Predicting
Predicting Project
Project Completion
Completion Date
Date
Using
Using Earned
Earned Value
Value Management
Management
NASA PM Challenge 2006
March 21-22
Galveston
Ray W. Stratton, PMP, EVP
714-318-2231
raystratton@ mgmt-technologies.com
© MMVI, Management Technologies
1
Presentation Outline
Refresher on EVM concepts
The Problem with SPI
The Earned Schedule Concept
Comprehensive Example
Research Results
Earned Schedule Maturity
Conclusions
References and Contact Information
© MMVI, Management Technologies
2
What Could be New About EVM?
In the 1900’s.....
EVM used in DoD, NASA, DoE, DoT, elsewhere
Provided Cost and Schedule Variances (CV, SV)
Computed Cost and Schedule Performance Indices
(CPI, SPI$)
Produced “Estimated Cost at Complete” (EAC)
21st Century Improvements to EVM
Provides SV in time units
Computes SPIt based on time units
Produces “Estimated Date of Completion”
© MMVI, Management Technologies
3
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
Refresher on
Earned Value Concepts
How earned value management is different
from cost reports and schedule updates.
© MMVI, Management Technologies
4
Basic EVM Concepts
Relate the money spent to the work done
Combines schedule status with cost information
Requires common “units” of measure
Fundamental concepts
The work has value equal to its budget
The project “equity” is the sum of the value of all
work done to-date
Building the equity (Earned Value) should equal the
budget planned for the work done (spend
appropriately)
The equity should accumulate at the rate planned
(stay on schedule)
© MMVI, Management Technologies
5
The Three Key Elements of EVM
Only three terms (two we already have)
Planned Value (PV)
Actual Cost (AC)
Earned Value (EV)
© MMVI, Management Technologies
6
The Planned Value (PV)
The schedule (plan) to build project equity from zero to
the total project value
The expected gain in project equity as each piece of
work is completed
Previously “budgeted cost of work scheduled” (BCWS)
Value
($ or hours)
© MMVI, Management Technologies
Planned
Value
Performance
Measurement
Baseline (PMB)
Time
7
The Earned Value (EV)
The project equity gained, regardless of the cost to
accomplish the work.
The gain in project equity as a result of completed work
Previously “budgeted cost of work performed” (BCWP)
Now
Value
($ or hours)
Planned
Value
Earned
Value
Time
© MMVI, Management Technologies
8
The Actual Cost (AC)
The total cost of the completed work
Previously “actual cost of work performed” (ACWP)
Now
Value
($ or hours)
Planned
Value
Actual
Costs
Earned
Value
Time
© MMVI, Management Technologies
9
Analyzing Variances From the Plan
Now
Value
($ or hours)
Planned
Value
Time
Actual
Costs
© MMVI, Management Technologies
Schedule
Variance
{}
Cost
Variance
Earned
Value
10
Cost Performance Index (CPI)
“Did the cost of adding “equity” equal the “equity”
added?”
CPI = progress/cost
CPI = EV/AC
Greater than 1.0 is good, less than 1.0 is bad
Example
CPI =
EV
WORK _ DONE
1000 hrs
=
=
= 0.66
AC COST _ OF _ WORK _ DONE 1500 hrs
(this is not good)
© MMVI, Management Technologies
11
Schedule Performance Index (SPI$)
“Did a month’s of schedule produce a month of
progress?”
SPI$ = progress/plan
SPI$ = EV/PV
Greater than 1.0 is good, less than 1.0 is bad
Example
SPI $ =
EV
WORK _ DONE
1200 hrs
=
=
= 1 .2
PV WORK _ PLANNED 1000 hrs
(this is good)
© MMVI, Management Technologies
12
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
The Problem with SPI
We’ve been using SPI for 30+ Years,
What could be wrong with it?
