A Critical Look At Critical Chain Project Management

1
A Critical Look At Critical Chain Project
Management
Best Practice Project Management

• Tzvi Raz, Ph.D., P.M.P.
• Tel Aviv University, Israel
• tzvir_at_tauex.tau.ac.il
• Robert Barnes, M.Sc., Dip. Mgmt, P.M.P.
• The iE3 Group Ltd, New Zealand
• robert.barnes_at_xtra.co.nz
• www.robertb.co.nz
• My presentation supported by

2
Agenda

• Introduction - What is CCPM?
• Critical analysis of CCPM
• based on the evidence in the research literature
  and in practice.
• CCPM - substitute for or addition to conventional
  PM.
• NB - I have significantly edited this
  presentation since the conference

3
Section 1. Introduction

• Critical Chain Project Management is -
• A fresh look at Project Management from Eliyahu
  Goldratt ( Critical Chain)
• Applies Theory Of Constraints (TOC) to Projects
• Claimed to achieve dramatic successes for
  projects, as TOC did for processes ( The
  Goal)
• Intuitively Plausible
• Seems to hold answers to many of the problems
  that have long troubled Project Managers
• Creating huge interest in the PM community
• eg, at least one paper on CCPM in most PM Network
  issues

4
Critical Chain Project Mgmt

• Is it New and Brilliant?
• or Largely Hype?
• First - what is it?
• (How does it differ from conventional PM?)

5
An Overview Of CCPM

• Step 1. Identify Critical Chain
• Initial Schedule -gt Critical Path. (A-gt B-gt D)
• Resource leveling -gt Critical Chain. (A-gtB-gt C-gt
  D)
• Critical Chain - The longest path through the
  network after resource leveling.

6
An Overview Of CCPM (contd.)

• Step 2. Make hidden buffers explicit.

Hidden Buffer
50
80
Unused Hidden Buffer is Wasted!
7
Why is hidden buffer wasted(1)?

• Origins in workers conflict -
• Want to finish tasks on time
• so need buffers
• Want to be productive (busy)
• need plenty of work, no buffers
• CONFLICT
• solve by taking on more work

8
Why is hidden buffer wasted (2)?
Non-critical Jobs
Another Job(Wkr B)
Another Job2(Wkr B)
Job 1(Wkr A)
Job 2(Wkr B)
Job 3(Wkr C)
Job 4(Wkr B)

• Wkr B knows that (s)hes not needed until day 6,
• gt Wkr B keeps busy with other work, and is not
  available until day 6
• gt if Wkr A finishes Job1 early, job 2 cant
  start anyway
• gt Wkr A knows that finishing early wont help
  the project
• gt rather than finishing early, Wkr A will use
  spare time on other tasks

9
Hence the schedule rule -

• If estimated durations are correct ON AVERAGE
• THE PROJECT WILL STILL BE LATE!
• Because
• Schedule gains are lost
• Schedule losses are passed on in full

10
CCPM Step 3. Buffer Management

• Step 3. Show buffers explicitly
• and pool them

Conventional Project Schedule
Task buffers are hidden
within individual tasks
Job 1
Job 2
Job 3
Job 4
CCPM Schedule
Buffers are pooled,
and made explicit
Project Buffer,
11
Feeding Buffers - the same principle, but on
non-critical paths
Project Buffer
Date 2
Date 1
Feeding Buffer
If Slack remains, then schedule as late as
possible
12
Resource Buffers - a Wake up call
Feeding Buffer
Critical Chain
Project Buffer
Alert Wkr A
Alert Wkr B
Resource Buffers
Alert Wkr C

• Adds neither Time nor Cost to the Project

13
CCPM Project Management (1)

• Focus on the Critical Chain
• Ensure resources give CC tasks priority
• Eliminate multitasking

14
CCPM Project Management (2)
Its OK for a task to be late
But not TOO Late
Focus on Buffer Consumption. Should be in
Proportion or better
15
Multiple Projects under CCPM

• Resources required to work simultaneously on
  several projects should give priority to the task
  of the one project that is in the greatest risk
  of missing its committed date, as measured by the
  remaining fraction of project buffer.

16
Section 2. A Critique of CCPM

• Does CCPM deliver?
• Anecdotal evidence suggests Yes
• Closer examination more ambiguous
• There are several points where CCPM assumptions
  can be challenged

17
1/ Hidden buffers may not exist

• CCPM assumes that
• all task owners pad estimates
• the actual duration expands to fill the time
  allotted.
• This is very Plausible,
• but theres no scientific support
• Research literature contains contrary evidence
  Hill et al. (2000)
• Personal experience - mixed. Some incorrigible
  optimists!
• Does reducing task duration reduce commitment?
• Personal experience - little effect, but suspect
  that there could be some risk of this
• Do we still get safety margin on top of buffer?
• Probably yes.

