- | 90% - 95% Repetitive construction operations: multi-story floor|
- | concrete framing, pipeline spreads, solar PV racking |
- | (High mechanized content, ~25% labor / 75% equipment)|
- | |
- | 100% Fully automated robotic machining / continuous |
- | chemical refining (Zero human learning effect) |
- +-----------------------------------------------------------------------------+
- 7. Step-by-Step Worked Numerical Case Studies
- Case Study 1: Crawford Unit Model — Batch Cost Forecasting
Problem: A modular fabrication yard is manufacturing 8 identical pre-assembled pipe racks (PARs). The first unit (Unit 1) required 1,200 labor-hours . The contractor's historical learning curve for pipe rack modularization is 85% (LR = 0.85).
Calculate the labor hours required to fabricate the 8th unit [y(8)] .
Calculate the labor hours required for the 5th unit [y(5)] .
Solution:
Step 1: Compute Slope Index (b): b = ln(0.85) / ln(2) = -0.1625189 / 0.6931472 = -0.234465
Step 2: Find Hours for Unit 8 via Doubling Rules:
- Unit 1: 1,200.0 hrs
- Unit 2 (1st doubling): 1,200 * 0.85 = 1,020.0 hrs
- Unit 4 (2nd doubling): 1,020 * 0.85 = 867.0 hrs
- Unit 8 (3rd doubling): 867 * 0.85 = 736.95 hrs
- Step 3: Find Hours for Unit 5 using Crawford Equation: y(5) = k * x^b = 1,200 * 5^(-0.234465)
- 5^(-0.234465) = e^[-0.234465 * ln(5)] = e^[-0.234465 * 1.609438] = e^(-0.377356) = 0.685673
- y(5) = 1,200 * 0.685673 = 822.81 labor-hours
- Case Study 2: Wright Cumulative Average Model — Program Bid Estimating
Problem: A defense contractor bids on a production batch of 4 specialized radar modules . Unit 1 requires 5,000 labor-hours at a direct labor rate of $80/hour . The contract specifies an 80% Wright Cumulative Average curve .
What is the cumulative average hours per unit for the 4 units?
What is the total direct labor cost for the entire 4-unit contract?
Solution:
Step 1: Calculate Cumulative Average for 4 Units (2 doublings):
- Y_bar(1) = 5,000 hrs
- Y_bar(2) = 5,000 * 0.80 = 4,000 hrs
- Y_bar(4) = 4,000 * 0.80 = 3,200 hours/unit
Step 2: Calculate Total Labor Hours for 4 Units (T4): T4 = x * Y_bar(4) = 4 * 3,200 hrs = 12,800 total labor-hours
- Step 3: Calculate Total Direct Labor Cost: Total Cost = 12,800 LH * $80/LH = $1,024,000
- 8. Disruptors, The Forgetting Curve & The Plateau Effect
In real-world construction and manufacturing, learning curve efficiency gains do not continue indefinitely and can be severely disrupted: +-----------------------------------------------------------------------------+
- | LEARNING CURVE DISRUPTORS & LIMITATIONS |
- | |
- | 1. DESIGN MODIFICATIONS & CHANGE ORDERS: |
- | - Altering drawings midway through production destroys muscle memory, |
- | invalidates prefabricated jigs, and shifts the curve upward. |
- | |
- | 2. THE FORGETTING CURVE (Production Breaks & Gaps): |
- | - Shutting down production for weeks/months results in skill decay. |
- | - Upon restart, unit hours spike backward toward initial Unit 1 levels.|
- | |
- | 3. CRAFT LABOR TURNOVER: |
- | - Replacing experienced workers with new hires dilutes collective gang |
- | proficiency and resets the learning progression. |
- | |
- | 4. THE PLATEAU EFFECT (Technological Floor): |
- | - After significant volume (e.g., 50-100 units), human learning gains |
- | reach a physical plateau dictated by machine cycle limits and |
- | material handling constraints. No further reduction occurs. |
- +-----------------------------------------------------------------------------+
Important
AACE CCP Exam Alert — Learning Curve Key Principles:
Wright's Model applies learning rate LR to the Cumulative Average [Y_bar(x) = k * x^s].
Crawford's Model applies learning rate LR to the Individual Unit [y(x) = k * x^b].
The exponent parameter is s = b = ln(LR) / ln(2). For an 80% curve, s = -0.322; for an 85% curve, s = -0.234; for a 90% curve, s = -0.152.
For the same nominal LR, Wright yields fewer total hours than Crawford.