2012 australian cost conference keynote cost estimation: a key component of affordable program...
TRANSCRIPT
2012 Australian Cost Conference KeynoteCost Estimation: A Key Component of Affordable Program Success
Dan Galorath
[email protected] Copyright 2012 Galorath Incorporated
Key Points
Cost is a key project
performance
parameter
Viable affordability
decisions yield project
achievements
Cost Estimation
with repeatable process is
best practice
2
© 2012 Copyright Galorath Incorporated 2
© 2012 Copyright Galorath Incorporated 3
Using Best Practices Are A Best Practice Themselves
• “Method or technique that has consistently shown results superior to those achieved with other means, and that is used as a benchmark”
• “Best" practice can evolve to become better as improvements are discovered
© 2012 Copyright Galorath Incorporated 4
Delusions of Success: How Optimism Undermines Executives' Decisions (Source: Richard Hartley, HBR)
• Problem:
Humans seem hardwired to be optimists
• Optimism from cognitive biases & organizational pressures• Exaggerate talents & degree of control
• Attribute negative consequences to external factors
• Anchoring (relying too heavily on one piece of information) magnifies optimism
• Most pronounced for new initiativesBest practice: Temper with “outside view”
Supplement traditional forecasting w/ statistical parametrics
Don’t remove optimism, but balance optimism & realism
© 2012 Copyright Galorath Incorporated 5
While Optimism Needs Tempering, So Does Short Sightedness (Source Northrop)
© 2012 Copyright Galorath Incorporated 7
Cost Overruns Are Everywhere
• “GAO: Staggering cost overruns dwarf modest improvements in Defense acquisition”
• R&D costs of weapons programs increased 42% over original estimates
• Average delay 22 months in delivering initial capabilities
• Evolving technical requirements
• Shortage of qualified government staff to manage
• "Every dollar of cost growth on a DoD weapon system represents a lost opportunity to pay for another national priority"
• “The Collins Class submarine program: Murphy was an optimist” http://epress.anu.edu.au/apps/bookworm/view/Agenda,+Volume+19,+Number+1,+2012/9601/Ergas.htm
• “BG Group shares hit by $5.4B cost overruns on Australian Liquified Natural Gas project”
© 2012 Copyright Galorath Incorporated 8
Software Is A Key Risk Item In Weapons Systems
• Navy Mobile User Objective Satellite Communication System delays to the Joint Tactical Radio System, a set of software-defined radios causes advanced MUOS capabilities to be drastically underused… GAO
• GAO identified 42 programs at risk for cost & schedule 1. military requirements changes
2. software development challenges
3. workforce issues
• National Institute of Standards and Technology (NIST)
• Software defects cost nearly $60 Billion Annually
• 80% of development costs involve identifying and correcting defects
Software, not Hardware or technology readiness levels were called out
© 2012 Copyright Galorath Incorporated 9
Example: Project Cost Alone Is not The Cost of IT Failure (Source: HBR)
• Case Study: Levi Strauss
• $5M ERP deployment contracted
• Risks seemed small
• Difficulty interfacing with customer’s systems
• Had to shut down production
• Unable to fill orders for 3 weeks
• $192.5M charge against earnings on a $5M IT project failure
“IT projects touch so many aspects of organization they pose a new singular risk”
http://hbr.org/2011/09/why-your-it-project-may-be-riskier-than-you-think/ar/1
© 2012 Copyright Galorath Incorporated 10
An ROI Analysis of A New System: Should We Fund This?
• Can we do better?• Will stakeholders tolerate a
loss for 3 years?• What is the risk?
