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?
QuickTim e™ and a Photo - JPEG decompres sor are needed to s ee 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
Formation & Selection
Phase 2 Planning
Phase 3
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
M TI
E
LE U D E CH S (
Due Date
)
Budget Constraint
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 2 = 0 1+r 1+r
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 M edium Low
Product Life
Annual Net Cash Inflow
Probability
20 years 10 years Abandon Project
$24 million $12 million None
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? Expected value of product at Year 4:
$299.09 $92.66 $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 Cash Flow $ (18.00) $ (18.00) $ (6.00) $ 75.63
What is the internal rate of return for this project?
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
Launch New Product
Probability = 1 - pc Commercial Failure (with net benefit = 0)
Develop New Product Probability = 1 - pt Technical Failure
Risk class 1
Risk class 2
DCF Example Revisited Product Demand High
0.3 Probability = pt Development Succeeds Research & Product Development
0.5 Market Development
0.2 Probability = 1 - pt Development Fails
Discount rate r1
Product Demand Medium
Product Demand Low
Drop project
Discount rate r2
Ranking/Scoring Models Profit abilit y/value 1) Increase in profitability? 2) Increase in market share? 3) Will add knowledge to organization that can be leveraged by other projects? 4) Estimated NPV, ECV, etc. Organizat ion's Strat egy 1) Consistent with organization's mission statement? 2) Impact on customers? Risk 1) Probability of research being successful? 2) Probability of development being successful? 3) Probability of process success? 4) Probability of commercial success? 5) Overall risk of project 6) Adequate market demand? 7) Competitors in market Organizat ion Costs 1) Is new facilit y needed? 2) Can use current personnel? 3) External consultants needed? 4) New hires needed? Miscellaneous Factors 1) Impact on environmental standards? 2) Impact on workforce safety? 3) Impact on quality? 4) Social/political implications
Scoring Attributes To convert various measurement scales to a (0, 1) range‌. xi - L LINEAR SCALE: value of attribute i is vi xi = U - L 1 - exp L - xi v x = . EXPONENTIAL SCALE: value of attribute i is i i 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
Extent of Product Change
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 Project A Project B Budget Limit (B t )
Y e a r (t) 2
3
4
($40)
$10
$20
$20
($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 Timetoj Expected duration = Âľ
M ost 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 Âľj = 6
Variance of task j =
Ďƒ2j
tpj - toj 2 = 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 450 400
300 250 200 150 100 50
N = No. of Team Members
N(N-1) 2
29
27
25
21
23
19
17
=
15
L = Number of Intra-team Links = N 2
11
13
9
7
5
3
0 1
L = No. of Intra-team Links
350
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 6.00 5.80 5.60 Group Harmony
5.40 5.20 5.00 4.80 4.60 4.40 4.20 4.00 1
2
3
4
5
6
Week High Performance (low cost) Teams
Low Performance (high c ost) Teams
7
Extent of Individual Contribution: High vs Low Performing Groups 6.00
Extent of Indiv idual Contributions
5.80 5.60 5.40 5.20 5.00 4.80 4.60 4.40 4.20 4.00 1
2
3
4
5
6
Week High Performance (low cost) Teams
Low Performance (high c ost) Teams
7
Decision Making Effectiveness: High vs Low Performing Groups
Decision Making Effectiv eness
6.00 5.50 5.00 4.50 4.00 3.50 3.00 1
2
3
4
5
6
Week High Performance (low cost) Teams
Low Performance (high c ost) Teams
7
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 Cost 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)
10.38*
6.94*
7.42*
11.45*
Organizational Structure
F-statistic Scheffe Results *Statistically significant at a p<0.01 level
A,B < C,D,E E<D 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: â&#x20AC;&#x153;A business relationship for mutual benefit between two or more parties with compatible or complementary business interests and/or goalsâ&#x20AC;? 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
â&#x20AC;&#x153;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.â&#x20AC;? 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 = α): Job B cannot start until α days after Job A is finished Start-to-start (SS = α): Job B cannot start until α days after Job A has started Finish-to-finish (FF = α): Job B cannot finish until α days after Job A is finished Start-to-finish (SF = α): Job B cannot finish until α 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 =
Task A 14 wks ES START = 0 LFSTART = 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 EARLIEST
Task or Milestone START
A B C D E F END
Duration ( ti ) 0 14 9 20 12 6 9 0
S tart Time (ES i) 0 0 0 0 14 26 26 35
Finish Time 0 14 9 20 26 32 35 35
L A TE S T
S tart 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
Duration ( ti )
Earliest 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
Lastest Finish Time (LFi) 0 14 14 29 26 35 35 35
Total S lack (TSi)
Critical Task?
