Lecture09 cc

Date/reference/classification 1

Session 9 Critical Chain

Single-Project Management Reducing Project Duration by 25%

& Increasing Due Date Performance Without Changes of Resource Capacity Dr. Thomas Lechler Phone: (201) 216-8174 Babbio Center 416 FAX: (201) 216-5385 email: [email protected]

Date/reference/classification

2 © T. Lechler, Ph.D., 2013

1. Paradigm Question: Process Value

How much would your organization gain if over 90%

of the projects are finished in time?

How much would your organization gain if all new

projects starting next month finished 15-25% sooner?

• Reduced project cost?

• Faster time-to-market?

• Greater responsiveness to customers?



1. Paradigm Question: Management Problems

Why is it so difficult to manage projects which deliver on time, within budget and with the full specification or scope intact ?

•__________________________________________________

•__________________________________________________

•__________________________________________________

•__________________________________________________

•__________________________________________________



1. Paradigm Question: Project Manager’s Dilemma

• PRESSURE TO INCREASE SAFETY TIME - in order to complete projects on time

• PRESSURE TO REDUCE OVERALL PROJECT SCHEDULE - in order to meet the

customers need for shorter lead times

Be a good

Project Manager

Complete Projects

on time

(to be realistic)

Add protection

to tasks

Don’t add

protection to

tasks

Prerequisite Requirement

Respond to

customers need for

short lead times

(quick response)

Objective



1. Paradigm Question: CPM Management of Uncertainty

Often projects

have difficulty

staying within

budget

Often scope or

specifications are

cut from project

The Way We Manage

Uncertainty in

Projects

Often projects

have difficulty

finishing on time

YOU CAN’T IMPOSE CERTAINTY ON UNCERTAINTY

YOU MUST LEARN TO MANAGE THE UNCERTAINTY



1. Paradigm Question: CPM Estimation

Average Work Uncertainty Multitasking Set-up Time

Task Estimate - (Duration)

Basic Estimation Equation: W = U * D

W: Estimate of Work

U: Units of Resources

D: Duration

Inflated Estimates!



1. Paradigm Question: CPM Assumptions

Basic assumptions of the CPM paradigm:

1. Each task is scheduled with 90% likelihood thus the project is

scheduled with 90% likelihood. But:

Due date on a big project may be less than 10% likely even if the

single tasks are scheduled with 90%!

2. The critical path does not change. But:

The critical path changes in many projects!

3. Good and bad luck average out. But:

Good luck is wasted and bad luck accumulates because

- Scheduled tasks are started with delays (Student Syndrome)

- Early task finish is often not reported (Parkinson’s Law)

=> What are the consequences?



1. Paradigm Question: CPM Assumption 1

Two scenarios:

1. All single tasks on the critical path are planned with 90% accuracy: The probability to achieve the planned finish date is: 38% (=0.9 EXP 9)

2. One of the 9 tasks could be only planned with 80% accuracy and one with 50%. The likelihood to achieve the planned finish date is: 18%!

Start

0d 9/1

(a) Initiate project

5d 9/1

(b) Assign PM

2d 9/8

(h) Design IPF

9d 9/12

(g) Conduct IPD

training

5d 9/28 (m) Conduct IPF

training

7d 10/5

(j) Prepare business

case

7d 10/5

(l) Finalize IPP

9d 10/18

(k) Write IPP

3d 10/13

(e) Develop project

schedule

10d 9/28

(i) Assign resources

1d 10/12

(d) Form PDT

12d 9/12

(f) Prepare expense

estimates

9d 9/28

Concept DCP

0d 10/28

(c) Establish

constraints

6d 9/8

(n) Reproduce

documents

8d 10/14




Critical Path is not stable!

Determine budget

4 1d

Wed 6/1/94 Wed 6/1/94

Selection

6 120h

Thu 6/2/94 Wed 6/22/94

Theme

7 2d

Thu 6/2/94 Fri 6/3/94

Date

8 3d

Mon 6/6/94 Wed 6/8/94

Site

9 1w

Thu 6/9/94 Wed 6/15/94

Costumes

10 1w

Thu 6/16/94 Wed 6/22/94

Hire

11 168h

Thu 6/23/94 Thu 7/21/94

Caterer

12 2d

Thu 6/23/94 Fri 6/24/94

Entertainment

13 9d

Mon 6/27/94 Thu 7/7/94

Keynote speaker

14 2w

Fri 7/8/94 Thu 7/21/94

Public relations

15 88h

Fri 7/22/94 Fri 8/5/94

Rent Equipment

18 40h

Mon 8/8/94 Fri 8/12/94

Invitation list

5 1d

Wed 6/1/94 Wed 6/1/94


10 © T. Lechler, Ph.D., 2013


– Finish early — do we gain?