© 2012 Copyright Galorath Incorporated 11
Software & IT Systems Are About Business Value
• “Software economics should provide methods for analyzing the choices software projects must make.” Leon Levy
• Business economics should provide methods for analyzing choices in which projects to fund
© 2012 Copyright Galorath Incorporated 12
Potentially 80% of Projects Don’t Return Adequate Value
Most projects cost more than they return, Mercer Consulting:
•“When the true costs are added up, as many as 80% of technology projects actually cost more than they return. It is not done intentionally but the costs are always underestimated and the benefits are always overestimated.” Dosani, 2001
© 2012 Copyright Galorath Incorporated 13
• Cutter Consortium Software Project Survey:
• 62% overran original schedule by more than 50%
• 64% more than 50% over budget
• 70% had critical product quality defects after release
• Standish Group CHAOS Report
• 46% challenged
• 19% failed
• 35% successful
•“ Fully one in six of the projects we studied was a black swan, with a cost overrun of 200%on average, and a schedule overrun of almost 70%”
~$875 billion spent on IT~$300 billion spent on IT projects ~$57 billion wasted annually
IT Has Similar Failures
Using Gates and Refining Estimates is a Best Practice (Adapted from K. Aguanno)
GreatIdea
Opportunity Analysis
Preliminary Business
Case
Committed Business
Case
AchieveBusiness
Case
Concept
- Describeidea &Possiblebenefits
MarketingAnalysis
FeasibilityStudy
Pilot or Proof of Concept-Determine
customer acceptance
-Interview focus groups, etc.
- Design solution- Estimate cost / schedule- Analyze risk- Determine feasibility / ROI
- Validate & commit to design & approach
-Revised estimates & schedule
-Risk reduction
-Baselined plan
-Build solution
- Deploy
-Achieve -business case
-Capture lessons learned -Including estimating
Full Execution or Deployment
Gate 1 Gate 2 Gate 3 Gate 4
Investment in time and money
© 2012 Copyright Galorath Incorporated 15
Causes of Project FailureSource: POST Report on UK Government IT Projects
Lack of a clear link between the project and the organisation’s key strategic priorities, including agreed measures of success.
1.
Lack of clear senior management and ministerial ownership and leadership 2.
Lack of effective engagement with Stakeholders 3.
Lack of skills and proven approach to project management and risk management
4.
Lack of understanding of and contact with the supply industry at senior levels within the organisation
5.
Evaluation of proposals driven by initial price rather than long-term value for money (especially securing the delivery of business benefits)
6.
Too little attention to breaking development and implementation into manageable steps
7.
Inadequate resources and skill to deliver the total delivery portfolio8.
Source: POST Report on UK Government IT Projects
© 2012 Copyright Galorath Incorporated 16
• June 28, 2010 Mandate
• September 14, 2010 Guidance
• November 3, 2010 Implementation• Five Specific Areas of Concern:
• Target Affordability and Control Cost Growth
• Reduce Non-Productive Processes and Bureaucracy
• Incentivize Productivity and Innovation in Industry
• Promote Real Competition
• Improve Tradecraft in Services Acquisition
US Better Buying Power Initiatives
Affordability Initiatives With “Should Cost” and “Will Cost”
Should Cost Performance
Cost Initiatives (Applied practices & improvements)
Will Cost Performance - =
© 2012 Copyright Galorath Incorporated 17
Characteristics of a Successful Should Cost Review (Source: AT Kearney)
© 2012 Copyright Galorath Incorporated 18
Key Points
Cost Estimation
with repeatable Process is
best practice
© 2012 Copyright Galorath Incorporated 19
© 2012 Copyright Galorath Incorporated 20
An Estimate Defined
• An estimate is the most knowledgeable statement you can make at a particular point in time regarding:
• Effort / Cost
• Schedule
• Staffing
• Risk
• Reliability
• Estimates more precise with progress
• A WELL FORMED ESTIMATE IS A DISTRIBUTION
20
© 2012 Copyright Galorath Incorporated 21
Viable Estimation Is Critical
• Estimating is critical for all kinds of systems
• Yet many treat is as a second rate process
• Everyone estimates…. Just most get it wrong and don’t have a process
• Having a repeatable estimation process is critical to both estimating AND to successful projects
• Estimation and measurement go hand in hand
© 2012 Copyright Galorath Incorporated 22
Estimation Methods 1 of 2
Model Category
Description Advantages Limitations
Guessing Off the cuff estimatesQuickCan obtain any answer desired
No Basis or substantiationNo ProcessUsually Wrong
AnalogyCompare project with past similar projects.
