Critical Chain Program Management and Earned Value Management

1
Critical Chain Program Managementand Earned
Value Management

• Gregg Tiemann, Boeing
• Eleanor Haupt, USAF
• ASC/Industry Cost/Schedule Workshop
• 5-7 April 2000

2
EVMS vs. CCPM?

• I only hope it is cast, not as EVM vs. CCPMbut
  as how to improve PM using the best tools for the
  project. Wayne Abba

3
Agenda

• Review of progress to date
• Answering the Challenges
• Quick Review of CCPM and EVMS Processes
• Issues and Recommendations
• Contractor Implementation
• Baselining
• Metrics
• Government Implementation
• Philosophy Implementation
• Earned Value criteria
• Benefits
• Open Discussion
• Review of Action Items

4
Review of Progress to Date
5
Status

• Solicited volunteers for working level team
• current e-mail roster
• 28 experts
• 4 major defense companies
• CCPM and EVMS experts
• Co-Leaders
• Gregg Tiemann, Boeing
• Eleanor Haupt, USAF, Wright-Patterson AFB
• Created website to share information
• http//home.att.net/ccpm
• Conducted working group meeting
• 9-10 March, WPAFB
• 4 panels baseline, metrics, government,
  implementation philosophy

6
Purpose of Team

• Explore, develop, and recommend methods to
  integrate critical chain scheduling and EVMS in a
  complex project environment
• Proposed methodology will be presented and
  discussed in workshops
• ASC/Industry Cost and Schedule Workshop (Apr 5-7)
• College of Performance Management Conference (May
  15-17, Clearwater Beach FL, www.cpm-pmi.org)
• Results to be documented in Schedule Reinvention
  Teams schedule resource handbook
• E. Haupt has lead

7
Attendeesof 9-10 MarWorking Meeting
8
Goals for 9-10 Mar Meeting

• Identify major issues
• Identify potential solutions to major issues
• make recommendations
• Identify any disconnects with current DoD policy
• EVMS criteria
• RFP language

9
Answering the Challenges
10
Answering the Charges

• Tell me how you will measure me and I will tell
  you how I will behave if you choose to measure
  me illogically, do not be surprised by illogical
  behavior. Eli Goldratt
• EVMS forces focus on local optima (task
  completion) instead of the global optima (project
  completion)
• EVMS fosters starting tasks early to earn value
  and results in bad multi-tasking.
• You are absolutly right. Using Earned Value in
  managing and controling a project is devostating
  as I've repeatedly proved in Critical Chain.
  Eli

• Agree with statement. The measure of EVMS
  success has always been getting the project done
  on time and at cost. This should be our
  motivating principle.
• Dont agree that it is black or white.
  Corrective actions should always be focused on
  how the task impacts the global goal of the
  entire project.
• Nonsense. The EVMS baseline should be laid in to
  a realistic schedule (CCPM or not).
• You are absolutely wrong. EVMS has been used
  successfully for over 30 years. It is now
  migrating to the private and worldwide sectors.
  EVMS is only as good as the baseline planning.
  Eleanor

11
Resolutions

• Resolved
• The purpose of critical chain scheduling is to
  achieve breakthrough success, reduce cycle time
  of projects, and ultimately reduce cost.
• The purpose of EVMS is the same, and in fact,
  EVMS baselines should be based on well planned,
  achievable schedules.
• We recognize that...
• - CC and EVMS are not mutually exclusive
• - that EVMS application is flexible enough to
  integrate with CC
• without driving additional reporting
  requirements

12
Quick Review of CCPM and EVMS Processes
13
Critical Chain Scheduling

• Lay out tasks, assuming 90 probability for
  durations
• Network the tasks, perform backward pass,
  establish late finish dates (critical path
  established)
• Identify resource constraints and critical chain
• Assign duration with 50 probability to tasks
• Deconflict the resource constraints
• Protect the critical chain
• Establish resource buffers on schedule legs
  before constraining resource
• Establish feeding buffers on schedule legs that
  feed into the critical chain
• Establish program buffer (assign half of interval
  between the end of the 50 and the 90 probable
  schedule)

14
Program Buffer
Normal Schedule, at around 90 probability for
task durations
Reduce durations to 50 probability
Divide difference by 2, assign as program buffer
resulting program buffer is about 33 of program
15
Resource Alerts Assure CC Resource Availability
Protects due date from CC variation
RB
RB
Task A
Task B
Task C
RB
RB
Task F
Task G
Program Buffer
Task J
Task D
Task E
FB
Due Date
Task H
Task I
FB
Aggressive Target Duration Estimates
Protects CC from non-critical task variation
16
EVMS Baselining