0
Yes
0
Yes
5 9
No No
0
Yes
3
No
0 0
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 February ID
Task Name
1
Start
2
Task A
3
Task B
4
Task C
5
Task D
6
Task E
7
Task F
8
Task G
9
Task H
10
Task J
11
End
March 21 24
27
1
4
7
10 13
16
April 19 22 25 28
31
3
6
9
May
12 15 18
21
24
27 30
3
6
9
12
15 18
3/1 Workers[5] Workers[7] Workers[3] Workers[12] Workers[2] Workers Workers[2] Workers[5] Workers[6] 5/10
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
Task A 14 wks ES START = 0 LFSTART = 0 ES B = 0 LFB = 14
ES F = 26 LFF = 35 ES D = 14 LFD = 26
Task F 9 wks
Task D 12 wks
START
Task B 9 wks
ES C = 0 LFC = 29
Task C 20 wks
ES END = 35 LFEND= 35
END
ES E = 26 LFE = 35
Task E 6 wks
Project Budget Example Task or Milestone ST ART
A B C D E F END
Duration (tj) 0 14 9 20 12 6 9 0
Early Start Time (ESj)
No. of No. of Latest Start Resource A Resource B Time (LSj) work ers workers
Material Costs
Direct Labor Labor + Materials Cost/wk
0 0 0 0 14
0 0 5 9 14
2 4 3 0
0 12 14 8
-
-
$ $ $ $
340 125 200
$ $ $ $
800 8,800 9,600 4,800
$ $ $ $
1,140 8,925 9,600 5,000
26
29
1
0
$
560
$
400
$
960
26 35
26 35
4 -
10 -
$
90
$
7,600 -
$
7,690 -
-
Cost for Resource A worker = $400/week Cost for Resource B worker = $600/week
Project Budget Example (contâ&#x20AC;&#x2122;d) Week
Early Start Times Task 1 A 1140 B 8925 C 9600 D E F
2
3
4
5
6
7
8
9
10
11
12
800 8800 9600
800 8800 9600
800 8800 9600
800 8800 9600
800 8800 9600
800 8800 9600
800 8800 9600
800 8800 9600
800
800
800
9600
9600
9600
19665 19665
19200 38865
19200 58065
19200 77265
19200 96465
19200 115665
19200 134865
19200 154065
19200 173265
10400 183665
10400 194065
10400 204465
Weekly Subtotals Cumulative
Week
Late Start Times Task A B C D E F Weekly Subtotals Cumulative
1
2
3
4
5
6
7
8
9
10
11
12
1140
800
800
800
800 8925
800 8800
800 8800
800 8800
800 8800 9600
800 8800 9600
800 8800 9600
800 8800 9600
1140 1140
800 1940
800 2740
800 3540
9725 13265
9600 22865
9600 32465
9600 42065
19200 61265
19200 80465
19200 99665
19200 118865
Cumulative Costs 450000 400000
300000
Range of feasible budgets
250000 200000 150000 100000 50000
Week Early Start Sc hedule
Late Start Schedule
33
31
29
27
25
23
21
19
17
15
13
11
9
7
5
3
0 1
Cumulativ e Cost
350000
Weekly Costs (Cash Flows) 25000
15000
10000
5000
Week Early Start Schedule
Late Start Schedule
33
31
29
27
25
23
21
19
17
15
13
11
9
7
5
3
0 1
Weekly Costs
20000
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
Task A 4 wks
Task B 8 wks
LS C = 12 Task C 5 wks
2 units
Start LS D = 6
LS E = 12
Task D 6 wks
Task E 2 wks
30 units
End LS F = 14 Task F 3 wks
Lot-Sizing Decisions in Projects â&#x20AC;˘ To minimize holding costs, only place orders at Late Starting Times â&#x20AC;˘ Can never reduce holding costs by delaying project Time 1
2
3
4
Demand:
2
Order option #1:
32
Order option #2:
2
5
6
7
8
9
10
11
12 30
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
M arginal Cost to Crash One Normal Cost Week $60 $85 $55 $120
$8 $5 $10 $4
Time-Cost Tradeoff Example (contâ&#x20AC;&#x2122;d) Project Duration (weeks)
Total Direct Cost
Critical Path(s)
Task(s) Reduced
22 21
Start-A-C-End
A
$320
20
Start-A-C-End
C
$338
19
Start-A-C-End
C
$348
18
Start-A-C-End
A, B
$361
Start-A-C-End Start-A-B-End Start-A-B-End Start-A-B-End Start-A-B-End
$328
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
Crash time =
tcj
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 tj - 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 12
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 Total Coordination Number of Weeks Weeks 0 33.33 1 17.67 3 14.11 6 14.33 10 16.67 15 20.56 21 25.76 28 32.17 36 39.70 45 48.33 55 58.03
10 No. of Programmers (Cost)