• Can the next task start early?

• Good luck is wasted.

– Finish late — do we lose?

• The next task is forced to delay

• Delays are accumulated.

Good luck is wasted and bad luck accumulates!

Task1: 10 Days

Task2: 10 Days

Task3: 10 Days +2 D

Delay 2 Days

Due Date

Delay 2 Days

+2 D

Task1: 10 Days

Task2: 8 Days

Task3: 10 Days

2 D

Due Date Not

2 Days Earlier

2 Days Earlier


11 © T. Lechler, Ph.D., 2013


Task safety wasted:

– as a student, did you start on assignments immediately?

– you had plenty of time, so you started later!

Task safety is wasted right at the beginning

3

10

90%

A

B

wasted

7

70%student

syndrome

50% Task estimates include

safety, but what if we

start later?

Student Syndrome


12 © T. Lechler, Ph.D., 2013


Report early finish after 5 days?

– There is always a bit more to do (polishing)

– The next person isn’t ready to start anyway

– Next time the estimate gets cut

We erroneously report when we’re done

actual50% 90%

1|0 |0 |5 1|5

time

pro

bab

ility

5 ?

Task estimates include

safety, but what if we

get lucky, and finish

early?

Parkinson’s Law


13 © T. Lechler, Ph.D., 2013


Slow starts and false finish reporting contaminate project metrics:

• Safety is systematically wasted!

• Initial expectations seem to be unrealistic…

expectedactual

B

3 2

student

syndrome

false finish

reporting

5

© 2000 ZULTNER & COMPANY

pro

ba

bili

ty

Combined Effects of:

Student Syndrome and Parkinson’s Law


14 © T. Lechler, Ph.D., 2013

1. Paradigm Question: CPM Limitations

Limitations of Critical Path Methods (CPM)

• CPM resource allocation leads to minimal single

project duration (local optimum)

• CPM does not explicitly take variation (risk) into

account

• CPM does not maximize the throughput of a multi-

project system (global optimum)

Critical Path Method leads to sub-optimal resource allocation!


15 © T. Lechler, Ph.D., 2013

1. Paradigm Question: CPM Vicious Cycle

The Vicious Cycle of CPM

– More and more safety is added

over time

How to break the vicious cycle?

– Applying Critical Chain

Project

Is Late

Cut Schedule

“it’s just too long”

Add

Safety

Struggle

to Deadline

Blame

Assigned

Goldratt, Viewer Notebook, 1999, pp.75


16 © T. Lechler, Ph.D., 2013

1. The Paradigm Question: System Problems

It’s the system, stupid!

“Pit a good person against a bad system and the system will

win all the time.”

(Rummler and Brache, 2002)

Deming’s 85/15

• 85% of faults are process related, and it is management’s

responsibility to solve them

• 15% of faults are the responsibility of individual

employee’s

• Most of the time management is focusing on the “15”

rather than the “85”, trying to find the guilty person rather

than to improve the process Deming, W.E., “Out of the Crisis”, MIT, CIA, Massachussetts, 1986


17 © T. Lechler, Ph.D., 2013

“If you do what you always

did, you’re gonna get what

you always got”

Yogi Berra

1. The Paradigm Question: Why change?


18 © T. Lechler, Ph.D., 2013

1. The Future of Project Management

• Critical Chain allows to shorten your projects

• minimum 15-25% reduction

• on large projects, often more than 25%

• NO added resources (and less overtime!)

• NO sacrifice of value or features

• NO increase in risk

• NO cutting of quality

Date/reference/classification 19

Critical Chain

Performance Impact


20 © T. Lechler, Ph.D., 2013

On-time performance N CPM N CC

Greater Than 90% 9 7% 8 53%

80% To 90% 12 10% 3 20%

70% To 80% 18 15% 1 7%

60% To 70% 20 17% 0 0

50% To 60% 18 15% 1 7%

40% To 50% 15 12% 0 0

Less Than 40% 27 22% 0 0

No Response 2 2% 2 14%

Results from: http://www.pdinstitute.com/surveys/surveyresults.htm

Question: On-time performance for the projects of my organization is:

CC promises higher due date performance!

