Transcript Document

Critical Chain
Project Management
“Cruise Control for Projects”
Skip Reedy
Theory of Constraints Jonah
TOCICO Certified Critical
Chain Project Management
© 2012
Skip Reedy
1
Standish Group
“Chaos Report”
“There is no reliable way
to measure project status
until it’s too late.”
2
© 2012
Skip Reedy
Theory of Constraints
Dr Eli Goldratt
Every system has something that limits it.
If you identify the system constraint and help it,
you get more output from the system.
Critical Chain
Project Management
This kink is the
constraint of this
simple system.
The constraint of a project
is the Critical Chain.
Help it to get done faster and
the project gets done faster.
3
© 2012
Skip Reedy
PMI
A GUIDE TO THE PROJECT
MANAGEMENT BODY OF KNOWLEDGE
Chapter 6 – Project Time Management
Critical Path
Critical Chain
Method
(CPM)
Project
Management
2 + pages
(CCPM)
1 + pages
4
© 2012
Skip Reedy
comment on PMI CCPM forum
"Critical
chain is crazy.
It is based on an assumption that we consistently
over estimate durations.
My experience is that we typically under estimate
durations … then [Critical Chain says to]
arbitrarily reduce durations to comply with critical
chain thinking …”
5
© 2012
Skip Reedy
What is a Project?
“A project is a combination of
dependent events and
statistical fluctuations,
convoluted by human behavior
and purported to accomplish a goal.”
6
© 2012
Skip Reedy
Statistical Fluctuations
How long will a project take?
Task estimates are needed to plan a project.
The estimates together determine the project length.
A
B
C
D
Project managers want their tasks to be on time,
because if the tasks are on time, their projects will be on time.
7
© 2012
Skip Reedy
Estimating
How long does it
take to Catch a Fish?
At Pike Place Market in Seattle?
8
© 2012
Skip Reedy
Test your estimating skill!
Estimating a task with little information or experience.
Write your estimates on the handout page.
1 How long will it take a horse and rider to travel
20 miles cross country?
© 2012
Skip Reedy
9
Variation in Tasks
Commute to Downtown
Occurrences
Leave Home
On-time Probability
90%
50%
60
0
25
50
Minutes
75
Duration Estimate
© 2012
Skip Reedy
100
120
10
Duration Estimates for Planning a Project
Optimistic
5 days
Likely
10 days
Pessimistic
20 days
Expected
11
5 days
A
A
10 days
A
20 days
20 days
Project Lengths
40 days
80 days
11
© 2012
Skip Reedy
Probability of a longer task
Subconsciously we think the probability curve
doesn’t move when we increase the estimate.
A
A
A
A
12
© 2012
Skip Reedy
Project Scheduling Software
Turns estimates into deadlines
A
A
1/26
2/9
2/10
Like Milestones
2/23
B
2/24
Like Deadlines
3/8
C
3/9
D
3/22
Software Schedule
Project
Start
1/26
3/22
Project
End
13
© 2012
Skip Reedy
Deadline Management
means “Be NOT Late”
Long estimates are self-fulfilling prophesies
14
© 2012
Skip Reedy
Deadlines and
Convoluting Behaviors
Parkinson’s Law
Work expands to fill the available time
when there isn’t enough to do.
A
15
© 2012
Skip Reedy
Deadlines and
Convoluting Behaviors
The Student Syndrome
Effort
Due Date
Time
16
© 2012
Skip Reedy
Multi-Tasking -
is the worst
This is how long each task or project should take if full attention is paid to each in turn.
Effort on Blue is interrupted by other projects.
Blue’s duration
Switching projects takes
time.
17
© 2012
Skip Reedy
Multi-Tasking demo
Delays caused by interruptions and switching tasks
Multi-task by writing a letter, number, letter, number, …
M
U
L
T
I
T
A
S
K
9
8
7
6
5
4
3
2
1
Total Time required ______
18
© 2012
Skip Reedy
Multi-Tasking demo
Do NOT Multi-task.
Write all the letters, then write all the numbers.
M
U
L
T
I
T
A
S
K
9
8
7
6
5
4
3
2
1
Total Time required ______
19
© 2012
Skip Reedy
Effort
The Student Syndrome & Multi-Tasking
Time
20
© 2012
Skip Reedy
The Student Syndrome &
severe Multi-Tasking
Effort
Or maybe like this!
Time
21
© 2012
Skip Reedy
Estimates become Actual Durations
Our allowances for “stuff” become part of the actual duration
Task Estimate is a Number
S/W schedules Start & End Dates
End Date = Due Date  Milestone = Deadline
Human Behaviors meet Deadlines
Tasks finish when Due
Pick a Number, get a Duration
22
© 2012
Skip Reedy
Lates Accumulate
Project
Start
Project
Due
3/22
A Late
A Pushed
C Late
A Pushed + C Pushed
Late
Project
Due
A
B
C

D
The probability curve and deadline move as tasks B, C & D are pushed by A.
23
© 2012
Skip Reedy
What do we know!
We don’t know when anything will finish.
Yet, tasks often get done about when they are estimated.
Allowances for delays & problems get used up.
EARLY usually doesn’t happen, and LATE often does.
Being on time is hard to do.
24
© 2012
Skip Reedy
The Pony Express
needs YOU!
25
© 2012
Skip Reedy
Short Estimates
Project
Duration
The schedule doesn’t fit the business case.
Business
Case
!!
This schedule fits the business case.
A

B
B

C
However, each task must start
before its predecessor ends.
C
In order to make it work, I’m asking
you to use aggressive 50% durations.