No. of Programmers 1 2 3 4 5 6 7 8 9 10 11
8
6 4
2 0 0
10
20
30
40
Total Number of Weeks
50
60
70
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 specificationsâ&#x20AC;Ś. 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? â&#x20AC;˘ Increased uncertainty (that leads to additional work) â&#x20AC;˘ 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 µj 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 = max s where there exists s paths to task k
µj tasks j on path s
Classic PERT Model (cont’d) Thus, expected project duration is defined as: µj
Expect Project Duration = E[ES END ] = tasks j on CP
σ 2j
Variance of Project Duration = Var[ES END ] = 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
S tart
Task A Requirements Analysis
Task C Hardware Acquisition
Task F Testing
End Task D User Training
Duration Estimates Task A B C D E F END
Description Requirements Analy sis Programming Hardware acquisition User training Implementation Testing End of project
Predecessors none A A A B, C E D, F
Optimistic 2 4 2 12 3 3 0
Pessimistic 14 12 13 18 7 7 0
Likely
Expected Duration
Variance
6 7 8 14 5 4 0
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
PERT Example #1 (contâ&#x20AC;&#x2122;d) Task B Programming Task E Implementation
S tart
Task A Requirements Analysis
Task C Hardware Acquisition
Task F Testing
End Task D User Training
Task B,C,D E F End
Path Start-A Start-A-C Start-A-C-E Start-A-C-E-F-End
PERT Expected Duration = PERT Variance =
Expected Early Start 6.67 14.50 19.50 23.83 23.83 8.250
Variance
Due Date
Zi
4.00 7.36 7.81 8.25
6 15 20 25
-0.33 0.18 0.18 0.41
Expected CP
= {Start, A, C, E, F, End}
Pr(zi) 0.37 0.57 0.57 0.66
PERT Example #2 Task A
Task C
µA = 4
µC = 10
σ 2A
=2
σ 2C = 5
END
START Task B
Task D
µB = 12
µD = 3
σ B2 = 4
σ 2D = 1
Example #3: Discrete Probabilities Task A (8.0) Task D (9.3) Task B (10.0)
S TART
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 (contâ&#x20AC;&#x2122;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 (contâ&#x20AC;&#x2122;d) Length of CP's 10 11 12 15 19 20 21 24 25
Cumulativ e 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 6.8%
Task B 32.0%
Task C 61.1%
Task D 38.8%
Expected Project Duration = 23.22
Monte-Carlo Simulation (PERT Example 1) Task A B C D E F END
Run 1 2 3 4 5 6 7
Task Duration (Uniform Dist) 4.99 4.75 3.38 12.20 5.94 5.34 0.00
Project Duration 31.07 27.41 23.97 28.93 26.85 28.82 28.77
197 198 199 200
30.37 29.78 25.33 29.70
Ave Var
27.13 16.777
Early Start
Latest Finish
Total Slack
Expected Duration
Variance
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
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 = sn where s is the sample standard deviation and n is the number of trials Using a normal approximation, a (1- Îą) twosided confidence interval is given by
+
X - zÎą/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 (contâ&#x20AC;&#x2122;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 â&#x20AC;˘ Project Buffer is placed at the end of the project to protect the customerâ&#x20AC;&#x2122;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 - µk
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 T2
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 T1 7 7 8 8 9 9
Prob 0.3 0.3 0.4 0.4 0.3 0.3
Values of T2 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
Prob
E[Delay] = max[0, D - T1 - E(T2)]
7 8 9
0.3 0.4 0.3 8
1 0 0 0.3
E[Makespan] 24 24 25 24.30
Can show that E[T(D)] ≥ E’[T(D)] ≥ D What are the implications for project managers?
Schoenbergerâ&#x20AC;&#x2122;s Hypothesis An increase in the variability of task durations will increase the expected project durationâ&#x20AC;Ś.
