Importance of Project Management
• Projects represent change and allow organizations to
effectively introduce new products, new
process, new programs
• Project management offers a means for dealing with
dramatically reduced product cycle times
• Projects are becoming globalized making them more
difficult to manage without a formal methodology
• Project management helps cross-functional teams to
be more effective
Management of IT Projects
• More than $250 billion is spent in the US each year on
approximately 175,000 information technology projects.
• Only 26 percent of these projects are completed on time and
within budget.
• The average cost for a development project for a large company
is more than $2 million.
• Project management is an $850 million industry and is expected
to grow by as much as 20 percent per year.
Bounds, Gene. “The Last Word on Project
Management” IIE Solutions, November, 1998.
What Defines a Project?
•
•
•
•
•
•
•
How does a project
differ from a
program?
Project Management versus Process Management
“Ultimately, the parallels between process and project
management give way to a fundamental difference:
process management seeks to eliminate variability
whereas project management must accept variability
because each project is unique.”
Elton, J. & J. Roe. “Bringing Discipline to Project
Management” Harvard Business Review, March-April,
1998.
Measures of Project Success
•
•
•
•
•
•
•
Was the movie
“Titanic”
a success?
Delayed Openings are a Fact of Life in the Foodservice,
Hospitality Industry
Disney's shipbuilder was six months late in delivering its new cruise ships,
and thousands of customers who had purchased tickets were stranded.
Even with that experience, their second ship was also delivered well after the
published schedules. Universal Studios in Orlando, Fla. had been building a
new restaurant and entertainment complex for more than two years. They
advertised a December opening, only to announce in late November that it
would be two or three months late.
Even when facilities do open close to schedule, they are rarely finished
completely and are often missing key components. Why do those things
happen? With all of the sophisticated computers and project management
software, why aren't projects completed on schedule?
Frable, F. Nation's Restaurant News (April 12, 1999)
IT Project Outcomes
More than 200%
late
101-200% late
6%
16%
51-100% late
21-50% late
9%
29%
Cancelled
8%
6%
Less than 20%
late
26%
On-Time
Source: Standish Group Survey, 1999 (from a
survey of 800 business systems projects)
Why do Projects Fail?
Studies have shown that the following factors
contribute significantly to project failure:
• Improper focus of the project management system
• Fixation on first estimates
• Wrong level of detail
• Lack of understanding about project management tools; too much
reliance on project management software
• Too many people
• Poor communication
• Rewarding the wrong actions
Why do IT Projects Fail?
• Ill-defined or changing requirements
• Poor project planning/management
• Uncontrolled quality problems
• Unrealistic expectations/inaccurate estimates
• Naive adoption of new technology
Source: S. McConnell, Construx Software Builders, Inc.
Have You Ever Lost Sight of the
Project Goals?
QuickTime™ and a
Photo - JPEG decompressor
are needed to see this picture.
Not all Projects Are Alike…
“[in IT projects], if you ask people what’s done and what remains to be
done there is nothing to see. In an IT project, you go from zero to 100
percent in the last second--unlike building a brick wall where you can see
when you’re halfway done. We’ve moved from physical to non-physical
deliverables….”
J. Vowler (March, 2001)
Engineering projects = task-centric
IT projects = resource-centric
Shenhar’s Taxonomy of Project Types
Degree of
Uncertainty/Risk
Super HighTech
ERP
implementation
in multi-national
firm
HighTech
New shrinkwrapped
software
MediumTech
LowTech
Advanced
radar
system
New
cellphone
Construction
Assembly
Projects
Auto repair
System
Projects
Array
Projects
System Complexity/Scope
High
Required Resources
Project Life Cycle
Phase 1
Phase 2
Phase 3
Formation &
Selection
Planning
Scheduling &
Control
Phase 4
Evaluation &
Termination
Time
Life Cycle Models: Pure Waterfall
Concept
Design
Requirements
Analysis
Architecture
Design
Detailed
Design
Coding &
Debugging
System
Testing
Source: S. McConnell
Rapid Development (Microsoft Press, 1996)
Life Cycle Models: Code & Fix
DESIGN
Design, Cost, Time Trade-offs
Required
Performance
Target
Budget
Constraint
Due Date
Optimal Time-Cost
Trade-off
COST
Optional Scope Contracts
Since it is widely accepted that you can select
three of the four dimensions (or perhaps only
two), what to do?
Fixed Scope Contract
specifies
Optional Scope Contract
specifies
SCHEDULE, COST, SCOPE
SCHEDULE, COST, QUALITY
(general design guidelines may be indicated)
Importance of Project Selection
“There are two ways for a business to succeed
at new products: doing projects right, and
doing the right projects.”
Cooper, R.G., S. Edgett, & E. Kleinschmidt.
Research • Technology Management, March-April, 2000.
Project Initiation & Selection
• Critical factors
1) Competitive necessity
2) Market expansion
3) Operating requirement
• Numerical Methods
1)
2)
3)
4)
Payback period
Net present value (NPV) or Discounted Cash Flow (DCF)
Internal rate of return (IRR)
Expected commercial value (ECV)
• Project Portfolio
1) Diversify portfolio to minimize risk
2) Cash flow considerations
3) Resource constraints
Payback Period
Number of years needed for project to
repay its initial fixed investment
Example: Project costs $100,000 and is expected
to save company $20,000 per year
Payback Period = $100,000 / $20,000 = 5 years
Net Present Value (NPV)
Discounted Cash Flow (DCF)
Let Ft = net cash flow in period t (t = 0, 1,..., T)
F0 = initial cash investment in time t = 0
r = discount rate of return (hurdle rate)
T
NPV =
•
t=0
Ft
1+rt
Internal Rate of Return (IRR)
Find value of r such that NPV is equal to 0
Example (with T = 2):
Find r such that
F0 + F1 + F2 = 0
1+r
1+r2
DCF Project Example*
Phase I
Research and Product Development
$18 million annual research cost for 2 years
60% probability of success
Phase II
Market Development
Undertaken only if product development is successful
$10 million annual expenditure for 2 years to develop marketing and
distribution channels (net of any revenues earned in test marketing)
Phase III
Sales
Proceeds only if Phase I and II verify opportunity.
Production is subcontracted and all cash flows are after-tax and occur
at year's end.
The results of Phase II (available at the end of year 4) identify the
product's market potential as indicated below:
Product
Demand
High
Medium
Low
Product Life
20 years
10 years
Abandon Project
Annual Net
Cash Inflow
$24 million
$12 million
None
Probability
0.3
0.5
0.2
*Hodder, J. and H.E. Riggs. “Pitfalls in Evaluating Risky Projects”, Harvard
Business Review, Jan-Feb, 1985, pp. 128-136.
DCF Project Example (cont’d)
Year
1
2
3
4
5 - 14
15 - 24
Expected Cash Flow (in $ million)
-18
-18
0.6 (-10) = - 6
0.6 (-10) = - 6
.6 (0.3 x 24 + 0.5 x 12) = 7.92
.6 (0.3 x 24) = 4.32
What is the internal rate of return for this project?
DCF Example Continued
What if you can sell the product (assuming that both Research and
Product Development AND Market Development are successful) to a
third party? What are the risks AT THAT POINT IN TIME?
Assume that discount rate r2 is 5%
Probability
What is 20 years of cash inflow at $24M/year?
What is 10 years of cash inflow at $12M/year?
$299.09
$92.66
Expected value of product at Year 4:
$136.06
0.3
0.5
DCF Example Continued
Expected cash flows (with sale of product at end of year 4) are now:
Year 1
Year 2
Year 3
Year 4
Outflow
$
18.00
$
18.00
$
10.00
$
10.00
Inflow
$
136.06
$
$
$
$
Net
(18.00)
(18.00)
(10.00)
126.06
Probability
1
1
0.6
0.6
Expected
Cas h Flow
$
$
$
$
What is the internal rate of return for this project?
(18.00)
(18.00)
(6.00)
75.63
Criticisms of NPV/DCF
1) Assumes that cash flow forecasts are accurate; ignores
the “human bias” effect
2) Fails to include effects of inflation in long term
projects
3) Ignores interaction with other proposed and ongoing
projects (minimize risk through diversification)
4) Use of a single discount rate for the entire project (risk
is typically reduced as the project evolves)
Expected Commercial Value (ECV)
Probability = pc
Commercial Success
(with net benefit =
NPV)
Probability = pt
Technical Success
Develop New
Product
Probability = 1 - pt
Launch New
Product
Probability = 1 - pc
Commercial
Failure (with net
benefit = 0)
Technical Failure
Risk class 1
Risk class 2
DCF Example Revisited
Product Demand
High
0.3
Probability = pt
Development
Succeeds
Research &
Product
Development
Market
Development
0.2
Probability = 1 - pt
Development Fails
Discount rate r1
0.5
Product Demand
Medium
Product Demand
Low
Drop project
Discount rate r2
Ranking/Scoring Models
Profit abilit y/value
1) Incr ease in p rofit abil ity?
2) Incr ease in ma rket sha re?
3) Will add know ledge to organ ization that can be leve raged by o ther projects?
4) Estim ated NPV, ECV, etc.
Organizat ion's Strat egy
1) Consistent wit h o rgan ization's mi ssion statement?
2) Impact on custome rs?
Risk
1) Probabilit y of research be ing succ essful?