Estimates are based on actual experience.
Truly similar projects must exist
Expert Judgment
Consult with one or more experts.
Little or no historical data is needed; good for new or unique projects.
Experts tend to be biased; knowledge level is sometimes questionable; may not be consistent.
Top Down Estimation
A hierarchical decomposition of the system into progressively smaller components is used to estimate the size of a software component.
Provides an estimate linked to requirements and allows common libraries to size lower level components.
Need valid requirements. Difficult to track architecture; engineering bias may lead to underestimation.
© 2012 Copyright Galorath Incorporated 23
Estimation Methods 2 of 2
Model Category Description Advantages Limitations
Bottoms Up Estimation
Divide the problem into the lowest items. Estimate each item… sum the parts.
Complete WBS can be verified.
The whole is generally bigger than the sum of the parts.
Costs occur in items that are not considered in the WBS.
Design To Cost
Uses expert judgment to determine how much functionality can be provided for given budget.
Easy to get under stakeholder number.
Little or no engineering basis.
Simple CER’s
Equation with one or more unknowns that provides cost / schedule estimate.
Some basis in data.
Simple relationships may not tell the whole story.Historical data may not tell the whole story.
Comprehensive Parametric Models
Perform overall estimate using design parameters and mathematical algorithms.
Models are usually fast and easy to use, and useful early in a program; they are also objective and repeatable.
Models can be inaccurate if not properly calibrated and validated; historical data may not be relevant to new programs; optimism in parameters may lead to underestimation.
23
“Best Value” Data Needs
(adapted from Boeing Value Front)
Models & Methods
Customer Need Priorities(Decision Analysis)
CustomerDesirability
Attribute Value
O/SDevProd
TOC Cost-Risk
Attribute Value
Distribution
UncertaintySimulation
© 2012 Copyright Galorath Incorporated 24
© 2012 Copyright Galorath Incorporated 25
Functional Focus Example: Ladies Purse
• Function ………………………………………... Hold stuff
• Cost…………………………………………… $400 at Nordstrom
• What else will perform the function?
• Paper bag - Cost = $0.05
• Go to plastic bag for more durability
• Cost = $0.10
• Add color………………………………………..Cost = $0.15
• Add strap.……………………………………….Cost = $0.25Misses one component of customer
satisfaction
© 2012 Copyright Galorath Incorporated 26
“Far Out” Higher TRL Level EstimationGoal: Better Cost For Highly Advanced Space Missions (15-20 Years in the Future)
Critical items at less than TRL 4…
Like asking Edison in 1876 “How much longer for the light bulb?”
•“Hard to say”
In 1879, once he had found a workable carbon filament, “How much will a production version of the light bulb cost to develop and produce Tom?”
•Then a TRL 4 question
TRL
Desiredcapability
Limits ofpotentialimpacts
1 2 3 4 5 6 7 8 9
A
ImpactAt TRL
1
ImpactAt TRL
3
ImpactAt TRL
7Early impactshave a muchgreater impacton the finalsystem
B
C
TRL 9
TRL 8
TRL 7
TTRRLL 66
TTRRLL 55
TRL 4
TRL 3
TRL 2
TRL 1
TRL9: Actual system “flight proven” thorough successful mission operations TRL8: Actual system completed and “flight qualified” through test and demonstration TRL7: System prototype demonstration in a space environment TRL6: System/subsystem model or prototype demonstration in a relevant environment TRL5: Component and/or breadboard validation in relevant environment TRL4: Component and/or breadboard validation in laboratory environment TRL3: Analytical and experimental critical function and/or characteristic proof-of-concept TRL2: Technology concept and/or application formulated TRL1: Basic principles observed and reported
Proposed Hyperspectral Imaging Satellite predicted fielding: 2016
This capability would be of interest to:• Military space asset planners • Government agencies• Commercial satellite producers• Advanced concept designers
© 2012 Copyright Galorath Incorporated 27
Black Swans (Unknowns)
• The match is unlikely to ever be perfect because some projects are affected by “unknown unknowns,” also called Black Swans, that is, events that are essentially unpredictable (e.g., a severe worldwide credit squeeze)
Score
Score
Score
© 2012 Copyright Galorath Incorporated 28
Dealing With the “Problem of Assumptions”
• Assumptions are essential but…
• Incorrect assumptions can drive an estimate to uselessness
• Use an assumption verification process
Estimates must have assumptions defined, but…Bad assumptions should not be justification for
bad estimates
© 2012 Copyright Galorath Incorporated 29
System Description (Parametrics Can Estimate More, Earlier) Adapted from CEBOK
“If you can’t tell me what it is, I can’t tell you what it costs.”