Schoenbergerâ&#x20AC;&#x2122;s Hypothesis Illustrated
Task A
END
START
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
Schoenbergerâ&#x20AC;&#x2122;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
Cost Plus Contract
Elements Cost Plus can be T&M with Firm price renegotiated Incentives with limits Incentives
Time & materials
Tornado Diagram Wage Rate Direct Labor Hours M aterial 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 M aterial Units Needed Early Completion Bonus M aterial 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 Task
Immediate Predecessors
Start A B C D E F G H I End
Start Start B A A C, D C, D G E, G F, H, I
Strike (wks)
Time
Cost
Time
Cost
0
-
0
-
3 1 5 2 2 5 4 1 1 0
$60 $50 $70 $60 $50 $90 $60 $40 $50 -
5 5 10 7 6 11 6 5 4 0
$40 $30 $40 $40 $30 $60 $30 $20 $20 -
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: â&#x20AC;&#x153;renewâ&#x20AC;? 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 Task C 9 wks
Task A 3 wks
Task D 5 wks
S TART
Task G 5 wks
Task B 2 wks
END
Task E 3 wks
Task 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 S tart 0 0 3 3 2 2 8
Late S tart 0 3 4 3 5 11 8
Resource Leveling: Early Start Schedule Early Start Schedule
Number of Workers Needed
25 20 Task G Task F Task E Task D Task C Task B Task A
15
10 5 0 1
2
3
4
5
6
7 Week
8
9
10
11
12
13
Resource Leveling: Late Start Schedule Late Start Schedule 18
Number of Workers Needed
16 14
Task G Task F Task E Task D Task C Task B Task A
12 10 8 6 4 2 0 1
2
3
4
5
6
7 Week
8
9
10
11
12
13
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
Jan 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
Worke rs
10 Ov erallocated:
Allocated:
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 S tart
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 S tart
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: Choose first available task
Rkj denotes the
Rkj
n
2) GRU: (Greatest) resource utilization =
n
3) GRD: (Greatest) resource utilization x task duration =
k
n
4) ROT: (Greatest) resource utilization/task duration =
Rkj / tj
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
Rkj tj
k
k
Resource Allocation Problem #2
Start
Task A1 6 days
Task A2 4 days
Task B1 3 days
Task B2 5 days
Task C1 2 days
Task C2 5 days
Gold Crew
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 (contâ&#x20AC;&#x2122;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 Task C1 2 days
Task B1 3 days
Task A1 6 days
End
Start
Task C2 5 days
Resource Buffers
Task B2 5 days
Task 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 S tart 0 6 6 11
Late S tart 0 6 8 11
Non-Renewable Resources: Graphical Solution
Cumulative Resources Supplied
40
Cumulative Re source s
36 32
Cumulative Resources Required
28 24 20 16 12 8 4 1
2
3
4
5
6
7
8
9
10
We eks
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
Realization #
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
Cumulativ e 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
â&#x20AC;˘ Assume that both Tasks A and B have possible durations: 8 days with probability = 0.5 10 days with probability = 0.5 â&#x20AC;˘ What is expected duration of project? (Is it 9 days?)
Project Monitoring and Control n
â&#x20AC;&#x153;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.â&#x20AC;? 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
What are appropriate performance metrics?
n
What data should be used to estimate the value of each performance metric?
n
How should data be collected? From which sources? At what frequency?
n
How should data be analyzed to detect current and future deviations?
n
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 â&#x20AC;&#x153;out of controlâ&#x20AC;??
Control System Example (cont’d) Week 2: Task expenses = 460 worker-hours Week 1 2
Planned Cost (BCWS) 400 400
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 cost (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 â&#x20AC;˘ Integrates cost, schedule, and work performed â&#x20AC;˘ 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
Task A (36 worker-hrs) 6
6
6
8
10
Task B (36 worker-hrs) 12
12
12 Task 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 (contâ&#x20AC;&#x2122;d) Progress report at the end of week #5: Cumulative Percent of Work Completed: Week
1
2
3
4
5
Task A Task B Task C
15%
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 (contâ&#x20AC;&#x2122;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
Cost Variance (CV)
0.4
-0.2
-4.6
-13.4
-11.8
Cumulative Scheduled Worker-Hrs (BCWS) Actual WorkerHrs Used (ACWP)
Example #2 (contâ&#x20AC;&#x2122;d) 140 BAC 120
BCWS
Performance Metric
100
80 Cost Variance
S chedule 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
2
3
4
Task A Task B Task C
20%
20%
20%
5
20% 20% Not started yet
20% 20%
W E E K Cumulative S cheduled Worker-Hrs (BCWS ) Actual WorkerHrs Used (ACWP) Earned Value (BCWP) S chedule Variance (S V) Cost 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 Week
7
8
9
10
Updating Forecasts: Pessimistic Viewpoint Assumes that rate of cost overrun will continue for life of projectâ&#x20AC;Ś.
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 lifeâ&#x20AC;Ś
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/)