2) Probabilit y of deve lopment being succe ssful?
3) Probabilit y of process succe ss?
4) Probabilit y of comm ercial succ ess?
5) Overall risk of project
6) Adequate market demand?
7) Competit ors in market
Organizat ion Costs
1) Is new facilit y needed?
2) Can use current personne l?
3) External con sultants needed?
4) New hires needed?
Miscellaneous Factors
1) Impact on env ir onmental standa rds?
2) Impact on workfo rce safety?
3) Impact on qua lit y?
4) Social/ polit ical im pli cations
Scoring Attributes
To convert various measurement scales to a (0, 1) range….
x -L
LINEAR SCALE: value of attribute i is vi xi = i
U- L
EXPONENTIAL SCALE: value of attribute i is vi xi =
1 - exp L - xi
.
1 - exp L - U
1.00
0.90
0.80
Attribute Value
0.70
0.60
Linear Scale
Exponential Scale
0.50
0.40
0.30
0.20
0.10
0.00
1
2
3
4
Response
5
6
7
Ranking/Scoring Example
Attribute
1) Does project increase market share?
Attribute
Weight (wi)
Measurement Scale
unlikely
1
2
2) Is new facility needed?
3
yes
3) Are there safety concerns?
likely
4
5
likely
15%
10%
no
unsure
30%
no
4) Likelihood of successful technical development?
unlikely
1
2
3
4
5
likely
20%
5) Likelihood of successful commercial development?
unlikely
1
2
3
4
5
likely
25%
Ranking/Scoring Example (cont’d)
Attribute
Project A
Project B
Project
Score (V j)
#1
#2
#3
#4
#5
4
2
yes
no
likely
unsure
4
3
1
4
0.75
0.25
0.25
0.75
0
0.5
0.75
0.5
0
0.75
0.413
0.525
0.97
0.64
0.64
0.97
0.00
0.88
0.97
0.88
0.00
0.97
0.581
0.845
Linear Scale
Project A
Project B
Exponential Scale
Project A
Project B
Analyzing Project Portfolios: Bubble Diagram
Prob of Commercial Success
High
Zero
High
Expected NPV
Low
Analyzing Project Portfolios: Product vs Process
Extent of Process Change
Source: Clark and Wheelwright, 1992
Key Elements of Project Portfolio Selection Problem
1. Multi-period investment problem
2. Top management typically allocates funds to different
product lines (e.g., compact cars, high-end sedans)
3. Product lines sell in separate (but not necessarily
independent) market segments
4. Product line allocations are changed frequently
5. Conditions in each market segment are uncertain from
period to period due to competition and changing
customer preferences
“Stage-Gate” Approach
Initiation
Define
Design
Initiation
Project Review
Charter
Work Statement
Risk Assessment
Purchasing Plan
Change Mgt
Detail Design
Schedule & Budget
Contingency Plan
Product &
Performance Reviews
Improve
Installation Plan
Facility Prep
Training Plan
Implementation
Control
Production close-out
Lessons learned
Post-project audit
Source: PACCAR Information Technology Division
Renton, WA
Project Selection Example
1
Y e a r (t)
2
3
4
Project A
($40)
$10
$20
$20
Project B
Budget
Limit (B t )
($65)
($25)
$50
$50
$120
$20
$40
$55
Phases of Project Management
n
n
Project formulation and selection
Project planning
u
u
u
u
u
n
Project scheduling
u
u
u
u
n
Summary statement
Work breakdown structure
Organization plan
risk management
Subcontracting and bidding process
Time and schedule
Project budget
Resource allocation
Equipment and material purchases
Monitoring and control
u
u
u
Cost control metrics
Change orders
Milestone reports
Project Planning
n
Summary Statement
u
u
u
u
u
n
Executive summary: mission and goals, constraints
Description and specifications of deliverables
Quality standards used (e.g., ISO)
Role of main contractor and subcontractors
Composition and responsibilities of project team
Organization Plan
u
u
u
u
u
u
Managerial responsibilities assigned; signature authority
Cross impact matrix (who works on what)
Relationship with functional departments
Project administration
Role of consultants
Communication procedures with organization, client, etc.
Importance of Project Planning
The 6P Rule of Project Management:
Prior Planning Prevents Poor Project
Performance
“If you fail to plan, you will plan to fail”
Anonymous
Work Breakdown Structure (WBS)
1) Specify the end-item “deliverables”
2) Subdivide the work, reducing the dollars and
complexity with each additional subdivision
3) Stop dividing when the tasks are manageable “work
packages” based on the following:
• Skill group(s) involved
• Managerial responsibility
• Length of time
• Value of task
Work Packages/Task Definition
The work packages (tasks or activities) that are defined
by the WBS must be:
• Manageable
• Independent
• Integratable
• Measurable
Design of a WBS
“The usual mistake PMs make is to lay out too many tasks;
subdividing the major achievements into smaller and
smaller subtasks until the work breakdown structure
(WBS) is a ‘to do’ list of one-hour chores. It’s easy to get
caught up in the idea that a project plan should detail
everything everybody is going to do on the project. This
springs from the screwy logic that a project manager’s job
is to walk around with a checklist of 17,432 items and tick
each item off as people complete them….”
The Hampton Group (1996)
Two-Level WBS
WBS level 1
WBS level 2
1.1 Event
Planning
1. Charity Auction
1.2 Item
Procurement
1.3 Marketing
1.4. Corporate
Sponsorships
Three-Level WBS
1. Charity Auction
WBS level 1
WBS level 2
1.1 Event
Planning
1.1.1 Hire Auctioneer
1.2 Item
Procurement
1.3 Marketing
1.4 Corporate
Sponsorships
1.2.1 Silent
auction items
1.3.1 Individual
ticket sales
1.2.2 Live auction
items
1.3.2 Advertising
1.1.2. Rent space
WBS level 3
1.1.3 Arrange for
decorations
1.2.3 Raffle items
1.1.4 Print catalog
Estimating Task Durations (cont’d)
• Benchmarking
• Modular approach
• Parametric techniques
• Learning effects
Beta Distribution
Probability density
function
Completion time of task j
Time
Optimistic Time toj
Expected duration = 
Most Likely Time = tm
Pessimistic Time tpj
Beta Distribution
For each task j, we must make three estimates:
toj most optimistic time
tpj most pessimistic time
tm
j most likely time
toj + tpj + 4tm
j
Expected duration j =
6
p
o2
t
j - tj
2
Variance of task j = j =
36
Estimating Task Durations: Painting a Room
Task: Paint 4 rooms, each is approximately 10’ x 20’. Use flat paint on walls,
semi-gloss paint on trim and woodwork. Each room has two doors and four
windows. You must apply masking tape before painting woodwork around the
doors and windows. Preparation consists of washing all walls and woodwork
(some sanding and other prep work will be needed). Only one coat of paint is
necessary to cover existing paint. All supplies will be provided at the start of the
task. Previous times on similar painting jobs are indicated in the table below.
hours
27
38
33
17
26
22
14
30
28
21
23
27
23
37
17
17
min
25
25
12
44
7
1
2
27
30
13
59
44
15
6
54
13
hours
31
19
26
30
25
24
32
32
13
42
22
32
32
27
26
21
min
52
15
27
27
21
28
58
1
43
45
57
15
31
15
11
52
What is your estimate of the average time you will
need? What is your estimate of the variance?
Estimating Task Durations with Incentives
Task: Consider the painting job that you have
just estimated. Now, however, there are
explicit incentives for meeting your estimated
times. If you finish painting the room before
your specified time, you will receive a $10
bonus payment. HOWEVER, if you finish
the painting job after your specified time, you
will be fined $1000.
Revised estimated time =
Estimating Task Durations with Incentives
Task: Consider the painting job that you have
just estimated. Now, however, there are
explicit incentives for meeting your estimated
times. If you finish painting the room before
your specified time, you will receive a $10
bonus payment. If you finish the painting job
after your specified time, there is no penalty.
Revised estimated time =
Role of Project Manager/Team
Client
Top
Management
Project Manager
Subcontractors
Project Team
Regulating
Organizations
Functional
Managers
Responsibilities of a Project Manager
To the organization and top management
• Meet budget and resource constraints
• Engage functional managers
To the project team
• Provide timely and accurate feedback
• Keep focus on project goals
• Manage personnel changes
To the client
• Communicate in timely and accurate manner
• Provide information and control on changes/modifications
• Maintain quality standards
To the subcontractors
• Provide information on overall project status
Project Team
What is a project team?
A group of people committed to achieve a
common set of goals for which they hold
themselves mutually accountable
Characteristics of a project team
•
•
•
•
Diverse backgrounds/skills
Able to work together effectively/develop synergy
Usually small number of people
Have sense of accountability as a unit
“I design user interfaces to please an audience of one.
I write them for me. If I’m happy, I know some cool
people will like it. Designing user interfaces by
committee does not work very well; they need to be
coherent. As for schedule, I’m not interested in
schedules; did anyone care when War and Peace came
out?”