© MMVI, Management Technologies
13
Two Serious Problems with SPI$
SPI$ fails in the end
SPI$ = EV/PV
At project completion EV = PV
SPI$ = EV/PV = PV/PV = 1.0 at project end
Implications:
At project start SPI is reliable
At some point SPI accuracy diminishes
Toward the project end it is useless
SPI$ does not show weeks or months of schedule
variance
What most people want to know and understand
© MMVI, Management Technologies
14
Graphing the Problem with SPI$
1.2
Project schedule
improvement or
false indications
due to SPI?
1.1
3 months
TIME
SPI$ 1.0
SPI$
.9
.8
© MMVI, Management Technologies
Original
Project
Finish
Actual
Project
Finish
15
Using CPI and SPI$ in
Estimating the Final Cost
Simple extrapolation
Final cost = cost of work done + (work left/past
performance)
Assumes future performance is same as past (generally
true)
EAC (CPI) = AC+(BAC-EV)/CPI = BAC/CPI
EAC (CPI, SPI$) = AC+(BAC-EV)/(CPI*SPI$)
Toward the end
of the project
how accurate
can this term be?
© MMVI, Management Technologies
16
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
The Earned Schedule Concept
Creating a Schedule Performance Index that
is accurate to the end of the project.
© MMVI, Management Technologies
17
At What Point Did I Expect to Attain my
Current Earned Value?
AC is not shown in this
example since this is
about Earned Schedule!
Budget
or
When work
was to be
done (PV)!
staff hours
Work done
(EV)
6 months
Time
10 months
Actual Time (AT) i.e., NOW
Earned Schedule (ES) i.e.,
How much have you accomplished and
when were supposed to?
© MMVI, Management Technologies
18
Some New Terms
AT = Actual Time (now, or data date)
ES = Earned Schedule (Point in time when current
progress was planned to occur)
PD = Planned Duration (planned project duration)
ED = Estimated Duration (estimated project duration)
PCD = Planned Completion Date (Planned project end
date)
ECD = Estimated Completion Date (Estimated project
end date)
© MMVI, Management Technologies
19
Some New Formulas
SVt = ES - AT (Schedule Variance in time units, note
subscript versus SV$)
SVt = 6 months – 10 months = -4 months
SPIt = ES/AT (SPI in time units)
SPIt = 6/10 = 0.6
© MMVI, Management Technologies
20
The Traditional and the New
Planned
Budget
(BAC)
Budget
Spent
BAC
Planned work
(PV)
Funds spent
(AC)
Schedule
Variance (SV$)
Cost
Variance
(CV)
Work done
(EV)
Work
Done
Planned Duration (PD)
Actual Time (AT)
Earned Schedule (ES)
© MMVI, Management Technologies
Schedule
Variance (SVt)
Planned
Completion
Date
21
Traditional and New EVM Analysis Tools
Traditional
EV
SPI $ =
PV
EV
CPI =
AC
EAC =
BAC
CPI
SV$ = EV − PV
New
ES
SPI t =
AT
ED =
PD
SPI t
ECD = start _ date + ED
SVt = ES − AT
CV = EV − AC
© MMVI, Management Technologies
22
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
Comprehensive Example
Putting all the pieces together.
© MMVI, Management Technologies
23
Comprehensive Example
Planned
Budget
(BAC)
Budget
Spent
Cost
Variance
(CV)
12 BAC
Planned work
(PV)
7 PV
Funds spent
(AC)
5 AC
4 EV
Work done
(EV)
Work
Done
6 ES
10 AT
18 PD
Planned Duration (PD)
Actual Time (AT)
Earned Schedule (ES)
© MMVI, Management Technologies
Schedule
Variance (SVt)
24
Comparing EV and ES Schedule Analysis
Traditional
EV $4
SPI $ =
=
= 0.57
PV $7
SV$ = EV − PV = $4 − $7 = −$3
ED = ?
© MMVI, Management Technologies
New
ES
6months
SPI t =
=
= 0.6
AT 10months
SVt = ES − AT = 6months − 10months
SVt = −4months
PD 18months
ED =
=
= 30months
SPI t
0.6
25
New Analysis Products
Schedule performance index {SPI(t) or SPIt}
Schedule Variance {SV(t) or SVt}
Estimated Duration (ED)
Estimated Completion Date (ECD)
© MMVI, Management Technologies
26
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
Research Results
Does Earned Schedule really work?
