<!-- argument-hint: [process task/step, WBS, CADRe, methodology (analogy/parametric/CER/build-up), GR&A, normalization, cost-risk/S-curve/UFE, JCL, sensitivity/trade/CAIV/affordability, economic analysis/NPV, chapter number] -->
NASA Cost Estimating Handbook (CEH) v4.0
Source: NASA (US Government work, public domain) | Chapters: 7
When to use
Use this skill to produce or review a NASA cost estimate end to end: scoping and defining what is being estimated, choosing an estimating methodology and tool, normalizing data, turning a deterministic point estimate into a confidence curve with a sized reserve, running a Joint Cost and Schedule Confidence Level analysis, supporting decisions with sensitivity/trade/affordability/CAIV analyses, and comparing alternatives with a discounted economic analysis. This is the cost-and-cost-risk complement to NASA's systems-engineering and risk material (nasa-risk, nasa-se-handbook, nasa-npr-7123) and to the broader government cost canon (gao-cost, gao-schedule).
Prerequisites: none — plain Markdown; no MCP server, API key, or licence tier needed at runtime.
How to Use This Skill
- Without arguments — load the core frameworks below: the 12-step process, the methodology choices, the cost-risk chain, the JCL, and the decision-support and economic-analysis toolkit.
- With a topic — ask about a process task (e.g. "GR&A", "normalization"), a methodology ("parametric CER", "engineering build-up"), the cost-risk chain ("S-curve", "UFE"), the JCL ("time-dependent vs. time-independent"), a decision-support method ("CAIV", "trade space"), or economic analysis ("NPV", "discount rate").
- With a chapter —
ch01 (process overview), ch03 (methodology), ch04 (cost-risk), ch05 (JCL), ch06 (decision support), ch07 (economic analysis).
Supporting files: glossary.md, patterns.md, cheatsheet.md.
Cost as a design variable. The handbook's spine is that cost is not a number reported after design — it is information that shapes decisions. Keep cost drivers visible, challenge estimates that drift from history, and always deliver uncertainty alongside the point value.
Core Frameworks & Mental Models
The NASA 12-Step Cost Estimating Process (3 parts)
The process is a twelve-step sequence in three parts. Presented in order for teaching, it is iterative and rarely linear in practice; GR&A is the most frequently revisited task.
| Part | Tasks | What it produces |
|---|
| 1 — Project Definition | 1 customer request · 2 WBS · 3 technical description | a scoped, commonly-understood definition of what is estimated |
| 2 — Cost Methodology | 4 GR&A · 5 methodology · 6 models/tools · 7 data & normalization | the approach and framework for the estimate |
| 3 — Cost Estimate | 8 develop · 9 cost-risk · 10 document · 11 present · 12 update | the estimate itself — produced, hardened, justified, briefed, maintained |
Governing policy an estimator must know: NPD 1000.5 and NPR 7120.5E (acquisition / life-cycle management), NPR 7123.1B (review entrance/exit criteria). NASA's process is deliberately consistent with — but tailored from — the GAO Cost Estimating and Assessment Guide (GAO-09-3SP), with a task-level mapping between them.
Project Definition — the data backbone (Part 1)
- Four critical elements — resources, data, schedule, expectations — jointly agreed between estimator and decision maker before a methodology is chosen.
- WBS — a product-oriented decomposition giving full coverage without double counting; mandated by NPR 7120.5E, standardized to Level 2, extended below via the CADRe standard WBS. The estimator owns the cost WBS and maps it back to the standard.
- CADRe (Cost Analysis Data Requirement) — captures programmatic/technical/cost data six times across the life cycle; stored in ONCE (One NASA Cost Engineering database), which also receives end-of-contract actuals to update cost models. This is the empirical grounding for future estimates.
- Technical baseline — one common definition of the system so the Project Office estimate and an independent-review estimate are comparable.
Estimating Methodologies (Task 5)
Three core methods plus a fourth data source; the choice follows data availability, which shifts across the life cycle:
- Analogy — adjust a similar fielded system's cost; fast but subjective, still needs normalization.
- Parametric — apply a Cost Estimating Relationship (CER), a regression-derived Y-vs-X equation whose drivers need both statistical correlation and a sound logical rationale; valid only inside its data range. Built via the seven-step parametric process.
- Engineering build-up — decompose the WBS into work packages, estimate each (labor and material separately), aggregate; defensible and severable but costly, slow, "what-if"-resistant.