Developer, Microsoft Corporation
As reported by MacCormack and Herman, HBR Case 9-600-097:
Microsoft Office 2000
Intra-team Communication
M = Number of project team members
L = Number of links between pairs of team members
If M =2, then L = 1
If M =3, then L = 3
Number of Intra-team Links
L = Number of Intra-team Links = N
2
=
N(N-1)
2
Importance of Communication
On the occasion of a migration from the east, men discovered a
plain in the land of Shinar, and … said to one another, “Come, let
us build ourselves a city with a tower whose top shall reach the
heavens….” The Lord said, …“Come, let us go down, and there
make such a babble of their language that they will not
understand one another’s speech.” Thus, the Lord dispersed
them from there all over the earth, so that they had to stop
building the city.
Genesis 11: 1-8
Project Performance and Group Harmony
What is the relationship between the design of
multidisciplinary project teams and project success?
Two schools of thought:
1) “Humanistic school” -- groups that have positive
characteristics will perform well
2) “Task oriented” school -- positive group
characteristics detract from group performance
Project Performance and Group Harmony (cont’d)
Experiment conducted using MBA students at UW and
Seattle U using computer based simulation of pre-operational
testing phase of nuclear power plant*
Total of 14 project teams (2 - 4 person project teams) with a
total of 44 team members; compared high performance (low
cost) teams vs low performance (high cost) teams
Measured:
Group Harmony
Group Decision Making Effectiveness
Extent of Individual’s Contributions to Group
Individual Attributes
*Brown, K., T.D. Klastorin, & J. Valluzzi. “Project Management
Performance: A Comparison of Team Characteristics”, IEEE Transactions on
Engineering Management, Vol 37, No. 2 (May, 1990), pp. 117-125.
Group Harmony: High vs Low Performing Groups
Extent of Individual Contribution: High vs Low
Performing Groups
Decision Making Effectiveness: High vs Low
Performing Groups
Project Organization Types
• Functional: Project is divided and assigned to appropriate functional
entities with the coordination of the project being carried out by
functional and high-level managers
• Functional matrix: Person is designated to oversee the project
across different functional areas
• Balanced matrix: Person is assigned to oversee the project and
interacts on equal basis with functional managers
• Project matrix: A manager is assigned to oversee the project and is
responsible for the completion of the project
• Project team: A manager is put in charge of a core group of
personnel from several functional areas who are assigned to the
project on a full-time basis
Project Organization Continuum
Functional Matrix
Functional
Organization
Project fully managed
by functional managers
Project Matrix
Balanced Matrix
Project Team
Organization
Project fully managed by
project team manager
A Business School as a Matrix Organization
Dean
Associate Dean for
Undergraduate
Program
Associate Dean for
MBA Programs
Director of
Doctoral Program
Accounting
Department Chair
Larry
Zelda
Diane
Marketing
Department Chair
Curly
Bob
Barby
Finance Department
Chair
Moe
Gloria
Leslie
Matrix Organizations & Project Success
• Matrix
organizations emerged in 1960’s as an
alternative to traditional means of project
teams
• Became
• Still
•
popular in 1970’s and early 1980’s
in use but have evolved into many different
forms
Basic question: Does organizational structure
impact probability of project success?
Organizational Structure & Project Success
• Studies by Larson and Gobeli (1988, 1989)
• Sent questionnaires to 855 randomly selected PMI members
• Asked about organizational structure (which one best describes the primary
structure used to complete the project)
• Perceptual measures of project success: successful, marginal, unsuccessful
with respect to :
1) Meeting schedule
2) Controlling cost
3) Technical performance
4) Overall performance
• Respondents were asked to indicate the extent to which they agreed with
each of the following statements:
1) Project objectives were clearly defined
2) Project was complex
3) Project required no new technologies
4) Project had high priority within organization
Study Data
• Classification of 547 respondents (64% response rate)
30% project managers or directors of project mgt programs
16% top management (president, vice president, etc.)
26% managers in functional areas (e.g., marketing)
18% specialists working on projects
• Industries included in studies
14% pharmaceutical products
10% aerospace
10% computer and data processing products
others: telecommunications, medical instruments, glass products,
software development, petrochemical products, houseware goods
• Organizational structures:
13% (71): Functional organizations
26% (142): Functional matrix
16.5% (90): Balanced matrix
28.5% (156): Project matrix
16% (87): Project team
ANOVA Results by Organizational Structure
N
Controlling
Cos t
Ave (SD)
Meeting
Schedule
Ave (SD)
Technical
Performance
Ave (SD)
Overall
Results
Ave (SD)
A
Functional
Organization
71
1.76 (.83)
1.77 (.83)
2.30 (.77)
1.96 (.84)
B
Functional Matrix
142
1.91 (.77)
2.00 (.85)
2.37 (.73)
2.21 (.75)
C
Balanced Matrix
90
2.39 (.73)
2.15 (.82)
2.64 (.61)
2.52 (.61)
D
Project Matrix
156
2.64 (.76)
2.30 (.79)
2.67 (.57)
2.54 (.66)
E
Project Team
87
2.22 (.82)
2.32 (.80)
2.64 (.61)
2.52 (.70)
Total Sample
546
2.12 (.79)
2.14 (.83)
2.53 (.66)
2.38 (.70)
F-statistic
10.38*
6.94*
7.42*
11.45*
Scheffe Results
A,B < C,D,E
E<D
Organizational Structure
*Statistically significant at a p<0.01 level
A,B < C < D,E A,B < C,D,E
A,B < C,D,E
Summary of Results
• Project structure significantly related to project success
• New development projects that used traditional functional organization
had lowest level of success in controlling cost, meeting schedule,
achieving technical performance, and overall results
• Projects using either a functional organization or a functional matrix had
a significantly lower success rate than the other three structures
• Projects using either a project matrix or a project team were more
successful in meeting their schedules than the balanced matrix
• Project matrix was better able to control costs than project team
• Overall, the most successful projects used a balanced matrix, project
team, or--especially--project matrix
Subcontracting = Business Alliance
n
When you subcontract part (or all) of a
project, you are forming a business
alliance....
Intelligent Business Alliances: “A business relationship for
mutual benefit between two or more parties with compatible
or complementary business interests and/or goals”
Larraine Segil, Lared Presentations
Communication and Subcontractors
What types of communication mechanism(s) will be
used between company and subcontractor(s)?
WHAT a company
communicates.....
HOW a company
communicates.....
How is knowledge
transferred?
Personality Compatibility
Subcontractor
Personality
Corporate
Personality
Project
Individual
Personality
Subcontracting Issues
• What part of project will be subcontracted?
n• What type of bidding process will be used? What type of
contract?
n• Should you use a separate RFB (Request for Bids) for
each task or use one RFB for all tasks?
n• What is the impact on expected duration of project?
n• Use a pre-qualification list?
n• Incentives? Bonus for finishing early? Penalties for
finishing after stated due date?
• What is impact of risk on expected project cost?
n
Basic Contract Types
n
Fixed Price Contract
u
n
Cost Plus Contract
u
n
Client pays a fixed price to the contractor irrespective of actual audited
cost of project
Client reimburses contractor for all audited costs of project (labor, plant,
& materials) plus additional fee (that may be fixed sum or percent of costs
incurred)
Units Contract
u
Client commits to a fixed price for a pre-specified unit of work; final
payment is based on number of units produced
Incentive (Risk Sharing) Contracts
General Form:
Payment to Subcontractor = Fixed Fee + (1 - B) (Project Cost)
where B = cost sharing rate
Cost Plus Contract
B=0
Fixed Price Contract
Linear & Signalling
Contracts
B=1
Why Use Incentive Contracts?
Expected Cost of Project = $100M
Two firms bid on subcontract
Firm 1
Firm 2
Fixed Fee (bid)
$5 M
$7 M
Project Cost
$105 M
$95 M
(inefficient producer)
What is result if Cost Plus Contract (B = 0) used?
Washington State Bid Code (WAC 236-48-093)
n
n
n
n
n
n
n
n
n
WAC 236-48-093: A contract shall be awarded to the lowest responsible and responsive
bidder based upon, but not limited to, the following criteria where applicable and only
that which can be reasonably determined:
1) The price and effect of term discounts...price may be determined by life cycle costing
if so indicated in the invitation to bid
2) The conformity of the goods and/or services bid with invitation for bid or request for
quotation specifications depicting the quality and the purposes for which they are
required.
3) The ability, capacity, and skill of the bidder to perform the contract or provide the
services required.
4) The character, integrity, reputation, judgement, experience, and efficiency of the
bidder.
5) Whether the bidder can perform the contract with the time specified.
6) The quality of performance on previous contracts for purchased goods or services.
7) The previous and existing compliance by the bidder with the laws relating to the
contract for goods and services.
8) Servicing resources, capability, and capacity.
Competitive Bidding: Low-Bid System
n
“In the low-bid system, the owner wants the most
building for the least money, while the contractor
wants the least building for the most money. The
two sides are in basic conflict.”
Steven Goldblatt
Department of Building Construction
University of Washington
The Seattle Times, Nov 1, 1987
Precedence Networks
Networks represent immediate precedence relationships
among tasks (also known as work packages or activities)
and milestones identified by the WBS
Milestones (tasks that take no time and cost $0 but indicate
significant events in the life of the project)
Two types of networks: Activity-on-Node (AON)
Activity-on-Arc (AOA)
All networks: must have only one (1) starting and one (1)
ending point
Precedence Networks: Activity-on-Node (AON)
A
C
Start
End
B
D
Precedence Diagramming
Standard precedence network (either AOA or AON) assumes that a successor
task cannot start until the predecessor(s) task(s) have been completed.