© MMVI, Management Technologies
27
Empirical Evidence
(Henderson)
Findings derived from a small set of real life project data
Commercial IT Infrastructure Expansion Project Phase 1
Earned Schedule, Independent Estimate At Completion (time) - IEAC(t)
as at Project Completion: Week Starting 15th July xx
Planned Schedule
Earned Schedule cum
IEAC(t) PD/SPI(t)
40
35
Duration (Weeks)
30
25
20
15
10
5
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Actual Time (Weeks)
Henderson, K., 2003, “Earned schedule: a breakthrough extension to earned value theory? A retrospective
© MMVI, Management Technologiesanalysis of real project data”, The Measurable News, Summer 2003, 21, 13-17
28
Academic Research
(Vanhoucke)
Created 3,100 project activity networks
Constructed a schedule for each network
Simulated (executed) activity duration and cost in a
controlled way (9 execution scenarios, each schedule
subjected to each scenario 100 times)
Monitored and forecasted at each reporting period
Analysed the data (3.100 x 9 x 100 = 2.790.000 project
data sets)
Vanhoucke M., Vandevoorde St., A Simulation and Evaluation of Earned Value
Metrics to Forecast the Project Duration, Ghent University, Working Paper
2005/312, June 2005
© MMVI, Management Technologies
29
The Nine Simulation Scenarios
Critical Path Activities
Real Duration < Planned
Duration
Real Duration = Planned
Duration
Real Duration > Planned
Duration
Non-Critical Path Activities
Real Duration < Planned
Duration
Real Duration = Planned
Duration
Real Duration > Planned
Duration
© MMVI, Management Technologies
30
Research Process
© MMVI, Management Technologies
31
Academic Research Results
“The results reveal that the Earned Schedule method
outperforms, on the average, all other forecasting
methods.”
Vanhoucke M., Vandevoorde St., A Simulation and Evaluation of Earned Value
Metrics to Forecast the Project Duration, Ghent University, Working Paper
2005/312, June 2005
© MMVI, Management Technologies
32
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
Earned Schedule Maturity
Current state of Earned Schedule
development and deployment
© MMVI, Management Technologies
33
Earned Schedule Maturity
First paper: March 2003, PMI-CPM Measurable News
Over twenty presentations and publications
A growing trend and management expectation
Yet to appear in commercial EVM software products
Early Adopters
USAF
Lockheed Martin
Boeing
Belgium
Management Technologies (education, processes)
© MMVI, Management Technologies
34
USAF Planning
Air Force understands the potential of Earned Schedule
Demonstrate Earned Schedule on a sample set of
programs
Build a body of evidence on a larger set of programs
Validate the Earned Schedule as a value added program
management tool
Integrate Earned Schedule as a tool in Air Force
Acquisition
© MMVI, Management Technologies
35
Lockheed Martin, Program #1
PD
PD
IEAC(t)
IEAC(t)
Linear (IEAC(t))
Linear (IEAC(t))
IEAC(t) = (PD ÷ SPIt)
55
55
50
50
PD = 48
45
PD = 54
45
1
1
2
2 3
3
4
4
5
6
5 6
© MMVI, Management Technologies
7
8
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
36
Lockheed Martin, Program #2
IEAC(t)
Linear (IEAC(t))
50
IEAC(t) = (PD ÷ SPIt)
45
40
PD = 39
AD = 42
35
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
© MMVI, Management Technologies
37
The Good News!
Previously
EVM was great at cost management
EVM was “ok” at schedule management, sometimes
Now
EVM can address both cost and schedule well
EVM can estimate
Cost at completion
Completion date!