-Mike Jeffers
“If you can tell me the range of what it might be, I can tell you the range of cost, schedule &
probability.”-Dan Galorath
© 2012 Copyright Galorath Incorporated 30
Uncertainty in the Cost Depends On Uncertainty of the Project Itself
Even though the entire project may be highly uncertain, tasks to the next gate should be estimatible within 10%.
SEER includes uncertainty in its estimates
Within 10%
31
Statistician Drowns in River with Average Depth of 3 Feet!
© 2012 Copyright Galorath Incorporated 31
Range vs. Point Estimates (Source US Army)
Range o
f R
isk
& U
nce
rtain
ty
Estimating Accuracy Trumpet
Technical and Program Maturity
RO
M -
30%
to
+75
%
Par
amet
ric
-1
0% t
o +
20%
An
alo
gy
-15
% t
o +
30%
En
gin
eeri
ng
-5
% t
o +
15%
Actu
al
-3%
to
+1
0%
Target Cost
Point estimate is most likely within range estimate with higher potential for cost
increase
Point estimate is most likely within range estimate with higher potential for cost
increase
Range estimate provides a degree of risk and uncertainty
Range estimate provides a degree of risk and uncertainty
Upper bound
Lower
bound
+75%
-30%
BAMaterielSolutionAnalysis
Systems Acquisition Sustainment
Technology Development
Production & Deployment
Operations & Support
Engineering and Manufacturing Development
C
Pre-Systems Acquisition© 2012 Copyright Galorath Incorporated 32
© 2012 Copyright Galorath Incorporated 3333
What is likely to happen
Feel lucky?
Firm Fixed Price?
Understand the risk before you commit!
© 2012 Copyright Galorath Incorporated 34
Dealing With Early Estimates
• Give a range: but they will hear the bottom of the range
• Give a high probability number:
• Will still be low in some cases and may be high in many cases but consider it a probably “not to exceed”
• Sticker shock may be a problem
• Give a category rather than an number: e.g over $10m Cat 1; over $5m Cat 2; over $1m Cat 3, etc.
Stakeholders always remember the first number even when told it is preliminary.
Developers will be optimistic by nature unless the process is tempered.
© 2012 Copyright Galorath Incorporated 35
Potential Accuracy at Various Gates (Adapted From Canada Treasury)
http://www.tbs-sct.gc.ca/itp-pti/pog-spg/irh-mei/irh-mei03-eng.asp
Estimate Accuracy Is a Function of Input Information Quality.Estimates Can Be Much Closer than Shown IF Data Is Available.
© 2012 Copyright Galorath Incorporated 36
Converting Uncertainty to Risk Is A Best Practice• Many treat as synonymous but…
• In risk situations, probabilities assigned based on data
• In uncertainty situations, we may assign probabilities • But there is no data to back them up
• If probability of rain tomorrow is 50%, that’s a risk
• We have historical data & scientific analyses about rain which make it reasonable to estimate the probability
• If we say Probability of success in harnessing nuclear fusion for routine energy production within the next ten years is 50%, that’s just a guess about uncertainty—evidence for the assignment is lacking or very weak
If you toss this thumbtack, what is the probability it will land this way instead of on its back?