Alternative relationships can be specified in many software packages:
Finish-to-start (FS = a): Job B cannot start until a days after Job A is
finished
Start-to-start (SS = a): Job B cannot start until a days after Job A has
started
Finish-to-finish (FF = a): Job B cannot finish until a days after Job A
is finished
Start-to-finish (SF = a): Job B cannot finish until a days after Job A
has started
Critical Path Method (CPM): Basic Concepts
Task A
7 months
Task B
3 months
Start
End
Task C
11 months
Critical Path Method (CPM): Basic Concepts
ESA = 0
LFA = 8
ESStart = 0
LFStart = 0
ESB = 7
LFB = 11
Task A
7 months
Task B
3 months
ESEnd = 11
LFEnd = 11
Start
End
Task C
11 months
ESC = 0
LFC = 11
ESj = Earliest starting time for task (milestone) j
LFj = Latest finish time for task (milestone) j
AON Precedence Network: Microsoft Project
Task A
Task B
2
7d
3
3d
Wed 12/20/00
Thu 12/28/00
Fri 12/29/00
Tue 1/2/01
Start
1
0d
Wed 12/20/00
Wed 12/20/00
End
Task C
4
11d
Wed 12/20/00
Wed 1/3/01
5
0d
Wed 1/3/01
Wed 1/3/01
Critical Path Method (CPM): Example 2
ES A =
LFA =
TaskA
14 wks
ES START = 0
LF START = 0
ES B =
LFB =
START
Task B
9 wks
ES C =
LFC =
Task C
20 wks
ES F =
LFF =
ES D =
LFD =
Task F
9 wks
Task D
12 wks
ES END =
LFEND=
END
ES E =
LFE =
Task E
6 wks
Example 2: Network Paths
Path
1
2
3
4
5
Tasks
START-A-D-F-END
START-A-D-E-END
START-B-D-F-END
START-B-D-E-END
START-C-E-END
Expected
Duration (wks)
35
32
30
27
26
Example 2: CPM Calculations
EARLI EST
Task or
Milestone
Duration
( ti )
Start Time
(ES i)
START
0
14
9
20
12
6
9
0
0
0
0
0
14
26
26
35
A
B
C
D
E
F
END
LATES T
Finish Time
0
14
9
20
26
32
35
35
Start Time
0
0
5
9
14
29
26
35
Finish Time
(LFi)
0
14
14
29
26
35
35
35
Example 2: Calculating Total Slack (TSi)
Total Slack for task i = TSi = LFi - ESi - ti
Task or
Milestone
START
A
B
C
D
E
F
END
Duration
( ti )
0
14
9
20
12
6
9
0
Earliest
Start Time
(ES i)
0
0
0
0
14
26
26
35
Lastest
Finish Time
(LFi)
0
14
14
29
26
35
35
35
Total Slack
(TSi)
Critical
Task?
0
0
5
9
0
3
0
0
Yes
Yes
No
No
Yes
No
Yes
Yes
Slack (Float) Definitions (for task i)
Total Slack (TSi)
= LFi - ESi - ti
Free Slack (FSi)
= ESi,min - ESi - ti
where ESi,min = minimum early start time of all tasks that
immediately follow task i
= min (ESj for all task j  Si)
Safety Slack (SSi)
= LFi - LFi,max - ti
where LFi,max = maximum late finish time of all tasks that
immediately precede task i
= min (LFj for all task j  Pi)
Independent Slack (ISi)
= max (0, ESi,min - LFi,max - ti)
Example #2: LP Model
Decision variables: STARTj = start time for task j
END = ending time of project (END milestone)
Minimize END
subject to
STARTj ≥ FINISHi
STARTj ≥ 0
for all tasks i that immediately precede task j
for all tasks j in project
where FINISHi = STARTi + ti = STARTi + duration of task i
Example #2: Excel Solver Model
Gantt Chart
Microsoft Project 4.0
Project Budgeting
• The budget is the link between the functional units and the project
• Should be presented in terms of measurable outputs
• Budgeted tasks should relate to work packages in WBS and
organizational units responsible for their execution
• Should clearly indicate project milestones
• Establishes goals, schedules, and assigns resources (workers,
organizational units, etc.)
• Should be viewed as a communication device
• Serves as a baseline for progress monitoring & control
• Update on rolling horizon basis
• May be prepared for different levels of aggregation (strategic,
tactical, short-range)
Project Budgeting (cont’d)
• Top-down Budgeting: Aggregate measures (cost,
time) given by top management based on
strategic goals and constraints
• Bottom-up Budgeting: Specific measures aggregated
up from WBS tasks/costs and subcontractors
Issues in Project Budgets
• How to include risk and uncertainty factors?
• How to measure the quality of a project budget?
• How often to update budget?
• Other issues?
Critical Path Method (CPM): Example 2
ES A = 0
LFA = 14
TaskA
14 wks
ES START = 0
LF START = 0
ES B = 0
LFB = 14
START
Task B
9 wks
ES C = 0
LFC = 29
Task C
20 wks
ES F = 26
LFF = 35
ES D = 14
LFD = 26
Task F
9 wks
Task D
12 wks
ES END = 35
LFEND= 35
END
ES E = 26
LFE = 35
Task E
6 wks
Project Budget Example
Task or
Milestone
Duration
(tj)
Early Start
Time (ESj)
Latest Start
Time (LSj)
No. of
Resource A
workers
START
0
14
9
20
12
6
9
0
0
0
0
0
0
5
2
4
0
12
$
$
340
125
$
$
800
8,800
$
$
1,140
8,925
0
14
26
9
14
29
3
0
1
14
8
0
$
$
$
200
560
$
$
$
9,600
4,800
400
$
$
$
9,600
5,000
960
26
26
4
10
$
90
$
7,600
$
7,690
35
35
-
-
A
B
C
D
E
F
END
No. of
Resource B
workers
Material
Costs
Direct Labor
Cost/wk
-
-
-
Cost for Resource A worker = $400/week
Cost for Resource B worker = $600/week
-
Labor +
Materials
-
Project Budget Example (cont’d)
Week
Early Start Times
Tas k
1
A
1140
B
8925
C
9600
D
E
F
2
3
4
800
8800
9600
800
8800
9600
800
8800
9600
800
8800
9600
800
8800
9600
800
8800
9600
800
8800
9600
800
8800
9600
19665
19665
19200
38865
19200
58065
19200
77265
19200
96465
19200
115665
19200
134865
19200
154065
19200
173265
Wee kly Su btotals
Cumul ative
5
Wee kly Su btotals
Cumul ative
7
8
9
10
11
12
800
800
800
9600
9600
9600
10400
183665
10400
194065
10400
204465
Week
Late Start Times
Tas k
A
B
C
D
E
F
6
1
2
3
4
5
6
7
8
1140
800
800
800
800
8925
800
8800
800
8800
800
8800
1140
1140
800
1940
800
2740
800
3540
9725
13265
9600
22865
9600
32465
9600
42065
9
10
11
12
800
8800
9600
800
8800
9600
800
8800
9600
800
8800
9600
19200
61265
19200
80465
19200
99665
19200
118865
Cumulative Costs
Range of
feasible budgets
Weekly Costs (Cash Flows)
Managing Cash Flows
• Want to manage payments and receipts
• Must deal with budget constraints on
project and organization requirements (e.g.,
payback period)
• Organization profitability
Cash Flow Example
Make payment
of $5000
M1
Task A
2 mos
Task D
8 mos
Receive payment
of $3000
Task C
4 mos
START
END
Task B
8 mos
Task E
3 mos
M2
Receive payment
of $3000
Cash Flow Example: Solver Model
Material Management Issues
When to order materials? How much to order?
Example:
• Single material needed for Task B (2 units) and Task E (30 units)
• Fixed cost to place order = S
• Cost of holding raw materials proportional to number of unit-weeks in
stock
• Cost of holding finished product greater than the cost of holding raw
materials
• Project can be delayed (beyond 17 weeks) at cost of $P per week
Material Management Example
LS A = 0
LS B = 4
LS C = 12
Task A
4 wks
Task B
8 wks
Task C
5 wks
2 units
Start
End
LS D = 6
LS E = 12
LS F = 14
Task D
6 wks
Task E
2 wks
Task F
3 wks
30 units
Lot-Sizing Decisions in Projects
• To minimize holding costs, only place orders at Late Starting Times
• Can never reduce holding costs by delaying project
Time
1
2
3
Demand:
4
2
5
6
7
8
9
10
11
12
30
Order option #1: 32
Order option #2:
2
30
Choose the option that minimizes inventory cost = order cost + holding
cost of raw materials
Time-Cost Tradeoffs
Time-Cost Tradeoff Example
A
C
Start
End
B
D
Task
Normal
Duration
A
B
C
D
7
6
15
10
Marginal Cost
to Crash One
Normal Cost
Week
$60
$85
$55
$120
$8
$5
$10
$4
Time-Cost Tradeoff Example (cont’d)
Project
Duration
(weeks)
22
21
Critical Path(s)
Start-A-C-End
Start-A-C-End
Task(s) Reduced
Total Direct
Cost
A
$320
C
$338
C
$348
A, B
$361
$328
Start-A-B-End
20
Start-A-C-End
Start-A-B-End
19
Start-A-C-End
Start-A-B-End
18
Start-A-C-End
Start-A-B-End
Linear Time-Cost Tradeoff
In theory, the normal or expected duration of a task can be reduced by
assigning additional resources to the task
Cost
Crash
Point
Crash
cost = Ccj
Slope (bj) = Increase in cost by
reducing task by one time unit
Normal
Point
Normal
cost = CN
j
c
Crash time = tj
Time
Normal time
N
= tj
Balancing Overhead & Direct Costs
Cost
Total Cost
Indirect
(overhead)
Costs
Direct
Costs
Crash
Time
Minimum Cost
Solution
Normal Time
Project
Duration
Time-Cost Tradeoff (Direct Costs Only)
Given Normal point with cost CNj and time tNj
c
c
and time tj
and Crash point with cost Cj
Ccj - CNj
Assume constant marginal cost of crashing task j = bj = c N
t j - tj
Decision Variables: Sj = Starting time of task j
END = End time of project
tj = Duration of task j
Minimize Total Direct Cost =
Sj ≥ Si + ti
tcj Š tj Š tNj
END = Tmax
tj, Sj ≥ 0
• bj tj
j
for all tasks i  Pj
for all tasks in project
General Time-Cost Tradeoffs
Minimize Total Costs =
• bj tj
+ I (END) + P L
j
where
I = indirect (overhead) cost/time period
P = penalty cost/time period if END is delayed beyond
deadline Tmax
L = number of time periods project is delayed beyond
deadline Tmax
QUESTION: HOW TO DEFINE L?