• Lay out tasks
• Network the tasks
• perform forward pass to establish early finish
  dates
• perform backward pass, establish late finish
  dates to meet milestones
• critical path established
• Program manager establishes assumptions for
  baseline dates
• early or late finishes or somewhere in between
• dependent on risk and overall schedule
  constraints
• Networked schedule (with milestone dates) used as
  basis for resource allocation
• Budgeted Cost of Work Scheduled (BCWS)

17
Issues and Recommendations

• Contractor Implementation

18
Baseline PanelPotential Issues

• sizing of buffers
• assign to buffers?
• is 50 the right probability?
• use of milestones
• change process/control
• initial program duration

19
Baselining

• Issue Sizing of buffers
• Discussion
• may be constrained by contractually mandated
  dates
• Recommendations

20
Baselining

• Issue should resources ( or hours) be assigned
  to buffers?
• Discussion
• con
• buffers may cover several WBS or OBS elements
• any task on this leg may cause usage of the
  buffer, therefore the scope of the increased work
  may not be known until it happens
• if the time buffer is not used, then it goes
  away, causing a paperwork drill to pull back
  resources
• student syndrome if I have funding reserve,
  Ill spend it
• pro
• accumulation of added costs as buffer is consumed
  without being able to claim earned value drives
  poor CPIs
• leaving about 1/3 of program value in management
  reserve invites raiding
• some buffer is usually consumed, therefore
  requiring budget resources
• Recommendations
• alternative one
• may allocate management reserve to some lower
  level to protect buffers
• would not have work scope, WBS identity, or
  phasing

21
Baselining

• Issue is 50 probability for durations the best
  assumption?
• Discussion
• CCPM advocates using 50 probability for
  establishing task durations
• irregardless of nature of work (development,
  production, etc.)
• May insert too much risk into complex development
  efforts
• Team noted that typical DoD programs may actually
  start at less than 50 probability
• Recommendations
• may wish to use different probabilities based on
  program phase
• 70 for highly complex development efforts
• 60 for normal development efforts
• 50 for production efforts
• on complex DoD programs, perform Monte Carlo
  simulation to determine probability of success
  BEFORE cutting duration times

22
Baselining

• Issue protecting contract milestone dates
• Discussion
• CCPM ignores all milestones except the end of the
  program
• typical contracts include milestone dates, and
  can be used for payment purposes
• government programs report milestone dates to OSD
  and report progress against this baseline
• Recommendations
• government should be flexible in agreeing to
  milestone activities without set dates
• e.g. Preliminary Design Review has a duration
  of 3 days, actual date will float
• should work with contractor for insight into
  predicting milestone dates
• government may need to find alternate way to
  assess progress and make payments
• government should establish OSD goals on more
  conservative schedule (e.g., 90)
• or, possible to insert buffers before milestone
  dates
• or, phase program buffers before major milestones
  and treat as mini projects
• e.g., activity from award to CDR is a mini
  project, with project buffer

23
Baselining

• Issue Baseline change process and control
• Discussion
• contractor may need more flexibility in
  processing baseline changes
• Recommendations
• contractors should establish rules for use of
  schedule buffers
• contractors may wish to establish more
  flexibility in approving management reserve
• e.g., allocate MR to IPT leads, with authority to
  approve at IPT level

24
Baselining

• Issue Initial program duration
• Discussion
• government contracts may be based on unrealistic
  milestones, high concurrency between development
  and production
• probability may be less than 50 to begin with
• Recommendations
• outside the purview of this group
• continue acquisition reform efforts to minimize
  schedule risk during initial program planning
  (before release of RFP)
• on complex DoD programs, perform Monte Carlo
  simulation to determine probability of success
  BEFORE cutting duration times

25
Metrics PanelPotential Issues

• Ignore SPI?
• What are the appropriate metrics?
• Create a new index?
• Change variance reporting?
• Impacts to EAC predictions
• need demonstration projects
• drill down into detail problem areas

26
Metrics

• Issue Can or should we ignore SPI?
• Discussion
• SPI compares earned value to planned value
• if durations (and budgeted resources) are
  established at 50 probability, then it is likely
  that earned value will lag more more than normal
• SPI will be low, driving a lot of variance
  reporting
• SPI can still be a valid indicator, pointing to
  problem areas
• Recommendations
• use SPI to drill down to problem areas
• select alternate metrics to establish variance
  reporting (see other charts)