2. Critical Chain Performance Impact: Average


21 © T. Lechler, Ph.D., 2013

2. CC Performance Impact: Selected Cases

• Harris Corporation

• Microelectronics plant, Mountaintop, PA

• Project Goal: To be operational in 27 months

• Facility Construction:

• Critical Chain Plan: 18 months

• Result with CC: 13 months, 4% over budget

• Industry norm: 28-30 months

• Facility Operation:

• Result with CC : Full production in 21 months

• Industry norm 46-54 months total


22 © T. Lechler, Ph.D., 2013


Lucent Technologies: Fiber Optic Cable Business Unit

Situation before CCM • Product realization cycle for

Fiber Optic Cable was same as

its competitors.

• Designs based on tools took

longer time to come to market.

• Many designers were overloaded

and multi-tasking on several

projects simultaneously.

Situation After CCM • Goals of implementation were

achieved.

• On-time delivery was markedly improved

• The organizations capacity to develop products markedly increased.

• With no increase in staff, the number of projects were completed.

• The production introduction interval was reduced by 50%.


23 © T. Lechler, Ph.D., 2013


Successful Cases of Companies That Used Critical Chain

• (http://www.goldratt.com/success.htm)

• Seagate Technology

• F-22 Project Raptor

• Lucent Technologies

• The Clowes Group

• Valmount Industries

• BAE Systems (this article can be viewed

in(www.ets.news.com)


24 © T. Lechler, Ph.D., 2013

CC-Project Before After

Warfighter Systems

Testing (US Air Force

Operational Test &

Evaluation Center)

18 projects in 6 months.

On time delivery unknown.

26 projects in 6 months.

75% projects on time;

30% reduction in cycle time.

Aircraft Repair and

Overhaul (US Naval

Aviation Depot, Cherry

Point)

Average turnaround time (TAT)

for H-46 aircraft was 225 days.

Throughput was 23 per year.

Reduced TAT to 167 days, a 25%

reduction while work scope was

increasing.

Throughput is 46 per year.

70% reduction in backlog

Submarine Maintenance

and Repair (US Naval

Shipyard, Pearl Harbor)

Job Completion Rate 94%

On-time delivery less than 60%.

Cost per job was $5,043.

Job Completion Rate now 98% On-time

delivery 95+%.

Cost per job reduced 33% Overtime

reduced by 49%

$9M saving in first year.



25 © T. Lechler, Ph.D., 2013

2. CC Performance Impact: Summary

The Impact of CC across approx. 80 cases:

• Increased systems throughput ~ 20%

• Reduced project schedule ~ 15% - 40%

• Increased on-time delivery ~ 93%

• Reduced backlog ~ 30% - 70%

• Reduced overtime ~ 20% - 50%

CC shows dramatic performance improvements!


26 © T. Lechler, Ph.D., 2013

3. What is CC? Critical Chain vs. CPM

Traditional PM scheduling problems:

• Resource conflicts

• Delays

• Uncertainty (scope change, context, resources)

CC offers a solution:

• Performance improvement with same resource base

• Reduces resource conflicts

• Reduces uncertainty

• Addresses multi-project environments


27 © T. Lechler, Ph.D., 2013

3. What is CC? Critical Chain vs. CPM

Critical Chain three level approach:

1. Philosophical Level: Theory of Constraints

2. Single-Project Level: Managing Variation

3. Multi-Project Level: Systems Approach

Critical Path one level approach:

1. Single Project Level: Managing Due Date

• Does not account for variation

• Does not account for behaviors

• Does not account for multi-project system

Critical Chain promises advantages over CPM!


28 © T. Lechler, Ph.D., 2013

3. What is CC? Philosophical Level: ToC

Theory of Constraints:

1. Systems perspective

2. Focus on the system’s bottle neck

3. Throughput mindset

4. Avoid sub-optimization

5. Use simple tools

Elyahu Goldratt, “The Goal,” 1988


29 © T. Lechler, Ph.D., 2013

3. What is CC? Philosophical Level: ToC Steps

Theory of Constraints:

Step 1: Identify the system's constraint(s)

Step 2: Decide how to exploit the system’s constraint(s)

Step 3: Subordinate everything else to the above decision

Step 4: Elevate the system’s constraint(s)

Step 5: If in the previous step, a constraint has been broken go

back to step 1, but do not allow inertia to become the

system’s constraint

Elyahu Goldratt, “The Goal,” 1988


30 © T. Lechler, Ph.D., 2013

3. What is CC? Philosophical Level: CC Estimation

Average Work

Touch Time

Uncertainty

Buffers

Multitasking

Minimize

Set-up Time

Basic Estimation Equation: W = U * D


31 © T. Lechler, Ph.D., 2013

• Realistic task estimates are 90% estimates.