Software Project Schedules
“Observe that for the programmer, as for the chef, the urgency of
the patron may govern the scheduled completion of the task, but it
cannot govern the actual completion. An omelet, promised in ten
minutes, may appear to be progressing nicely. But when it has not
set in ten minutes, the customer has two choices--wait or eat it
raw. Software customers have the same choices.
The cook has another choice; he can turn up the heat. The
result is often an omelet nothing can save--burned in one part, raw
in another.”
F.P. Brooks, “The Mythical Man-Month”,
Datamation, Vol 20, No 12 (Dec, 1974), pp.
44-52.
Coordination Costs (Software Development Project)
n
n
n
Assume you want to develop program that will require (approximately) 50,000 lines of
PERL code
A typical programmer can write approximately 1500 lines of code per week
Coordination time is M (M-1)/2 weeks
No. of
Programmers
1
2
3
4
5
6
7
8
9
10
11
No. of
Weeks
Coding
33.33
16.67
11.11
8.33
6.67
5.56
4.76
4.17
3.70
3.33
3.03
No. of
Coordination
Weeks
0
1
3
6
10
15
21
28
36
45
55
Total
Number of
Weeks
33.33
17.67
14.11
14.33
16.67
20.56
25.76
32.17
39.70
48.33
58.03
Brook’s Law
“Adding manpower to a late
software project makes it later.”
n
F.P. Brooks, “The Mythical Man-Month”,
Datamation, Vol 20, No 12 (Dec, 1974),
pp. 44-52.
Compressing New Product Development
Projects
Traditional Method
Design follows a sequential pattern where
information about the new product is slowly
accumulated in consecutive stages
Stage 0
Stage 1
Stage N
New Product Development Process
Overlapped Product Design
Allows downstream design stages to start before
preceding upstream stages have finalized their
specifications….
Stage 0
Stage 1
Stage N
Issues and Tradeoffs
What are the tradeoffs when moving from a
traditional sequential product design process
to an overlapped product design process?
• Increased uncertainty (that leads to additional
work)
• Can add additional resources to tasks to reduce
duration--but costs are increased
Classic PERT Model Defined
• Since task durations are now random variables, time of any
milestone (e.g., end of project) is now RV
• Assume all tasks are statistically independent
• Use values of j to identify expected critical path
• Since time of event (e.g., ESk) is now sum of independent RV’s,
central limit theorem specifies that ESk is approximately
normally distributed with mean E[ESk] and variance Var[ESk]
Expected early start time of task k = EES k =
where there exists s paths to task k
max
s
•
tasks j on path s
j
Classic PERT Model (cont’d)
Thus, expected project duration is defined as:
j
•
Expect Project Duration = E[ES END ] =
tasks j on CP
•
Variance of Project Duration = Var[ES END ] =
 2j
tasks j on CP
Using central limit theorem and standard normal distribution:
P ESEND Š Tmax = P z Š
Tmax - E ESEND
Var ESEND
PERT Example #1
Task B
Programming
Task E
Implementation
Start
Task A
Requirements
Analysis
Task C
Hardware
Acquisition
Task F
Testing
End
Task D
User
Training
Duration Estimates
Tas k
A
B
C
D
E
F
END
Desc ription
Requi remen ts Analysi s
Progra mmin g
Hardware acq uisi tion
User traini ng
Imp leme ntation
Tes ting
End of p roject
Predec essors
non e
A
A
A
B, C
E
D, F
Optimistic
2
4
2
12
3
3
0
Pess imis tic
14
12
13
18
7
7
0
Lik ely
6
7
8
14
5
4
0
Expected
Duration
6.6 7
7.3 3
7.8 3
14.33
5.0 0
4.3 3
0.0 0
Variance
4.0 0
1.7 8
3.3 6
1.0 0
0.4 4
0.4 4
0.0 0
PERT Example #1 (cont’d)
Task B
Programming
Task E
Implementation
Start
Task A
Requirements
Analysis
Task C
Hardware
Acquisition
Task F
Testing
End
Task D
User
Training
Ta sk
B,C,D
E
F
End
Path
Sta rt-A
Sta rt-A-C
Sta rt-A-C-E
Sta rt-A-C-E-F-End
PERT Expec ted Duration =
PERT V aria nce =
Ex pecte d
Ea rly Sta rt
6.67
14 .50
19 .50
23 .83
23 .83
8.250
Varia nce
4.00
7.36
7.81
8.25
Expecte d CP
Due Date
6
15
20
25
Zi
-0.33
0.18
0.18
0.41
= {Start, A, C, E, F, End}
Pr(zi)
0.37
0.57
0.57
0.66
PERT Example #2
Task A
Task C
A = 4
C = 10
 2A
=2
 2C = 5
END
START
Task B
Task D
B = 12
D = 3
 2B = 4
 2D = 1
Example #3: Discrete Probabilities
Task A
(8.0)
Task D
(9.3)
Task B
(10.0)
START
END
Task C
(19.0)
Task A
Task B
Task C
Task D
Value
Prob
Value
Prob
Value
Prob
Value
Prob
7
0.333
2
0.2
5
0.2
3
0.3
8
0.333
12
0.8
15
0.2
12
0.7
9
0.333
25
0.6
Example #3 (cont’d)
Task A
Combination
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
35
36
Value
7
7
7
7
7
7
7
7
7
7
7
7
8
8
8
8
8
8
8
8
8
8
8
8
9
9
9
9
9
9
9
9
9
9
9
9
Task B
Prob
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
Value
2
2
2
2
2
2
12
12
12
12
12
12
2
2
2
2
2
2
12
12
12
12
12
12
2
2
2
2
2
2
12
12
12
12
12
12
Task C
Prob
0.2
0.2
0.2
0.2
0.2
0.2
0.8
0.8
0.8
0.8
0.8
0.8
0.2
0.2
0.2
0.2
0.2
0.2
0.8
0.8
0.8
0.8
0.8
0.8
0.2
0.2
0.2
0.2
0.2
0.2
0.8
0.8
0.8
0.8
0.8
0.8
Value
5
5
15
15
25
25
5
5
15
15
25
25
5
5
15
15
25
25
5
5
15
15
25
25
5
5
15
15
25
25
5
5
15
15
25
25
Task D
Prob
0.2
0.2
0.2
0.2
0.6
0.6
0.2
0.2
0.2
0.2
0.6
0.6
0.2
0.2
0.2
0.2
0.6
0.6
0.2
0.2
0.2
0.2
0.6
0.6
0.2
0.2
0.2
0.2
0.6
0.6
0.2
0.2
0.2
0.2
0.6
0.6
Value
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
3
12
Critical
Prob
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
0.3
0.7
Path
A, D
A, D
C
A, D
C
C
B, D
B, D
B, D
B, D
C
C
A, D
A, D
C
A, D
C
C
B, D
B, D
B, D
B, D
C
C
A, D
A, D
C
A, D
C
C
B, D
B, D
B, D
B, D
C
C
Prob of
Length
CP
0.004
0.009
0.004
0.009
0.012
0.028
0.016
0.037
0.016
0.037
0.048
0.112
0.004
0.009
0.004
0.009
0.012
0.028
0.016
0.037
0.016
0.037
0.048
0.112
0.004
0.009
0.004
0.009
0.012
0.028
0.016
0.037
0.016
0.037
0.048
0.112
of CP
10
19
15
19
25
25
15
24
15
24
25
25
11
20
15
20
25
25
15
24
15
24
25
25
12
21
15
21
25
25
15
24
15
24
25
25
PATHS
A,D
B, D
C
0.004
0.009
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.004
0.009
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.004
0.009
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.037
0.016
0.037
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.037
0.016
0.037
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.037
0.016
0.037
0.000
0.000
0.000
0.000
0.004
0.000
0.012
0.028
0.000
0.000
0.000
0.000
0.048
0.112
0.000
0.000
0.004
0.000