Parkinsons Law
18
2/ How big should explicit buffers be?

• CCPM does not provide any scientific or objective
  basis for determining the buffer size
• Default 1/3 suggested, based on probability
  distribution
• No empirical basis for using this probability
  curve
• CCPM uses statistical assumption of independent
  values (for task duration and buffer) - NOT TRUE
• Experience
• Discuss with task owner, get agreement
• Varies widely
• Optimists, pessimists - estimating errors dont
  cancel

19
3/ CCPM gives incorrect Resource Graph

• When do you need Worker A?

Gantt Chart
Resource Graph. Not this
But This
20
4/ What if theres insufficient room for feeding
buffer?
Conventional Plan - had free float
Date 2
Date 1
Should excess buffer be added to the critical
chain? Correct answer is NO, but what do you do
with it? (CCPM ignores this problem). Indicates
a near-critical path - could easily become the CP.
21
5/ The Critical Chain

• How good is the resource-leveled schedule that
  serves as the basis for CCPM ?
• As good as the algorithm used for its calculation
• The critical chain may change
• Feeding buffer exceeds the free slack of the
  feeding chain
• During project execution the critical chain may
  change as resource availability changes or as
  buffers are used
• This is true of conventional PM also
• but CCPM may add complexity

22
6/ Schedule control

• CCPM achieves schedule control by monitoring the
  extent of buffer penetration.
• The remaining-work estimate is also subjected to
  inflation by safety margins.
• CCPM assigns priority to the task belonging to
  the chain with the highest rate of buffer
  penetration.
• Ignores other important criteria, such as a
  projects strategic or financial impact.

23
7/ Multi-Project Management

• CCPM staggers the projects around the
  constraining resource.
• At any given point in time there could be several
  constraining resources, each leading to a
  different schedule.
• The premise that there is a single constraining
  resource comes from the steady state of
  manufacturing and operations environments.
• Consequently we doubt the applicability of the
  solution obtained with CCPM.

24
8/ CCPM Focus

• Project success and project management success
  are not equivalent.
• Project management success meeting agreed goals
  on time and budget.
• Project success benefits to customers and
  stakeholders
• Both conventional and CCPM deal with project
  management success rather than project success
• but CCPM focuses exclusively on ON TIME
• Focus
• CCPM meeting schedule goals by dealing with
  schedule uncertainty
• Conventional both cost and schedule.

25
9/ CCPM is presented as Alternative

• CCPM is presented as a revolutionary concept that
  replaces current project management knowledge and
  practices
• It is not properly integrated with the accepted
  body of knowledge and state of the practice.
• Dilemma to organizations that are new to project
  management choose between CCPM or mainstream
  methodologies.

26
10/ CCPM may have high Adoption Costs

• Specialized software tool needed
• CCPM presented as a methodology to be adopted as
  a whole ?massive re-education
• Culture change may be difficult -
• Give up ownership of the task duration
• Rely on common buffers to absorb deviations in
  individual task performance.
• Replace Due Date by Estimated Completion Date
  Range as represented by feeding and project
  buffers.
• No multi-tasking.

27
Section 3. Concluding remarks.

• CCPM has adapted the concepts of bottlenecks and
  material buffers from TOC for operations
  management, calling them critical chain and
  time buffers in projects.
• CCPM concepts are not necessarily new
• But does this matter? In our view, NO. CCPM
  would be worthy of study even it were only a
  packaging of old ideas
• and theres more to it than that!

28
The Key Question -

• Is CCPM as a methodology indeed superior to the
  currently accepted project management
  methodologies?
• Some dramatic successes have been claimed.
• However, we are not aware of studies comparing
  CCPM with a properly-applied conventional project
  management methodology.
• We DO NOT believe that CCPM is a better
  alternative to conventional PM.
• We DO believe that CCPM has good ideas which
  should be included in conventional PM
• PMBOK should take it more seriously

29
A Parallel
30
Good ideas from CCPM (1)

• Account explicitly for duration uncertainty
• Focus on Project on time, not Task on time
• Always consider resource availability
• What is the shortest realistic plan?
• Critical Chain is a useful term
• Focus on the key tasks and resources
• Reduce or eliminate multitasking
• Management focus

31
Good ideas from CCPM (2)

• Constantly monitor the amount of slack (buffer)
  in your schedule
• Provide flexible advance notice of upcoming work
  to critical resources

32
Finally, our suggestions

• Incorporate those CCPM elements that are
  applicable to your environment within a broader
  project management methodology.
• Consider the broader context
• Project success vs. meeting due date
• Reducing uncertainty vs. buffer management
• Adopt a comprehensive methodology
• Consider costs and benefits as well as time

33
Thank you for Listening

• Are there any Questions?