- EVM extrapolation — a fourth source once a program produces actuals.
- Delphi — a documented expert-judgment fallback only when empirical data are absent.
GR&A, Models, and Normalization (Tasks 4, 6, 7)
- Ground Rules and Assumptions (GR&A) — global (whole-estimate: dollars, schedule, inclusions/exclusions, quantities) and element-specific; documented continuously as they evolve, agreed with the P/pM and stakeholders.
- Models mechanize a methodology but don't replace judgment — no single model fits all purposes; calibrate parametric models to comparable history, validate before use, cross-check with a second method, keep current.
- Normalization makes raw data comparable: inflation (BY/CY/RY dollars; NASA New Start Inflation Index for new efforts, DCAA forward pricing for contracted work), plus learning curves, production rate, scope, anomalies, and reporting-system mapping. Collect risk data here — uncertainty drives the later cost-risk analysis.
The Cost-Risk Chain (Tasks 8–9)
Point estimate → S-curve (CDF) → UFE reserve.
- The deterministic LCC point estimate is the start of cost-risk work, not the deliverable — summing most-likely values understates the total.
- The Task 9 cost-risk assessment produces a credible S-curve / Cumulative Distribution Function across the cost range; the confidence level is read off it.
- Unallocated Future Expense (UFE) is the reserve sized to reach the management-chosen confidence; phase it where risk realizes (PDR→CDR, Integration & Test).
- Document to the rebuild standard (Basis of Estimate complete enough for an independent analyst to reproduce the number); keep the estimate a living baseline (Task 12, with EVM against actuals).
Joint Cost and Schedule Confidence Level (JCL)
The probability of finishing at or under both the cost target and the schedule target — jointly, not as two separate confidences. Produced via a Probabilistic Cost-Loaded Schedule (PCLS) in six steps (0 goals · 1 schedule · 2 cost-load · 3 risk · 4 uncertainty · 5 results).
- Time-Dependent (TD) cost grows when the schedule slips (the "marching army"); Time-Independent (TI) cost (e.g. materials) does not.
- Risk = discrete probability-and-impact event; uncertainty = indefiniteness of the baseline (triangular distribution; usually dominates variance). Modeled together but kept distinct.
- The scatterplot's lower-left quadrant fraction is the JCL; frontier lines mark chosen confidence levels; the chart is a snapshot of the current plan only.
- Policy: ABC = 70% JCL, MA ≥ 50%, reviewed by the Standing Review Board, set at KDP-C (projects) / KDP-I (programs).
Decision-Support and Economic Analysis (Ch 6–7)
- Five decision-support categories: Sensitivity, Trade Studies, Affordability, CAIV, Economic Analysis (procedures in Appendix N).
- CAIV / Affordability invert the default flow: cost becomes an input/constraint and design converges on cost while meeting threshold performance.
- Economic Analysis layers discounting and cost-vs-benefit on the LCCE: seven steps, NPV as the criterion (most-positive wins), OMB Circular A-94 for discount rates, and inflation (NNSI) kept distinct from the time-value-of-money.