If you toss this thumbtack, what is the probability it will land this way instead of on its back?
• This chapter discusses a 1972 GAO report on cost estimating
• We reported that cost estimates were understated and causing unexpected cost growth
• Many of the factors causing this problem are still relevant today
• We also discuss a 12 step process for producing high quality cost estimates
GAO Publication: Characteristics of credible cost estimates and a reliable process for creating
them
© 2012 Copyright Galorath Incorporated 37
Generalized 10 Step System Estimation Process 2011
1. Establish Estimate Scope
2. Establish Technical Baseline, Ground
Rules, Assumptions
4. Refine Technical Baseline Into
Estimable Components
4. Collect data / estimation inputs
5. Estimate Baseline Cost, Schedule, Affordability Value
6. Validate Business Case Costs &
Benefits (go / no go)
6. Quantify Risks and Risk Analysis
8. Generate a Project Plan
9. Document Estimates and Lessons
Learned
10. Track Project Throughout
Development
© 2012 Copyright Galorath Incorporated 38
Contractors At Least Level 3 Would Be Acceptable
Why should we care? Maturity is related to estimate viability… Better estimation process more likely to
be successful in execution© 2012 Copyright Galorath Incorporated 39
Key Points
Viable Affordability
decisions yield project
achievements
40
© 2012 Copyright Galorath Incorporated 40
Requirements And Features
Analysis
Preliminary Design
Process
•Human Factors•Security•Reliability•Availability•Survivability
•Supportability•Testability•Producibility•Reuse•Transportability
Other Factors
Alternative 1Alternative 2Alternative 3Alternative 4
SEER Assessment
•Performance•Schedule•Risk Assessment•Life Cycle Cost
1
23
4 5
SelectionProcess
OK?Optimized?
Yes
CostPerformance ScheduleRisk
Bottoms Up Estimation as Required
NoIterate
Should Cost: Trade Study FlowShould Cost: Trade Study Flow
© 2012 Copyright Galorath Incorporated 42
© 2012 Copyright Galorath Incorporated 43
Manual Estimates: Human Reasons For Error (Metrics Can Help)
• Manual Task estimates yield SIGNIFICANT error without ranges
• Desire for “credibility” motivates overestimate behavior (80% probability?)
• So must spend all the time to be “reliable”
• Best practice approach force 50% probability & have “buffer” for overruns
• Technical pride sometimes causes underestimates
43
Balancing Resources & Schedule Is A Best Practice
For a given Size, Complexity and Technology
Minimum Time
Optimal Effort(Lower Effort for Longer Schedule)
Calendar Time
Eff
ort
Mon
ths
Effort Increasedue to Longer Schedule
© 2012 Copyright Galorath Incorporated 44
0 4 8 12 16 20
Schedule ProbabilityExample Application 1
Probability
Time (calendar months)
1%10%20%30%40%50%60%70%80%90%99%
Understand Project Risks Include Them In Planning Decisions (Example SEER-SEM Outputs)
0 1800 3600 5400 7200 9000
Effort (person-hours)
1%10%20%30%40%50%60%70%80%90%99%
Effort ProbabilityExample Application 1
Probability
0 12 24 36 48 60
Defects ProbabilityExample Application 1Probability
Defects (count)
1%10%20%30%40%50%60%70%80%90%99%
• © 2012 Copyright Galorath Incorporated 45
Understanding & Tracking Defects, Growth And Other Metrics
Health and Status Indicator shows status
and trends from the previous snapshot
•Including Size Growth and Defect Discovery/Removal
Rate
•User defined control limits to control the transition
between red-yellowyellow-green
Track defect
discovery and
removal rates
against expected
rates
Increased defect
reporting rate shows a
worsening trend
Track software
size growth © 2012 Copyright Galorath Incorporated 46
© 2012 Copyright Galorath Incorporated 47
Goal Question Metric Approach Best Practice
• Combine goal-orientation bottoms up, decision-support & other operational management techniques
• www.weather.com to decide to bring an umbrella is decision support
Goal
Question
Development
Contractors
Metric