• They include safety

– to assure on-time completion despite unknowns

In 90% of the cases the task will be finished EARLIER!

In 10% of the cases the task will be finished LATER!

55

50%

90%

1|0 |0 |5 1|5

time

prob

abili

ty

10%

10

expected safety commitment

task duration

4. What is CC? Single-Project Level: Local vs. Global Safety


32 © T. Lechler, Ph.D., 2013

4. What is CC? Single-Project Level: Local vs. Global Safety

Is trying to keep every task on-time an efficient way to make the project deadline on time?

Probability of meeting the due date 73%!

5

50%

90%

5

50%

90%

5

50%

90%

105

5 10

105

traditional

committed

project

duration:

30 days

D

E

A

D

L

I

N

E


33 © T. Lechler, Ph.D., 2013

4. What is CC? Single-Project Level: Buffer Sizing

55

25 75

8.665

25 25+ + =

project buffer

safetytake

square

root

square standard deviation

50% 90%

5 5

5

55

duration

55

safety

duration

duration

safety

Single-Project Level: Local vs. Global Safety

How much buffer is needed for a 90% Due Date estimate?

Goldratt’s Heuristic: Buffer Size = 50% of accumulated safety!


34 © T. Lechler, Ph.D., 2013

4. What is CC? Single-Project Level: Buffer Sizing

An efficient way to manage risk:

– pool task variation in a project buffer,

– the project schedule is shorter,

– risk is not increased!

555

expected duration safety

8.66

50%

90%

project buffer

savings

15 6

© 2000 ZULTNER & COMPANY

Critical

Chain

committe

d project

duration:

24 days

D

E

A

D

L

I

N

E


35 © T. Lechler, Ph.D., 2013

4. What is CC? Single-Project Level: Buffer Management

2|02|5 1|0 |0

95 5 5

-1 day

minimum buffer required

6 5 85

3 105

+2 days

510

-5 days

finished!

2

1

3

1|5 |53|0

8.6

7.1

5.0

actual buffer

+1 buffer-days

+5 buffer-days

+5 buffer-days

BUFFER STATUS

How many buffer-days

are needed to meet the

due date with 90%

probability?

How could good and

bad luck averaged

out?

Buffers absorb risks

CC Buffer Metric: Actual Buffer/Minimum Buffer Required


36 © T. Lechler, Ph.D., 2013

6:Prog 3:HW

5:HW

Feeding Buffer

2:CS

6:Prog 3:HW

3:Eng 5:HW 4:CS 2:CS Project Buffer

Resource Leveled Critical Path (in Red)

Critical Chain (in Red) Buffered Schedule

Individual activities are scheduled at their average durations (no safety

margin) 15%-25% decrease of project duration

4:CS 3:Eng

CPM Network

CC Network

4. What is CC? Single-Project Level: Feeding & Project Buffers

Due Date


37 © T. Lechler, Ph.D., 2013

4. What is CC? Single-Project Level: Resource Buffers

For Relay Runners

– resources get a countdown

– assure the resource can start immediately

– activities start as soon as possible NOT when the schedule says

3

10

A

B

Resource

Buffer

7

start!

10

current task next task

3

10

2

1

three days to go

two days to go

one day to go

good luck strikes—early finish

the plan


38 © T. Lechler, Ph.D., 2013

4. What is CC? Single-Project Level: Relay Runner Effects

Resources seem to be working “faster” but

• Overtime goes down

• Speed may go down

• Throughput goes way up

• Quality goes up (slightly to a lot)

• Projects take much less time

• Resources will be idle more often

• Resource utilization goes down


39 © T. Lechler, Ph.D., 2013

4. What is CC? Single-Project Level: Computation Steps

Critical Chain computation steps for single project schedule:

• Compute baseline schedule using average activity durations and ALAP

• Aggregate safety margins into Project Buffer

• Identify Critical Chain (CC)

• Protect CC using Feeding Buffers

• Try to keep baseline schedule and CC fixed during execution

• Use buffers as proactive warning system during execution


40 © T. Lechler, Ph.D., 2013

7. What is CC? Summary

CC (1996):

• Feeding Buffer

• Resource Buffer

• Project Buffer

CC Metrics

• Actual Buffer/ Minimum required buffer

• Due date performance (% of milestones finished or mean project duration and its standard dev.