27
Metrics

• Issue What are the appropriate metrics?
• Discussion
• Need appropriate metrics to serve as early
  warning system
• Should be appropriate to new TOC culture
• manage the buffers
• Recommendations
• Focus on analyzing schedule variances through
  buffer consumption
• Program manager should establish simple tiered
  warning system, e.g.,
• 0 - 33 buffer consumed green (okay)
• 34 - 66 buffer consumed yellow (watch plan)
• 67 - 100 buffer consumed red (act)
• Tailor tiered system to different buffers or risk
  areas in program
• tighter control needed for resource buffers?
• Tailor color coding to percent complete

28
Early Warning - Notional
project buffer consumed
0 25 50
75 100
critical chain percent complete
29
Metrics

• Issue Create a new index?
• Discussion
• SPI metric not appropriate for culture of buffer
  management
• Recommendations
• Buffer Performance Index (BPI)
  complete
• 
  buffer consumed
• BPI of 1 or greater is good
• Others?

30
Metrics

• Issue How should variance reporting change?
• Discussion
• typical variance reporting thresholds would
  result in many more variances to report
• need to tailor CPR or C/SSR DID
• Recommendations
• contractor defined significant variances
• eliminate reporting on traditional schedule
  variances
• report on buffer consumption
• need to evaluate potential for additional cost
  variances to occur

31
Metrics

• Issue Impacts to EAC predictions
• Discussion
• some EAC formulas rely on a combination of CPI
  and SPI as a performance factor
• SPI factors that are significantly worse could
  overstate EAC
• Recommendations
• use new index (BPI?) in place of SPI in EAC
  formulas, e.g.,
• EAC ACWP (BAC - BCWP)
• (CPIBPI)

32
Metrics

• Issue need demonstration projects
• Discussion
• theory and proposed recommendations need to be
  tried out in a demo project
• concern can CCPM work for a very large and
  complex DoD project?
• Recommendations
• conduct smaller scale demo projects
• track alternate metrics
• analyze final results vs. metrics
• analyze organizational behavior in relation to
  metrics - what drove behavior?
• recommend best metrics for future projects
• dont implement policy until we have a few demo
  projects completed

33
Metrics

• Issue need to be able to drill down into detail
  problem areas
• Discussion
• balance need for insight into problem areas vs.
  desire to control amount of reporting
• Recommendations
• provide drill down to the detail necessary to
  understand cost and performance

34
Issues and Recommendations

• Government Implementation

35
Government PanelPotential Issues

• Status reporting
• impacts to FPRA business base
• Paying subcontractors
• Funding profiles
• Incorporation into IBR process
• Leveraging larger MR for out-of scope changes

36
Government

• Issue potential impact to government program
  status reporting
• Discussion
• quarterly OSD reports (Defense Acquisition
  Executive Summary) reports CV and SV and program
  office explanations
• other reports go to Program Executive Officers
  and service staffs
• SV could be misleading
• use of SPI to forecast EAC could result in
  Nunn-McCurdy breaches
• Recommendations
• seek OSD approval to report alternate metrics
• avoid use of SPI in EAC calculations

37
Government

• Issue Potential impacts to FPRA business base
• Discussion
• Are there significant changes from the baselined
  business base assumptions and the realized
  business base? In other words, we forecast
  indirect rates on the 50 probable schedule, but
  real performance causes a significant shift in
  the business base for the actual or applied rate.
• Recommendations
• Continue process of baselining rates at beginning
  of program and updating yearly to new forecast.
• May wish to allocate MR to protect potential
  difference in rates
• Educate DCMC at next quarterly DCMC earned value
  conference

38
Government

• Issue Timing of payment to subcontractors if
  they are on time, but prime is not
• Discussion
• What happens when a subcontractor delivers on
  time to the 50 probable schedule, but the
  primes performance has slipped, resulting in
  material sitting on dock prior to actual need
  date?
• Will government allow payment to subcontractor in
  advance of revised need date?
• Recommendations
• Get DCMC/DCAA buy-in to allow payment for on-time
  delivery
• Educate administrative contracting officers