0.012
0.028
0.000
0.000
0.000
0.000
0.048
0.112
0.000
0.000
0.004
0.000
0.012
0.028
0.000
0.000
0.000
0.000
0.048
0.112
6.8%
32.0%
61.1%
Example #3 (cont’d)
Length of
CP's
10
11
12
15
19
20
21
24
25
Cumulative
Prob
0.004
0.004
0.004
0.108
0.019
0.019
0.019
0.224
0.599
Prob
0.00
0.01
0.01
0.12
0.14
0.16
0.18
0.40
1.00
Criticality Indices
Task A
Task B
Task C
Task D
6.8%
32.0%
61.1%
38.8%
Expected Project Duration = 23.22
Monte-Carlo Simulation (PERT Example 1)
Ta sk
A
B
C
D
E
F
END
Run
1
2
3
4
5
6
7
Ta sk Dura tion
(Uniform Dis t)
4.99
4.75
3.38
12 .20
5.94
5.34
0.00
Proje ct Duration
31 .07
27 .41
23 .97
28 .93
26 .85
28 .82
28 .77
19 7
19 8
19 9
20 0
30 .37
29 .78
25 .33
29 .70
Ave
Var
27 .13
16 .777
Ea rly
Start
La test
Finis h
Total
Sla ck
Ex pecte d
Duration
0
4.99
4.99
4.99
9.74
15 .68
21 .02
4.99
9.74
9.74
21 .02
15 .68
21 .02
21 .02
0.00
0.00
1.36
3.83
0.00
0.00
0.00
6.67
7.33
7.83
14 .33
5.00
4.33
0.00
4.00
1.78
3.36
1.00
0.44
0.44
0.00
t(B)
1
0
1
0
1
0
0
t(C)
0
1
0
1
0
0
1
t(D)
0
0
0
0
0
1
0
t(E)
1
1
1
1
1
0
1
t(F)
1
1
1
1
1
0
1
0
1
1
0
1
0
0
1
0
0
0
0
1
1
1
1
1
1
1
1
48 .5%
42 .0%
Project Makespan
95% Confidence interval
99% Confidence interval
9.5%
Lower Limit
26.56
26.37
90 .5%
Upper Limit
27.72
27.90
Varia nce
90 .5%
Calculating Confidence Intervals
For a confidence interval, we can use the sample mean X
and the estimated standard error of the mean sX = s
n
where s is the sample standard deviation and n is the
number of trials
Using a normal approximation, a (1- a) twosided confidence interval is given by
+
X - za/2 sX
New Product Development Projects
Lease
Mfg/Office
Space
Beta test fails (with
probability of 0.25)
and rework is needed
Design of
physical unit
Assemble prototype
Beta test
prototype
START
END
Identify/hire
staff
Electronics
design
Software
Beta test fails (with
probability of 0.25)
and rework is needed
New Product Development Projects (cont’d)
Lease
Mfg /Office
Space
Design of
physical unit
Prob = .75
Assemble prototype
START
Beta test
prototype
Prob = .25
Identify/hire
staff
Electronics
design
Beta test fails and
rework is needed
Software
END
Critical Chain and the Theory of Constraints (TOC)
Project “Goal” (according to Goldratt): Meet Project Due Date
• Use deterministic CPM model with buffers to deal with any
uncertainties,
• Place project buffer after last task to protect the customer’s
completion schedule,
• Exploit constraining resources (make certain that resources are
fully utilized),
• Avoid wasting time slack time by encouraging early task
completions,
• Carefully monitor the status of the buffer(s) and communicate
this status to other project team members on a regular
basis, and
• Make certain that the project team is 100 percent focused on
critical chain tasks
Project Buffer Defined
• Project Buffer is placed at the end of the project to protect the
customer’s promised due date
Task B
Programming
Task E
Implementation
Start
Task A
requirements
analysis
Task C
Hardware
acquisition
Task F
Testing
Project
Buffer
Task D
User
User
training
PERT Example #1 Revisited with Project Buffer
End
Calculating Project Buffer Size
For those “who want a scientific approach to sizing
buffers....”
For tasks k on critical chain, we can calculate project buffer
using following formula that project will be completed
within worst-case duration estimates around 90 percent of
the time:
Buffer =
•
tasks k on critical chain
tpk - k
2
Implications of Project Uncertainty
Task A
END
START
Task B
Assume that the duration of both tasks A and B are described by a
normal distribution with a mean of 30 days
What is the probability that the project will be completed within 30
days?
Uncertainty and Worker Behavior
Consider a project with two tasks that must be completed serially
The duration of each task is described by a RV with values Ti (i = 1, 2)
Start
Task 1
Task 2
End
Values of T 1
Prob
Values of T 2
Prob
7
8
9
0.3
0.4
0.3
8.0
14
18
0.5
0.5
16
Parkinson’s Law (Expanding Work)
“Work expands so as to fill the time available for its
completion”
Professor C.N. Parkinson (1957)
Set a deadline D = 24 days
So T(D) = project makespan (function of D) where
E[T(D)] = E(T1) + E(T2) + E[max(0, D - T1 - T2)]
Values of T 1
7
7
8
8
9
9
Prob
0.3
0.3
0.4
0.4
0.3
0.3
Values of T 2
14
18
14
18
14
18
Prob
0.5
0.5
0.5
0.5
0.5
0.5
E[T(D)] = 25 days
Project
Makespan
24
25
24
26
24
27
Prob
0.15
0.15
0.2
0.2
0.15
0.15
Procrastinating Worker
Set a deadline D = 24 days
E’[T(D)] = E(T1) + E(T2) + E{max[0, D - T1 - E(T2)]}
Values of T 1
7
8
9
Prob
0.3
0.4
0.3
8
E[Delay] =
max[0, D - T 1 - E(T2)]
E[Makespan]
1
0
0
0.3
24
24
25
24.30
Can show that E[T(D)] ≥ E’[T(D)] ≥ D
What are the implications for project managers?
Schoenberger’s Hypothesis
An increase in the variability of task durations will
increase the expected project duration….
Schoenberger’s Hypothesis Illustrated
Task A
END
S TART
Task B
Duration of
Task A
12
14
16
14.0
Probability
0.1
0.8
0.1
Duration of
Task B
Probability
10
15
0.5
0.5
12.5
Schoenberger’s Hypothesis Illustrated
Realization
1
2
3
4
5
6
Task A
Duration
12
14
16
12
14
16
Task B
Duration
10
10
10
15
15
15
Probability
0.05
0.4
0.05
0.05
0.4
0.05
Max (A, B)
12
14
16
15
15
16
Expected duration equals 14.55 days
Increasing the variance of Task A:
Duration of
Task A
12
14
16
14.0
Probability
0.3
0.4
0.3
Duration of
Task B
10
15
Probability
0.5
0.5
12.5
Results in an increased expected duration = 14.65 days
Risk Management
• All projects involve some degree of risk
• Need to identify all possible risks and outcomes
• Need to identify person(s) responsible for managing
project risks
• Identify actions to reduce likelihood that adverse
events will occur
Risk Analysis
Risk Exposure (RE) or Risk Impact =
(Probability of unexpected loss) x (size of loss)
Example: Additional features required by client
Loss: 3 weeks
Probability: 20 percent
Risk Exposure = (.20) (3 weeks) = .6 week
How to Manage Project Risks?
Preventive Actions
• Actions taken in anticipation of adverse events
• May require action before project actually begins
• Examples?
Contingency Planning
• What will you do if an adverse event does occur?