Chapter Index
| # | Section | Key content |
|---|
| ch01 | Introduction & Process Overview | Why cost estimating exists; cost as a design variable; the 12-step / 3-part process; governing policy (NPD 1000.5, NPR 7120.5E, NPR 7123.1B); mapping to the GAO guide |
| ch02 | Project Definition (Tasks 1–3) | Customer request and the four critical elements; the WBS (coverage without double counting, Level 2 standard, CADRe extension); CADRe and ONCE; the technical baseline |
| ch03 | Cost Methodology (Tasks 4–7) | GR&A; analogy / parametric (CER) / engineering build-up / Delphi; models and tools; data gathering and normalization; inflation (BY/CY/RY) and the NASA New Start Inflation Index |
| ch04 | Cost Estimate (Tasks 8–12) | The LCC point estimate; probabilistic estimating; the S-curve / CDF; cost-risk assessment and UFE; the Basis of Estimate and the rebuild standard; presentation and updating; EVM |
| ch05 | Joint Cost & Schedule Confidence Level (JCL) | The PCLS method; the six steps; time-dependent vs. time-independent cost; risk vs. uncertainty; the scatterplot, quadrants and frontier lines; ABC/MA policy thresholds |
| ch06 | Decision-Support Analyses | Sensitivity analysis (five steps); cost-performance trade studies and the trade space; make-vs-buy and lease-vs-buy; affordability; Cost As an Independent Variable (CAIV) |
| ch07 | Economic Analysis | Discounting and the time-value-of-money; NPV and measures of merit; the seven-step EA process; OMB Circular A-94; nominal vs. real rates; cost-effectiveness when benefits won't quantify |
Topic Index
- Affordability analysis → ch06
- Analogy method → ch03
- Basis of Estimate (BOE) / rebuild standard → ch04
- CADRe (Cost Analysis Data Requirement) → ch02, ch04
- CAIV (Cost As an Independent Variable) → ch06
- Cost as a design variable → ch01
- Cost Estimating Relationship (CER) → ch03, ch04
- Cost-risk assessment → ch04
- Cross check → ch04
- Cumulative Distribution Function (CDF) / S-curve → ch04
- Decision support (five categories) → ch06
- Delphi method → ch03
- Discounting / discount rate → ch07
- Earned Value Management (EVM) → ch04, ch03
- Economic Analysis (EA) → ch07
- Engineering build-up method → ch03
- Four critical elements → ch02
- Frontier line (JCL) → ch05
- Ground Rules and Assumptions (GR&A) → ch03, ch04
- Inflation / dollar types (BY / CY / RY) → ch03
- Joint Cost and Schedule Confidence Level (JCL) → ch05
- Lease-versus-buy analysis → ch06
- Life-Cycle Cost (LCC) point estimate → ch04
- Make-versus-buy analysis → ch06
- Methodology selection (data availability) → ch03
- NASA New Start Inflation Index (NNSI) → ch03, ch07
- Net Present Value (NPV) → ch07
- Nominal versus real rate → ch07
- Normalization → ch03
- OMB Circular A-94 → ch07, ch06
- ONCE database → ch02
- Parametric method → ch03
- Point estimate → ch04, ch06
- Probabilistic cost estimate → ch04
- Probabilistic Cost-Loaded Schedule (PCLS) → ch05
- Process overview (12 steps, 3 parts) → ch01
- Project Definition tasks → ch02
- Risk versus uncertainty → ch05
- Sensitivity analysis → ch06, ch04
- Technical baseline / description → ch02
- Time-Dependent vs. Time-Independent cost → ch05
- Trade study / trade space → ch06
- Unallocated Future Expense (UFE) → ch04, ch05
- Work Breakdown Structure (WBS) → ch02
Supporting Files
- glossary.md — key cost-estimating terms, alphabetical, with chapter references
- patterns.md — reusable patterns (define-before-price, method by data availability, normalize-then-compare, point-estimate-to-S-curve, rebuild-standard documentation, the JCL/PCLS, affordability/CAIV, the seven-step EA) with When/How/Trade-offs
- cheatsheet.md — the 12-step process table, methodology selection table, dollar types, the cost-risk chain, the JCL, decision-support analyses, economic analysis formulas, tells & smells
Scope & Limits
Covers: NASA's cost estimating process per the NASA Cost Estimating Handbook v4.0 — the 12-step, three-part process; Project Definition (customer request, WBS, CADRe/ONCE, technical baseline); methodology selection (analogy, parametric/CER, engineering build-up, EVM, Delphi); GR&A, model/tool selection, and data normalization/inflation; the deterministic-to-probabilistic cost-risk chain (point estimate, S-curve/CDF, UFE) and documentation/presentation/updating; JCL analysis via probabilistic cost-loaded schedules; the five decision-support analyses (sensitivity, trade studies, make/lease-vs-buy, affordability, CAIV); and economic analysis (discounting, NPV, OMB Circular A-94).
Does not cover in depth: the handbook's appendix-depth procedures, worked examples, and tool-specific instructions (named here as Appendices C, F, G, H, I, J, K, N and pointers to ACEIT/PCEC/NAFCOM/SEER/PRICE — not reproduced); detailed regression/statistical method for building CERs; schedule construction (see nasa-schedule / gao-schedule and the NASA Schedule Management Handbook); formal risk identification and handling beyond the cost interface (see nasa-risk); and the full GAO and DoD cost guides (see gao-cost; complemented by dau-se-guidebook).
Source version: NASA Cost Estimating Handbook (CEH) Version 4.0, February 2015. Public-domain US Government work; guidance is NASA-specific and not a substitute for the current controlled handbook in a live estimating engagement.