39
Government

• Issue Government funding profiles can be a
  constraint
• Discussion
• Funding profile for development contracts are
  incremental
• Profiles are usually laid in before RFP is
  released
• Potential problem could result from mismatch
  between longer funding profile and more
  aggressive CCPM schedule
• Recommendations
• Program office should attempt to build funding
  profile to CCPM schedule
• Seek approval to treat development money as 2
  year money (which it is) for flexibility
• Establish contingent liabilities as needed

40
Government

• Issue How do we incorporate this into the
  Integrated Baseline Review process?
• Discussion
• How do we revise the IBR process?
• How do we perform a more global look at
  constraining resources?
• How do we evaluate a 50 probable schedule?
• Recommendations
• Alternatives
• Joint IPT participation in baseline development,
  (not a single point review)
• Phased IBR (work scope, then schedule, then
  budget)
• Phased IBR (global and top level assessments,
  then look at control accounts)

41
Government

• Issue Temptation to use Management Reserve (if
  much larger) for subsequent contract changes
• Discussion
• If not educated on CCPM, program office could
  lean on contractors to use large pot of MR for
  future changes
• Recommendations
• Education is key
• May wish to allocate and identify MR based on 90
  schedule

42
Issues and Recommendations

• Implementation Philosophy

43
Implementation Philosophy PanelPotential Issues

• Customer, Govt, DCMC Education
• Tools
• Impact to IPTs?
• Bargaining agreements?
• Apply to Subcontractors?

44
Implementation Philosophy

• Issue Customer education
• Discussion Commercial customer may not like
  size of buffers
• Recommendations

45
Implementation Philosophy

• Issue Government and DCMC education
• Discussion
• Recommendations

46
Implementation Philosophy

• Issue Availability of tool sets
• Discussion
• ProChain and Concerto are currently availability
  schedule tools for CCPM
• Flexible enough to handle the toughest DoD
  project?
• Impact to current EVMS analysis tools
• Recommendations

47
Implementation Philosophy

• Issue Impact to IPT Concept?
• Discussion
• discussed
• no impact seen
• Recommendations

48
Implementation Philosophy

• Issue Impact to bargaining agreements?
• Discussion
• Discussed
• no impact seen
• Recommendations

49
Implementation Philosophy

• Issue How does the prime apply CCPM to
  subcontractors?
• Discussion
• Recommendations
• Apply CCPM according to degree of risk

50
Benefits
51
Benefits of CCPM and EVMS Integration

• Better baselines and program execution
• Schedule and cost
• Better focus on enterprise resources during IBR
• Better tailoring of metrics (e.g. focus on buffer
  consumption vs. strict thresholds)
• Time saving
• CC forcing re-examination of tools that have been
  around but ignored or not adequately utilized
• Results in better project management

52
Earned Value Criteria
Youre under suspicion for violation of EVMIG
2.3 sub para 2a, questions 139 and 152
53
Review of Earned Value Criteria

• EVMIG criteria reviewed by entire team
• EV Criteria
• 2.2 May need to apply CCPM to subcontractors
  based on risk and complexity (prime contractor
  decision)
• Alternatively place feeding buffer where sub
  schedules feed into critical chain or have
  supplier place buffer on their schedule
• 2.5 Any buffer should have a unique WBS
  organization ID (e.g. program buffer may be
  assigned to the Program Mgr)
• 2.6 Should only have ONE schedule
• 2.7 Recommend insertion of buffers before key
  milestones (e.g. design reviews) IF there is a
  contractually required date
• 2.8 Recognized DoD awards contracts with
  aggressive schedules (RFP requirement or
  competitive environment)
• 2.12 LOE effort for time period of program
  buffer may be placed in management reserve at
  lower levels (e.g. at level of CAM for LOE
  activities)
• Note number refers to criteria contained in
  Earned Value Management Implementation
  Guide, Chapter 2,
  Criteria.

54
Review of Earned Value Criteria

• EV Criteria
• 2.13 Current process of baselining rates at
  beginning of program and updating yearly to new
  forecast should not significantly change if using
  CCPM as baseline schedule. There may be more
  variability between the baseline and new rates.
  May wish to allocate MR for this potential
  difference.
• Recommendation Have DCMC/DCAA look at this
• 2.23 conclusion variance analysis supports
  schedule variance on buffer consumption
• 2.24 no change (rate variance)
• 2.25 tailor reporting in DID

55
Review of Earned Value Criteria

• CONCLUSION
• No current problems seen implementing CCPM on a
  program that requires compliance with EV criteria

56
The final chart...

• Open Discussion
• Review of Action Items