• Operating expenses (# of hours invested)

• Inventory (amount of work in process not finished yet, hours invested in unfinished work orders/activities)

• Project Quality (# development cycles and # of changes)


41 © T. Lechler, Ph.D., 2013

7. What is CC? Summary

Critical Chain Advantages (CC)

• Provides a systems approach for managing multiple projects sharing a set of resources

• Improved system throughput (global optimum)

• Explicitly takes variation (risk) into account

• Efficiently! (reduces time to market)

• Provides a visible, and powerful way to manage risk and likelihood of on-time delivery

• A base for real risk management (reduces % of late projects)

Critical Chain promises advantages over CPM!


42 © T. Lechler, Ph.D., 2013


43 © T. Lechler, Ph.D., 2013

Implementing CC: Creating a Project Schedule

The Five Focusing Steps of ToC

1. Identify the constraint

2. Exploit the constraint

3. Subordinate everything else to the above decision

4. Elevate the constraint

5. Don’t let inertia become the constraint (If the constraint is broken, go to 1)


44 © T. Lechler, Ph.D., 2013

Implementing CC: Step 0

Before creating the CC project schedule

• Define the project and its purpose

• Define the fundamental measurements for the project


45 © T. Lechler, Ph.D., 2013


1. Identify the Critical Chain

A. Lay out the project network with all tasks as late as possible (latest completion time of tasks)

B. Working backwards, identify the contention (usually resource) to address next

C. Remove the contention by adding more resources, or moving tasks earlier

D. Continue until all conflicts are resolved

E. Identify the longest dependent chain


46 © T. Lechler, Ph.D., 2013

Implementing CC: Step 2-4

2. Exploit the Critical Chain

Use 50% task times and a project buffer

3. Subordinate everything else

• Protect the CC with feeding buffers

• Resolve new conflicts from buffers by moving tasks earlier

4. Elevate (shorten) the Critical Chain

Add resources, change procedures, etc.


47 © T. Lechler, Ph.D., 2013


5. Go back to step 1. Do not allow inertia to become the constraint!

We have now applied the Five Focusing steps of Theory of Constraints

(ToC) to schedule a project


48 © T. Lechler, Ph.D., 2013

Implementing CC: Case 1

• Variation…

– Non-critical path

• if it slips, could impact the critical path

– Protect the critical path — with a feeding buffer

• absorbs non-critical path variation

• prevents the critical path from shifting

10 10 16

20

1616

A

5 5 8

FB 1188

C

PB 1410

CPM Project Plan

CC Project Plan


49 © T. Lechler, Ph.D., 2013

20 60

20 40 20

60

Airframe +10

Component Aircraft

Complete

20 30

30


Let’s establish a Critical Chain Schedule

Aircraft Project – Traditional Schedule

Engine +20

Critical Path: 140 Months

Aircraft

Start


50 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15


Cut task estimates by 50%

Aircraft

Start


51 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component

Aircraft

Complete

10 15

15


Push all tasks to start as late as possible.


52 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component

Aircraft

Complete

10 15

15


Eliminate Resource Contention


53 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15


Resource contention eliminated by moving

Component branch 15 months backward.

+15


54 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15



+15


55 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15



Engine branch 5 months backward.

+5

+15


56 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15



+5

+15


57 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15



Airframe branch 5 months backward.

+5

+5

Aircraft

Start +15


58 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15


The longest path through the network considering both

TASK and RESOURCE DEPENDENCIES is the Critical

Chain (CC = 70 months).

+5

+5

Aircraft

Start

CC

+15


59 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15

Implementing CC: Buffer Management (Project Buffer)

Using Buffer Management: Project Buffer

Project Buffer = 38.08 months (accurate calculation)

Total project duration = 108 months

+5

+5

Aircraft

Start

Project Buffer

CC

+15


60 © T. Lechler, Ph.D., 2013

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10 15

15

Implementing CC: Buffer Management (Feeding Buffer)

Using Buffer Management: Feeding Buffers

Locations of feeding buffers

+5

+5

Project Buffer

+15


61 © T. Lechler, Ph.D., 2013

10

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10

Implementing CC: Buffer Management (Feeding Buffer)

Using Buffer Management: Feeding Buffers

Durations of feeding buffers moves project start to an

earlier start date!

Total project duration: 129 months

15

30

15 PB=38 FB=26.5

FB=30

FB=10


62 © T. Lechler, Ph.D., 2013

Implementing CC: Conclusions

10 30

10 20 10

30

Airframe

Engine

Component Aircraft

Complete

10

Use Project Buffer

Use “Float” as “quasi feeding buffers”

Total project duration = 108 months compared to

Standard schedule with 140 months

Aircraft

Start

Project Buffer

CC

15

15 FB=5

FB=5

Education

Lecture09 cc