SPIt can record project outcome in lessons learned
© MMVI, Management Technologies
38
Credit to Discoverers, Advocates, &
Researchers
Henderson Kym, Earned Schedule: A Breakthrough Extension to
Earned Value Theory? A Retrospective Analysis of Real Project
Data,The Measurable News, Summer 2003
Robert Handshuh, New Concept in Earned Value - Earned
Schedule, PMI Southeast Regional Conference USA, June 2005
Walt Lipke & Kym Henderson, Earned Schedule Status Update
and Early Adaptor Applications Feedback, 17th IPMC November
2005
Vandevoorde St., Vanhoucke M., A Comparison of different
project duration forecasting methods using earned value
metrics, Ghent University, working paper 2005/312, June 2005
Approved for Publishing in “International Journal of Project
Management”
State of the Art Report on Forecasting Duration Methods
© MMVI, Management Technologies
39
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
Conclusions
So What?
© MMVI, Management Technologies
40
What We Know About Earned Schedule
Faithful to the project’s end
Provides a reliable means to calculate an estimated
completion date
Provides a useful schedule performance metric for
project history and lessons learned
May provide a better estimated completion date than
critical path methods!
© MMVI, Management Technologies
41
Why use ES?
More intuitive schedule information
Schedule variance in time units (intuitive to most people)
Cost variance in resource ($$) units
SPIt retains utility to project end
SPIt does not automatically creep toward 1.0 near
project end
SPIt captures the final project schedule performance
data
Can finally estimate Project Completion Date
© MMVI, Management Technologies
42
•Refresher on EVM concepts
•The Problem with SPI
•The Earned Schedule Concept
•Comprehensive Example
•Research Results
•Earned Schedule Maturity
•Conclusions
•References and Contact Information
Reference and Contact Information
Tell me more!
© MMVI, Management Technologies
43
Seminal Documents
Lipke, Walter, “Schedule is Different”, The Measurable
News, March and Summer 2003
Henderson, Kym, “Further Development in Earned
Schedule”, The Measurable News, Spring 2003
Henderson, Kym, Earned Schedule: A Breakthrough
Extension to Earned Value Theory? A Retrospective
Analysis of Real Project Data, The Measurable News,
Summer 2003
The Measurable news is a quarterly publication of
the Project Management Institute’s College of
Performance Management
www.pmi-cpm.org
© MMVI, Management Technologies
44
Other References
Earned Schedule Status Update and Early Adopter Applications Feedback
Presentation facilitated by Walt Lipke and Kym Henderson, 17th IPMC
November 2005
Earned Schedule Leads to Improved Forecasting:
Presentation by Walt Lipke, intended for 17th IPMC November 2005
Earned Schedule in Action:Paper by Kym Henderson, published PMI CPM
Journal, The Measurable News "Spring" 2005
Forecasting Project Schedule Completion by Using Earned Value Metrics,
Presentation by Ing. Stephan Vandevoorde, Senior Project Manager,
Fabricom Airport Systems and Prof. Dr. Mario Vanhoucke, Ghent University,
Belgium, Early Warning Signals Conference, V.U.B. Brussels, June 2005
Not Your Father's Earned Value: Earned Schedule overview paper by Ray
Stratton, published on Projects@Work (http://www.projectsatwork.com) 24
Feb 2005
Connecting Earned Value to the Schedule:Paper by Walt Lipke, published
PMI CPM Journal, The Measurable News, "Winter" 2004
© MMVI, Management Technologies
45
One-Stop Shopping
Comprehensive Library of EVM and Earned Schedule
Papers
Presentations
http://sydney.pmichapters-australia.org.au/
Click “Education,” then “Presentations and Papers” for .pdf copies
© MMVI, Management Technologies
46
In Closing ………
If you are not practicing Earned Schedule as part of your
Earned Value Management System ………
………you are practicing 20th Century
Earned Value Management!
© MMVI, Management Technologies
47
Discussion and Questions
Ray W. Stratton, EVP, PMP
President, Management Techologies
PMI CPM Vice President, Research and Standards
714-318-2231
raystratton@mgmt-technologies.com
www.mgmt-technologies.com
© MMVI, Management Technologies
48