Parametric Cost Estimating

Handbook

Chapter 1
Background and Definitions
 History
 NASA and DoD
– Problems with cost estimation, especially in RDTE and
first unit production

 Work by the RAND Corporation:

– Started a cost analysis department in 1950
– Cost estimation for aircraft industry
 Cost Realism

 Foundation of the estimate

– System of logic
– Assumptions of the future
– Reasonableness of the historic basis

 Key Questions
– See Chapter 1 pages 11-12.
 Chapter 2
Collection and Normalization of Parametric Data
 Consistent Scope
 Anomalies

 Improved Technology
 Chapter 2
Pitfalls in the Use of Parametric Models
 Using the model outside the rang of the data
 Anomalies

 Improved Technology
 Cost Estimating Process
Project Definition

Receive Customer Build or Obtain/Participate in

Request and Obtain WBS Development of Project
Understand Project Technical Description

Cost Methodology
Develop Select Select/Build Gather and
Ground Rules and Cost Estimating Cost Model Normalize Data
Assumptions Methodology

Estimate

Develop Develop and Document Present Update Cost

Point Estimate Incorporate Cost Probabilistic Estimate Estimate on a
Risk Assessment Cost Estimate Results Regular Basis
 Selecting Cost Estimating Methodology
Method Pre-Phase A Phase A Phase B Phase C/D Phase E

Parametric

Analogy

Engineering
Build Up

Legend Primary Applicable Not

Applicable
Parametric Cost Estimating Methodology
Define
Estimating Collect
“Hypothesis” “Relationship” Evaluate &
Data Normalize Analyze Data
Data For Candidate
Relationships

Perform
Statistical Test
(Regression) Select Cost
Relationships
Analysis Estimating
Relationship
Parametric Strengths and Weaknesses
Strength Weakness
CER’s excellent for “what if” Difficult for others to understand

Statistically sound predictors Must fully document process of

that give info on the confidence data collection, development,
of their predictive ability conclusions and validation

Eliminates reliance on opinions Collecting appropriate data and

by using actual data developing CER’s can be time
consuming and expensive

Defensible, use of scientific Loses predictive ability outside

method the range of the data
Analogy Cost Estimating Methodology
 Performed on the basis of comparison and extrapolation to
like items or efforts.
 Cost data from one past program that is technically
representative of the program to be estimated serves as the
basis of the estimate.
 Cost data is subjectively adjusted upward or downward
depending on whether the subject system is felt to be more
or less complex than the analogous program.
Analogy Strengths and Weaknesses
Strength Weakness
Based on actual historical data Relies on a single data point

Fairly quick Can be difficult to identify

appropriate analog

Readily understood Requires “normalization” to

ensure accuracy

Accurate for minor deviations Relies on extrapolation and/or

from the analogy expert judgement for
“adjustment factors”
 Engineering Build Up Methodology
(Grass Roots)
 Rolls up individual estimates for each element into the overall
estimate.
 Cost involves the computation of the cost of a WBS element by
estimating at the lowest level of detail (often referred to as the
“work package” level) wherein the resources to accomplish the
work effort are readily distinguishable and discernible.
 Labor requirements are often estimated separately from material
requirements.
 Overhead factors for cost elements such as Other Direct Costs
(ODCs), General and Administrative (G&A) expenses, materials
burden, and fee are generally applied to the labor and materials
costs to complete the estimate.
Grass Roots Cost Estimating Methodology

Segregate into
Decompose
CES/WBS Estimate
CES/WBS into
Individual Aggregate
“Work Packages”
“Work Packages” “Work Packages”

Test for
Perform Aggregate into Omissions &
“Sanity Check” “Total Estimate” Duplications
 “Grass Roots” Strengths and
Weaknesses
Strength Weakness
Intuitive Costly, takes considerable time and cost
to create a WBS build-up estimate
Defensible Not readily responsive to “what if”
requirements
Credibility provided by visibility New estimates must be “built-up” for
into the BOE of each cost element each alternative scenario
Severable; the entire estimate is not Cannot provide statistical confidence
compromised by the miscalculation level
of an individual cost element
Provides excellent insight into major Does not provide good insight into cost
cost contributors drivers
Reuse; easily transferable for use Relationship/links among cost elements
and insight into individual projects must be “programmed” by the analyst
 Data Types and Sources
Data Category Data Type Data Sources
Cost Data Historical Costs Basic Acct Records
Labor Costs Cost Reports
CERs from prior projects Historical Databases

Contracts (Secondary)
Cost Proposals (Secondary)

Technical/Operational Physical Characteristics Functional Specialists

Performance Characteristics Technical Databases

Performance Metrics Engineering Specifications
Technology Descriptors Engineering Drawings
Major Design Changes Performance/Functional Specs
Operational Environment End User Operators
 More Data Types and Sources
Data Category Data Type Data Sources
Project Data Devlmnt & Prodctn Sched Project Database
Quantities Produced Functional Organizations
Production Rates Project Management Plan

Equivalent Units Major Subcontractors

Breaks in Production

Significant Design Changes

Anomalies (strikes, disasters)