Organizational Goal
Question
Development
Organizations
Metric
© 2012 Copyright Galorath Incorporated 48
Reasons Many Don’t Want To Provide Data
• They could be proven wrong
• It could be used against them
• Data often doesn’t exist
• Even if processes dictate data requirements
• If it exists, it may not be clean
• It may give away corporate productivity & bid strategy
© 2012 Copyright Galorath Incorporated 49
Data Must Be Used With Caution
• Run sanity checks on data
• A million lines of code can’t be developed in 3 months
• Ongoing issue between our statisticians and engineers
• Some Statisticians claim.. “That is what the data says, so it must be right”
• Sometimes even if it is obviously wrong
© 2012 Copyright Galorath Incorporated 50
Data Doesn’t Have To Be Perfect To Be Useful: But Is Has To Be Viable
• 80 Calories per serving
• 2.5 Servings per can
• 4 Ounces, Condensed, 8 Ounces With Water
© 2012 Copyright Galorath Incorporated 52
• Correlation does not imply causation• Just because two data points may sit side by side
doesn’t mean they are the same or will have the same outcome
• Casual analysis is a recognized error in medicine
Perhaps ???
The Error of Causal AnalysisCreating a False Association
Tumor Can Cause Headache
Headache doesn’t mean a tumor
Use Historical Measurement to Evaluate Your Estimate!
It’s easy to dig deeper and deeper to justify an estimate!© 2012 Copyright Galorath Incorporated 53
© 2012 Copyright Galorath Incorporated 54
Estimation Best Practices
• Decide Why You Want An Estimate
• Map Estimation Goals To Estimate Process Maturity & Develop Plan To Achieve The Maturity
• Have A Documented, Repeatable Estimation Process
• Make The Estimating Process As Simple As Possible; But No Simpler
• Be Proactive: The Process Is Important, The Tools Go Along With The Process
• Get Buy-in From Program Managers
• Hold People Accountable: Center Of Excellence Can Prepare Estimate But Program Managers Must Own Them
• Tie The Estimate To The Plan
© 2012 Copyright Galorath Incorporated 55
Estimation Best Practices 2
• Evaluate Total Ownership Cost; Not Just Development
• Estimate A Range And Pick A Point For The Plan
• Re-estimate The Program When It Changes
• Avoid Death Marches: Programs With Unachievable Schedules Are Likely To Fail And Drain Morale
• Keep A History: Start An Enterprise Database NOW…
• Business Case: Evaluate ROI In Addition To Costs
• Convert Expert Spreadsheets Into A Common Language
© 2012 Copyright Galorath Incorporated 56
Estimation Best Practices 3
• Track Progress Vs. Estimate Throughout The Life Cycle
• Estimate Schedule As Well As Effort (Cost) For Complete Picture
• Tie The Business Case Into The Estimating Process
• Attack Non-productive Rework As Part Of The Process
© 2012 Copyright Galorath Incorporated 57
Estimation Best Practices 4
• Have clear definitions: What does “complete” mean? What activities are included and excluded (E.g. development only or total ownership; help desk included or excluded, etc.) Which labor categories are included and excluded in the estimate (e.g. are managers included? Help desk? Etc.)
• Measure what you care about
• Estimating & tracking rework can help control costs
• Don’t ignore IT infrastructure and IT services costs
• Tracking defect sources can go along with the process
© 2012 Copyright Galorath Incorporated 58
Conclusions
• Cost estimation and analysis are VITAL core processes
• Best practices ferret out what the cost is REALLY anticipated to be
• Risk and uncertainty must be taken into account
• Best practice project management understands the difference and acts to reduce uncertainty, or convert it to risk
• Applying affordability analysis to the business case yields the best value
© 2012 Copyright Galorath Incorporated 59
Additional Information
• www.galorath.com
• Dan on estimating BLOG: www.galorath.com/wp
• Email: [email protected]
• Phone: +1 310 414-3222 x614