• “Trigger point” invokes contingency plan
• Frequently requires additional costs
Risk and Contracts
High
Low
Low
High
Degree of Risk
Contractor
Client
Fixed Price Contract
Elements
can be
Firm price renegotiated
Incentives
Cost Plus Contract
T&M
with limits
Cost Plus
with
Incentives
Time &
materials
Tornado Diagram
Wage Rate
Direct Labor Hours
Material Units Needed
$1260
$1290
$1760
1760
$1700
$1720
$1265
$1680
$1310
1310
$1350
Early Completion Bonus
Material Unit Cost
$1690
$1350
Interest rates
Energy costs
$1640
$1380
$1400
Overhead
$1200
$1300
$1620
$1625
$1400
$1500
$1600
Project Cost ($000's)
$1700 $1800
Sensitivity Chart
Wage Rate
Direct Labor Hours
Material Units Needed
Early Completion Bonus
Material Unit Cost
Interest rates
Energy costs
Overhead
0.85
0.73
0.62
-0.45
0.42
0.28
0.19
0.10
-0.5
0
0.5
1.0
Rank Order Correlation with Total Project Cost
Van Allen Company
Strike
(wks)
Prob
3
0.45
4
0.3
5
0.25
E[Strike Duration]
Expected
Duration
1.35
1.20
1.25
3.80
Resource Allocation & Leveling
Resource Leveling: Reschedule the noncritical
tasks to smooth resource requirements
Resource Allocation: Minimize project
duration to meet resource availability constraints
Resource Allocation & Leveling
Three types of resources:
1) Renewable resources: “renew” themselves
at the beginning of each time period (e.g.,
workers)
2) Non-Renewable resources: can be used at
any rate but constraint on total number
available
3) Doubly constrained resources: both
renewable and non-renewable
Resource Leveling
Tas k C
9 wks
Tas k A
3 wks
Tas k D
5 wks
START
Tas k G
5 wks
Tas k B
2 wks
END
Tas k E
3 wks
Tas k F
2 wks
Task
A
B
C
D
E
F
G
Workers
7
3
2
10
4
5
6
Duration (t j)
3
2
9
5
3
2
5
Early Start
0
0
3
3
2
2
8
Late Start
0
3
4
3
5
11
8
Resource Leveling: Early Start Schedule
Resource Leveling: Late Start Schedule
Resource Leveling: Microsoft Project
Dec 17, '00
T
Dec 24, '00
W
T
10
10
F
S
S
Dec 31, '00
M
T
W
T
F
10
10
10
10
10
S
S
Ja n 7, '01
M
T
W
T
F
10
10
16
16
16
S
S
M
T
W
T
F
16
16
21
21
21
25
20
15
10
5
Workers
10
Overall oca ted:
Al located :
S
Renewable Resource Allocation Example
(Single Resource Type)
3 workers
6 workers
Task A
4 wks
Task C
1 wk
Task E
4 wks
START
Task B
3 wks
Task D
5 wks
5 workers
8 workers
7 workers
Maximum number of workers available = R = 9 workers
END
Resource Allocation Example: Early Start Schedule
Task C:
6 workers
Task A:
3 workers
Start
End
Task B:
5 workers
Task E:
7 workers
Task D:
8 workers
Week
1
2
3
4
5
6
7
8
9
10
11
12
No. of Workers/wk
Cumulative Workers
"Wasted" worker-wks
8
8
1
8
16
1
8
24
1
11
35
-
14
49
-
8
57
-
8
65
-
8
73
-
7
80
-
7
87
-
7
94
-
7
101
-
Maximum number of workers available = R = 9 workers
Resource Allocation Example: Late Start Schedule
Task A:
3 workers
Start
Task C:
6 workers
End
Task B:
5 workers
Task E:
7 workers
Task D:
8 workers
Week
1
2
3
4
5
6
7
8
9
10
11
12
No. of Workers/wk
Cumulative Workers
"Wasted" worker-wks
5
5
-
5
10
-
5
15
-
11
26
-
11
37
-
11
48
-
11
59
-
14
73
-
7
80
2
7
87
2
7
94
2
7
101
2
Maximum number of workers available = R = 9 workers
Resource Allocation Heuristics
n
Some heuristics for assigning priorities to available tasks j, where
number of units of resource k used by task j
n
1) FCFS:
n
2) GRU: (Greatest) resource utilization =
n
3) GRD: (Greatest) resource utilization x task duration =
Rkj denotes the
Choose first available task
• Rkj
• Rkj
k
tj
k
• Rkj / tj
n
4) ROT: (Greatest) resource utilization/task duration =
n
5) MTS: (Greatest) number of total successors
n
6) SPT: Shortest processing time = min {tj}
n
7) MINSLK: Minimum (total) slack
n
8) LFS: Minimum (total) slack per successor
n
9) ACTIMj: (Greatest) time from start of task j to end of project = CP - LSj
n
10) ACTRESj: (max) (ACTIMj)
n
11) GENRESj: w ACTIMj + (1-w) ACTRESj where 0 ≤ w ≤ 1
k
Resource Allocation Problem #2
Start
Tas k A1
6 days
Tas k A2
4 days
Tas k B1
3 days
Tas k B2
5 days
Tas k C1
2 days
Gold Crew
Tas k C2
5 days
Purple Crew
End
How to schedule tasks to minimize project makespan?
Priority scheme: schedule tasks using total slack (i.e., tasks with
smaller total slack have higher priority)
Gold Crew
Purple
Crew
Task A1
Task B1
1
2
3
4
5
6
7
1
2
3
4
5
6
7
8
Task C1
9
10
11
12
9
10
11
12
Task A2
8
13
14
15
16
17
15
16
17
Task B2
13
14
18
19
20
Task C2
18
19
20
Resource Allocation Example (cont’d)
But, can we do better? Is there a better priority scheme?
Gold Crew
Purple
Crew
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Microsoft Project Solution (Resource Leveling Option)
Solution by: Microsoft Project 2000
Critical Chain Project Management
• Identify the critical chain: set of tasks that determine the overall
duration of the project
• Use deterministic CPM model with buffers to deal with uncertainty
• Remove padding from activity estimates (otherwise, slack will be
wasted). Estimate task durations at median.
• Place project buffer after last task to protect customer’s completion
schedule
• Exploit constraining resource(s)
• Avoid wasting slack times by encouraging early task completions
• Have project team focus 100% effort on critical tasks
• Work to your plan and avoid tampering
• Carefully monitor and communicate buffer status
Critical Chain Buffers
Project Buffer
: placed after last task in project to protect schedule
Feeding Buffers
: placed between a noncritical task and a critical task
when the noncritical task is an immediate predecessor of the critical task
Resource Buffers
resource type
: placed just before a critical task that uses a new
Critical Chain Illustrated
Feeding Buffers
Tas k C1
2 days
Tas k B1
3 days
Tas k A1
6 days
End
Start
Tas k C2
5 days
Resource Buffers
Tas k B2
5 days
Tas k A2
4 days
Non-Renewable Resources
12 units
Task B
5 wks
6 units
8 units
Task A
6 wks
START
Task D
2 wks
END
Task C
3 wks
10 units
Task
A
B
C
D
Duration
6
5
3
2
No. of Nonrenewable
Resources Units
Needed
6
12
10
8
Early Start
0
6
6
11
Late Start
0
6
8
11
Non-Renewable Resources: Graphical Solution
Cumulative Resources
Supplied
40
Cumulative Resources
36
32
Cumulative Resources
Required
28
24
20
16
12
8
4
1
2
3
4
5
6
7
8
9
10
Weeks
11
12
13
14
15
16
17
18
19
20
Resource Allocation Problem #3
Issue: When is it better to “team” two or more
workers versus letting them work separately?
• Have 2 workers, Bob and Barb, and 4 tasks: A, B, C, D
• Bob and Barb can work as a team, or they can work separately
• When should workers be assigned to tasks? Which configuration
do you prefer?
How to Assign Project Teams?
A
C
Start
End
B
D
Configuration #1
Bob and Barb work jointly on all four tasks; assume that they can complete each
task in one-half the time needed if either did the tasks individually
Configuration #2
Bob and Barb work independently. Bob is assigned to tasks A and C; Barb is
assigned to tasks B and D
Bob and Barb: Configuration #1
TASK A
Duration
6
5
4
Expected
duration
Prob
0.33
0.33
0.33
5.0
TASK B
Duration
9
6
Prob
0.667
0.333
TASK C
Duration
12
7
8.0
Prob
TASK D
Duration
0.6
0.4
10
6
10.0
Configuration #1
Bob and Barb work jointly on all four tasks.
What is the expected project makespan?
Prob
0.25
0.75
7.0
Bob and Barb: Configuration #2
Bob and Barb work independently. Bob is assigned to tasks A and C; Barb is
assigned to tasks B and D
Realiza tion #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
A
6
6
6
6
6
6
6
6
5
5
5
5
5
5
5
5
4
4
4
4
4
4
4
4
B
9
9
9
9
6
6
6
6
9
9
9
9
6
6
6
6
9
9
9
9
6
6
6
6
C
12
12
7
7
12
12
7
7
12
12
7
7
12
12
7
7
12
12
7
7
12
12
7
7
D
10
6
10
6
10
6
10
6
10
6
10
6
10
6
10
6
10
6
10
6
10
6
10
6
Bob
A+C
Barb
B+D
18
18
13
13
18
18
13
13
17
17
12
12
17
17
12
12
16
16
11
11
16
16
11
11
19
15
19
15
16
12
16
12
19
15
19
15
16
12
16
12
19
15
19
15
16
12
16
12
max
(A+C,
B+D)
19
18
19
15
18
18
16
13
19
17
19
15
17
17
16
12
19
16
19
15
16
16
16
12
Prob
0.03
0.10
0.02
0.07
0.02
0.05
0.01
0.03
0.03
0.10
0.02
0.07
0.02
0.05
0.01
0.03
0.03
0.10
0.02
0.07
0.02
0.05
0.01
0.03
Bob and Barb: Configuration #2
Bob and Barb work independently. Bob is assigned to tasks A and C; Barb is
assigned to tasks B and D
max (A+C,
B+D)
12
13
15
16
17
18
19
Prob
0.07
0.03
0.20
0.20
0.17
0.17
0.17
Cumulative
Prob
0.07
0.10
0.30
0.50
0.67
0.83
1.00
Expected Project Makespan: 16.42
Parallel Tasks with Random Durations
Task A
START
END
Task B
• Assume that both Tasks A and B have possible durations:
8 days with probability = 0.5
10 days with probability = 0.5
• What is expected duration of project? (Is it 9 days?)
Project Monitoring and Control
n
“It is of the highest importance in
the art of detection to be able to
recognize, out of a number of acts,
which are incidental and which are
vital. Otherwise your energy and
attention must be dissipated instead
of being concentrated.”
Sherlock Holmes
Status Reporting?
One day my Boss asked me to submit a status
report to him concerning a project I was working
on. I asked him if tomorrow would be soon enough.
He said, "If I wanted it tomorrow, I would have
waited until tomorrow to ask for it!"
New business manager, Hallmark Greeting Cards
Control System Issues
n
n
n
n
n
What are appropriate performance metrics?