 Normalizing Data
 Once data has been collected it needs to be normalized which
involves analyzing the raw data collected and adjusting it to
make it consistent.
 Inconsistencies include:
– changes in dollar values over time (inflation)
– learning or cost improvements for organizational efficiency, and if more
than one unit is being produced
– effects of production rates on the data set being analyzed
– physical characteristics such as size and weight
– Complexity of system.
 Cost Risk
 Identify and Quantify Cost-Risk
– Identify and assess risk
– Translate risk assessment into cost impact
– Perform S-curve and CRM scenario-based cost risk
– Incorporate CRM scenario-based and S-curve cost risk in Cost
Analysis Data Requirement (CADRe) Part C Life Cycle Cost
Estimate (LCCE)
 Establish Cost-Risk Reporting
– Develop RFP CADRe & EVM Data Requirements Description
(DRD) and equivalent project plan requirements
– Evaluate EVM and LCC DRD’s in proposals/project plans
– Do Integrated Baseline Review
 Manage Cost-Risk Using Reported Data
– Do EVM performance measurement and CADRe S-curve analysis
– Compile end-of-contract cost-risk data for database updates, data
evaluation and analysis and cost-risk algorithm updates
? ?
Key Risk Analysis Questions

 What can happen?

 How likely is it?

 What are the consequences if it does happen?

 Risk Components
 Root Cause (future tense / yet to happen):
– If eliminated or corrected would prevent a potential
consequence from occurring.

 Probability (likelihood):
– Assessed with the formulation of the “Future Root Cause”

 Consequence (effect):
– Impact of that future occurrence.
Reference: Risk Management Guide
for DOD Acquisition August 2006
 Risk Management Process

Risk
Identification

Risk Risk
Analysis Tracking
Risk
Mitigation
Planning

Risk
Mitigation
A Process, Not an Event Plan Implementation
 Risk Reporting Matrix
Risk Title (P, S, or C)
Risk Causal Factor
Mitigation Approach
5

4
Likelihood

1
1 2 3 4 5
Consequence
 Levels of Likelihood Criteria
Level Likelihood Probability of Occurrence

1 Not Likely ~10%

2 Low Likelihood ~30%

Likelihood

3 Likely ~50%

4 Highly Likely ~70%

5 Near Certainty ~90%

 Levels & Types of Consequence Criteria
Level Technical Performance Schedule Cost

Minimal or no consequence to technical performance Minimal or no Minimal or no

1
impact impact

Minor reduction in technical performance or Able to meet key Budget or UPC

2 supportability, can be tolerated with little or no impact dates increase
to program Slip <__months <__ (1% Budget)

Minor sched slip. Able Budget or UPC

Moderate reduction in technical performance or to meet key M/S with
3 supportability with limited impact on program objectives increase
no sched float
<__ (5% Budget)
Slip <__months

Significant degradation in technical performance or Program Critical Budget or UPC

major shortfall in supportability; may jeopardize path affected increase
4 program success
Slip <__months <__ (10% Budget)

Severe degradation in technical performance; Cannot Cannot meet Exceeds APB

5 meet KPP or key technical/supportability threshold; program key M/S threshold
will jeopardize program success >__ (10% Budget)
Slip >__months
 Risk Mitigation Planning
Risk Handling
 Four Fundamental Strategies
– Avoidance
– Control
– Assumption
– Transfer
 Handling also requires Risk Monitoring activities:
– Test and evaluation
– Earned Value reporting
– Technical Performance Measures
– Program Metrics
– Network Scheduling (PERT, CPM, etc)
– Technical Reviews (entrance / exit criteria)
 Cost Risk Approaches

 Analytic Approach
– Scenario Based Method (SBM)
– Formal Risk Assessment/Method of Moments

 Simulation Approach
 Scenario Based Method
 SBM:
– A variation of sensitivity analysis where one specifies a well-
defined set of conditions or scenarios that would create a condition
that management would like to guard against

– It postulates on specified scenarios that, if they occurred, would

result in costs higher than the level planned or budgeted.

– The scenarios do not have to represent “worst cases,” rather they

should reflect a set of conditions a PM or decision-maker would
want to budget for, should any or all of these conditions occur.

– Requires an eight-step process which is partly subjective in nature

and makes use of methodologies addressed in IMSE 5205
(Engineering Risk-Benefit Analysis)
 Formal Risk Assessment (FRISK):
Method of Moments
 FRISK:
– A statistical approximation method that utilizes triangular
distributions of inputs that approximate the total system using a log-
normal distribution

– Base on two fundamental postulates:

 The sum of a series of triangular distributions is a lognormal
distribution.
 There is a pairwise Pearson correlation between cost elements

– The method supports cost risk analysis by allowing the user to

statistically sum WBS elements to obtain a probability distribution
of total costs.

– The WBS elements are estimated by triangular distributions

 Simulation Approach
 Monte Carlo Simulation:
– Calculates numerous scenarios of a model by repeatedly picking
random values from the input variable distributions for each
“uncertain” variable and calculating the results.

– Typically run 2500 – 10000 iterations.

– Results of the simulation include risk-adjusted estimates and

corresponding statistical estimate distributions.

– The estimate distributions provide the decision-maker with a range

of possible outcomes with a minimum and maximum value or
bounds.

– Often uses commercially available products such as Crystal Ball or

@Risk amongst others.

– The simulation follows a seven-step process (See pages 2-21 thru 2-

27 in the NASA Cost Estimating Handbook
 The Triangular Distribution
 Requires three estimates for each WBS element
– Minimum (Optimistic)
– Maximum (Pessimistic)
– Most Likely

Optimistic Most Likely Pessimistic

Cost Cost Cost