D
?
D
Critical Chain
tasks

26
© 2012
Skip Reedy
The Critical Chain
27
© 2012
Skip Reedy
Good Statistics
Help the Pony Express
The variability of a series of tasks is less than
that of the sum of the individual task variabilities.
Series
flow time
probability
Central Limit Theorem
28
© 2012
Skip Reedy
The Feeding Chains
29
© 2012
Skip Reedy
Critical Chain Buffers
30
© 2012
Skip Reedy
Buffer Management
Fever Chart
Late
Due
Project
Fever
Chart
Early
31
© 2012
Skip Reedy
Work Rules
Critical Chain Project Management
Prioritize work (tasks & projects)
Use aggressive task estimates
Identify the Critical Chain
Aggregate safety in buffers
Use Buffer Management
Work quickly on one task
32
© 2012
Skip Reedy
Multi-Project Critical Chain
33
© 2012
Skip Reedy
Multi-Project Fever Chart
34
© 2012
Skip Reedy
Consider Unloading Ships...
(Projects)
5 ships + 5 persons @ 5 person-days to unload each
= 5 days
35
© 2012
Skip Reedy
Staggering the Projects
Put all five resources on
ship 1 the first day, ship
2 the second day, etc.
Initial Results:
Ship
New (days) Old (days) Saved
1
1
5
4
2
2
5
3
3
3
5
2
4
4
5
1
5
5
5
0
Nobody loses. Four of five clients done sooner.
Cost = $ 0
Continuing Results:
Each arriving ship is
unloaded in 1 day.
36
© 2012
Skip Reedy
Confluence
37
Boeing Commercial Airplanes
Application of CCPM to 777-300 ER Airframe Design
55% of the Airplane
10,000 projects
85% of the Airframe
1,000 people at (peak)
20% larger than any
previous derivative
$500,000,000 program
Adapted from a presentation by Stephen R. Johnson
777 Airframe Integration Team Chief Engineer (AFIT)
38
© 2012
Skip Reedy
777-300 ER
Challenges & Results
Load data > 2 months late
Experienced lead engineers resisted CCPM for
more than 6 months
26% faster than its very aggressive schedule
50% reduction* in engineering errors
All end-items shipped on time
Improved engineering release compliance ~15%
* Relative to the last 777 derivative: 777-300
39
© 2012
Skip Reedy
The Next Derivative
with CCPM
 777-200 LR airframe design was
completed 5 months earlier than a
very aggressive schedule
40
© 2012
Skip Reedy
Be EARLY, NOT LATE
Critical Chain is EARLY
3 Ideas to Use Today
1. The Critical Chain is the constraint. Focus on it
so it’s always being worked.
2. Pink Paper Pyramids keep the distractions away.
3. Ask 3 Status Questions: What’s going on? What
are you doing about it? and Do you need any help.
41
© 2012
Skip Reedy
Questions?
42
Skip Reedy
Skip is certified in Critical Chain Project Management.
He is a member of the Project Management Institute and the Theory of
Constraints International Certification Organization. He was with Boeing
Commercial Airplanes in Seattle as a Manufacturing Engineer and Design
Engineer for 8 years and then as an internal CCPM consultant for six years.
For the last eight years he is an independent project management consultant
with CCPM Consulting LLC.
He is a mechanical engineer from the University of Detroit with an MBA from
Loyola University of Chicago. He is a Theory of Constriants Jonah and is
certified by TOCICO.
Skip can be contacted at 480-648-1862
[email protected]
http://ccpmconsulting.com
43
© 2012
Skip Reedy
Marine Corps Logistics Base
TOC & CCPM implementation 8 minute
video is available at
http://www.ccpmconsulting.com/marine-corps-logistics-base-albany-ga/
44
© 2012
Skip Reedy
Traditional Project Management
Estimates become NOT LATE
Task Estimates
Confident estimates are long enough to get the work done, allowing for typical delays and problems.
Probability as a Number
An estimate is not a prediction. We don’t know how long anything will take.
It’s a single number for a task that has the possibility of a range of durations.
Crazy Project Management Software
Project management software schedules estimates with start and end dates.
End dates become due dates, and deadlines. Deadlines must be met.
Convoluted Normal Behaviors
Parkinson’s Law, the Student Syndrome, Multitasking, Deadlines reinforce be NOT LATE.
LATEs, not EARLYs
Tasks are rarely EARLY and often LATE. LATE tasks push successors to be LATE.
LATEs accumulate. EARLYs don’t.
Allowances are Ineffective
Allowances for delays and problems (Safety, Contingency, or Buffer time) are effective against being NOT
EARLY and NOT LATE. This safety is wasted.
Increasing the safety makes projects longer
LATE tasks encourage future estimates to be longer. Longer estimates make longer projects.
Longer projects struggle to be NOT LATE.
Safety in tasks creates longer projects without improving on-time results.
45
© 2012
Skip Reedy
Shameless Advertising
Original Pink
Pyramid
purported to accomplish a goal.
Fold in
Management
110 # card stock works well.
Using a straight-edge, run a dull
knife along/between the fold lines.
Suggested warnings:
What I am working on is more
important than whatever it is
you are about to ask me.
So don’t do it!
Tape together
close to the corners.
If you are delivering money,
please interrupt.
Fold in
Critical
A project is a combination of depende
© 2012
Skip Reedy
46