What data should be used to estimate the value of each
performance metric?
How should data be collected? From which sources? At
what frequency?
How should data be analyzed to detect current and future
deviations?
How should results of the analysis be reported? To whom?
How often?
Controlling Project Risks
Key issues to control risk during projecct:
(1) what is optimal review frequency, and
(2) what are appropriate review acceptance levels
at each stage?
“Both over-managed and under-managed
development processes result in lengthy design
lead time and high development costs.”
Ahmadi & Wang. “Managing Development Risk in
Product Design Processes”, 1999
Project Control & System Variation
Common cause variation: “in-control” or normal
variation
Special cause variation: variation caused by forces
that are outside of the system
According to Deming:
• Treating common cause variation as if it were special cause variation
is called “tampering”
• Tampering always degrades the performance of a system
Control System Example #1
n
Project plan: We estimate that a task will
take 4 weeks and require
n
1600 worker-hours
At the end of Week 1, 420 worker-hours
have been used
Is the task “out of control”?
Control System Example (cont’d)
Week 2: Task expenses = 460 worker-hours
Planned Cost
Week
(BCWS)
1
400
2
400
Actual Cost
420
460
Cumulative
Actual Cost
(ACWP)
420
880
470
Cost (in worker-hours)
460
450
440
430
420
410
400
390
380
370
1
2
3
4
Week
Is the task “out of control”?
Control System Example (cont’d)
Week 3: Task expenses = 500 worker-hrs
Week
Planned cost
(worker-hours )
Actual cost
(worker-hours )
Cumulative cos t
(worker-hours )
1
2
3
400
400
400
420
460
500
420
880
1380
600
Worker-hours
500
400
300
200
100
0
1
2
3
4
Week
Is the task “out of control”?
Earned Value Analysis
• Integrates cost, schedule, and work performed
• Based on three metrics that are used as the basic
building blocks:
BCWS: Budgeted cost of work scheduled
ACWP: Actual cost of work performed
BCWP: Budgeted cost of work performed
Schedule Variance (SV)
Schedule Variance (SV) = difference between value of
work completed and value of scheduled work
Schedule Variance (SV) = Earned Value - Planned Value
= BCWP - BCWS
Cost Variance (CV)
Cost Variance (CV) = difference between value of
work completed and actual
expenditures
Cost Variance (CV) = Earned Value - Actual Cost
= BCWP - ACWP
Earned Values Metrics Illustrated
Worker-Hours
Present time
Planned Value
(BCWS)
Actual Cost
(ACWP)
BAC
Cost Variance
(CV)
Earned Value
(BCWP)
Schedule Variance
(SV)
Week 1
Week 2
Week 3
Week 4
Week 5
Week 6
Relative Measure: Schedule Index
Schedule Index
(SI ) =
BCWP
BCWS
If SI = 1,
then task is on schedule
If SI > 1,
then task is ahead of schedule
If SI < 1,
then task is behind schedule
Relative Measure: Cost Index
Cost Index (CI) =
BCWP
ACWP
If CI = 1,
then work completed equals
payments (actual expenditures)
If CI > 1,
then work completed is ahead
of payments
If CI < 1,
then work completed is behind
payments (cost overrun)
Example #2
W E E K
1
2
3
4
5
6
7
8
9
10
Tas k A (36 worker-hrs)
6
6
6
8
10
Tas k B (36 worker-hrs)
12
12
12
Tas k C (56 worker-hrs)
10
10
12
12
12
Weekly
Scheduled
Worker-Hrs
Cumulative
6
6
6
20
22
22
10
12
12
12
Scheduled
Worker-Hrs
(BCWS)
6
12
18
38
60
82
92
104
116
128
Example #2 (cont’d)
Progress report at the end of week #5:
Cumulative Percent of Work Completed:
Week
1
Task A 15%
Task B
Task C
2
3
4
5
30%
40%
60%
80%
25%
65%
Not started yet
Worker-Hours Charged to Project:
Week
1
2
3
4
5
Task A
Task B
Task C
5
6
8
10
15
10
10
Not started yet
Example #2 (cont’d)
Progress report at the end of week #5:
W E E K
1
2
3
4
5
6
7
8
9
10
6
12
18
38
60
82
92
104
116
128
5
11
19
44
64
Earned Value
(BCWP)
5.4
10.8
14.4
30.6
52.2
Schedule
Variance (SV)
-0.6
-1.2
-3.6
-7.4
-7.8
Cos t Variance
(CV)
0.4
-0.2
-4.6
-13.4
-11.8
Cumulative
Scheduled
Worker-Hrs
(BCWS)
Actual WorkerHrs Used
(ACWP)
Example #2 (cont’d)
140
BAC
120
BCWS
Performance Metric
100
80
Cost
Variance
Schedule
Variance
60
ACWP
40
BCWP
20
0
1
2
3
4
5
6
Week
7
8
9
10
Using a Fixed 20/80 Rule
Cumulative Percent of Work Completed:
Week
1
Task A 20%
Task B
Task C
2
3
4
5
20%
20%
20%
20%
20%
20%
Not started yet
W E E K
Cumulative
Scheduled
Worker-Hrs
(BCWS)
Actual WorkerHrs Used
(ACWP)
Earned Value
(BCWP)
Schedule
Variance (SV)
Cos t Variance
(CV)
1
2
3
4
5
6
7
8
9
10
6
12
18
38
60
82
92
104
116
128
5
11
19
44
64
7.2
7.2
7.2
14.4
14.4
1.2
-4.8
-10.8
-23.6
-45.6
2.2
-3.8
-11.8
-29.6
-49.6
Using a Fixed 20/80 Rule
140
120
Cost (in Worke r-hours)
100
80
BCWS
ACWP
60
40
BCWP
20
0
1
2
3
4
5
6
We e k
7
8
9
10
Updating Forecasts: Pessimistic Viewpoint
Assumes that rate of cost overrun will continue
for life of project….
Estimate at Completion
(EAC) = ACWP BAC = 1 BAC .
BCWP
CI
= (64/52.2) 128 = 1.23 x 128 = 156.94 worker-hrs
Updating Forecasts: Optimistic Viewpoint
Assumes that cost overrun experienced to date
will cease and no further cost overruns will be
experienced for remainder of project life…
Estimate at Completion
(EAC) = BAC - CV = 128 + 11.8 = 139.8 worker -hrs .
Multi-tasking with Multiple Projects
How to prioritize your work when you have multiple
projects and goals?
Consider two projects with and without multi-tasking
Project A
A-1
B-1
Project B
A-2
B-2
A-3
B-3
A-4
B-4
Due-Date Assignment with Dynamic Multiple Projects
• Projects arrive dynamically (common situation for both
manufacturing and service organizations)
• How to set completion (promise) date for new projects?
• Firms may have complete control over due-dates or only partial
control (i.e., some due dates are set by external sources)
• How to allocate resources among competing projects and tasks (so
that due dates can be realized)?
• What are appropriate metrics for evaluating various rules?
What Does the Research Tell Us?
• Study by Dumond and Mabert* investigated four due date assignment
rules and five scheduling heuristics
• Simulated 250 projects that randomly arrive over 2000 days
• average interarrival time = 8 days
• 6 - 49 tasks per project (average = 24); 1 - 3 resource types
• average critical path = 31.4 days (range from 8 to 78 days)
• Performance criteria: 1) mean completion time
2) mean project lateness
3) standard deviation of lateness
4) total tardiness of all projects
• Partial and complete control on setting due dates
* Dumond, J. and V. Mabert. “Evaluating Project Scheduling and Due Date Assignment Procedures:
An Experimental Analysis” Management Science, Vol 34, No 1 (1988), pp 101-118.
Experimental Results
• No one scheduling heuristic performs best across all due date
setting combinations
• Mean completion times for all scheduling and due date rules not
significantly different
• FCFS scheduling rules increase total tardiness
• SPT-related rules do not work well in PM (SASP)
• Best to use more detailed information to establish due dates
Project Management Maturity Models
• Methodologies to assess your organization’s current level of
PM capabilities
• Based on extensive empirical research that defines “best
practice” database as well as plan for improving PM process
• Process of improvement describes the PM process from
“ineffective” to “optimized”
• Also known as “Capability Maturity Models”
PM Maturity Model Example*
1)
Ad-Hoc
The project management process is described as disorganized, and occasionally even
chaotic. Systems and processes are not defined. Project success depends on individual effort.
Chronic cost and schedule problems.
2)
Abbreviated: Some project management processes are established to track cost, schedule,
and performance. Underlying disciplines, however, are not well understood or consistently
followed. Project success is largely unpredictable and cost and schedule problems are the norm.
3)
Organized: Project management processes and systems are documented, standardized, and
integrated into an end-to-end process for the company. Project success is more predictable. Cost
and schedule performance is improved.
4) Managed: Detailed measures of the effectiveness of project management are collected and used
by management. The process is understood and controlled. Project success is more uniform.
Cost and schedule performance conforms to plan.
5) Adaptive:
Continuous improvement of the project management process is enabled by feedback
from the process and from piloting innovative ideas and technologies. Project success is the
norm. Cost and schedule performance is continuously improving.
* source: The Project Management Institute PM Network (July, 1997), Micro Frame Technologies, Inc. and
Project Management Technologies, Inc. (http://pm32.hypermart.net/)