2010palisadeconf_michaelbrink_scheduleriskanalysis
DESCRIPTION
SchedulingTRANSCRIPT
Palisade Risk Conference
Introduction to Schedule Risk Analysis
using a Risk Driver Approach
Sydney, 20 & 21 October 2009
Presenter: Michael Brink
TBH CAPITAL ADVISERS CAPITAL ADVICE FOR THE LIFE OF THE PROJECT
CAPITAL
ADVISERS
Objectives and Purpose
• Purpose of presentation is to provide an introduction to Schedule Risk Analysis as a tool to understand project schedule and completion risk and to introduce the risk driver approach to Schedule Risk Analysis
• The Risk Driver approach to Schedule Risk Analysis in the presentation draws largely on approach outlined in the text “Practical Schedule Risk Analysis” by David Hulett.
Scope and Limitations of Presentation
• Time allocated is about 45 minutes so presentation is necessarily high level
• Assumption is attendees have some familiarity with Schedule Risk Analysis via tools such as @RISK for MS Project (ideally sat through Rishi’s MS Project @RISK presentation yesterday?)
• Hopefully provides enough of the flavour of the risk driver approach
• Anyone seeking a more detailed discussion I am in the process of developing a 2 day hands on workshop on this topic
Critical Path Method (CPM) raises a number of issues”
• Experience indicates that CPM scheduling does not always reliably identify the path that ultimately determines project completion date
• Possible drivers of this are: – Project Scheduling by is nature is a difficult discipline – Rules of Scheduling complex – Often owners/manager set unrealistic deadlines – Historically schedules have used point estimates for duration i.e.
deterministic approach
• Schedule Risk Analysis offers an approach that can address some of the inherent CPM weaknesses.
• Risk Driver Approach is a further refinement to Schedule Risk Analysis that focuses on specific risks driving schedule risk
Questions that Schedule Risk Analysis seeks to answer
• What is the likelihood of achieving target completion dates
• What level of contingency time should be allowed to provide a level of completion certainty acceptable to the project owners
• Where are the greatest risks in the project schedule
Dealing with Schedule Uncertainty “There are no facts about the future”
• Uncertainty in project schedule duration arises primarily due to: – uncertainty due to estimation error; and
– uncertainty due to risks and impact on schedule durations
• The Risk Driver Approach to Schedule Risk Analysis is focused on explicitly analysing schedule uncertainty arising from these two sources
Dealing with Schedule Uncertainty Uncertainty arising from Estimation Error
• Activity duration usually estimated based on:
– Knowledge of work done
– Resources available
– Productivity of resources available
– Reliance on other parties
• Expressed in terms of:
– Percentage below and above
– E.g. -10% +15%
Dealing with Schedule Uncertainty Uncertainty arising from Project Risks
• Project Risks leading to schedule uncertainty include:
– Technology risk
– Resource availability risk
– Resource productivity risk
– Supplier performance or delivery risk
– Regulatory risk
– Schedule Completion pressure
Applying Monte Carlo analysis to Schedule analysis
• Activity duration best represented by probability distribution
• Selecting best/most suitable probability distributions beyond scope of this presentation (triangular distribution is used in all following examples) but usual options are: – Triangular
– Beta distribution
– Normal
– Uniform
Visualising random sampling from a triangular distribution (Lotto model)
Most Likely Best Case Worst Case
Monte Carlo Random Sampling visualised
For each input distribution each Monte Carlo iteration
involves a single random draw from the bucket of balls
Typical Schedule Risk Analysis outputs
0
0 X <=14/06/2012
95%
0
0.2
0.4
0.6
0.8
1
01/04/12 26/04/12 21/05/12 15/06/12 10/07/12
Distribution for Complete/Finish
Schedule Contingency P80
P80 Completion Date
Deterministic Completion Date
Before discussing Risk Driver Approach will briefly consider:
• Need for high quality Schedule
– Limited time to discuss
– Specialised skill of TBH
• Collecting Risk Data
– Cultural biases to avoid
– Risk Interviews vs Risk Workshops
– Selecting appropriate distributions
– Focus on values in the tails min and max
Overview of the Risk Driver Approach
• Starts with risks that are prioritised in Risk Register, uses these risks to drive activity duration uncertainty
• Risk Interviews focuses on risks rather than impacted activity durations
• Probability distributions for each activity are derived based on probability and impact of all risks that are assigned to that activity
• Risks broadly considered in two categories – Risks that apply 100% of the time, and – Risks that apply less than 100% of the time
• Total Contingency reserve is explained in terms of risks that cause it and mitigation strategies can focus on risks rather than activities
Mechanics of the Risk Driver Approach
• Risk Register used to determine probability of occurrence and impact on schedule if risk occurs for prioritised risks
• Impact may be positive (=opportunity) or negative ( = threat) • Risks assigned specifically to each activity impacted
– One risk can impact several activities, and – One activity can be impacted by several risks
• Activities impacted by a single identical risk will be 100% correlated • Once completed overall project risk can be measured in terms of schedule
contingency • Impact of mitigation is measure against schedule contingency • Risk Driver approach is based on assumption that the risk drivers are base
level risk independent of each other (i.e. not correlated) but activities may be correlated
• Risk Interviews conducted at the same base risk driver summary level rather than at an individual activity level
Risk No 1 P=40% Factors
.95, 1.1, 1.15
Duration Activity A
Duration Activity B
Correlation between durations of Activity A
and B is 100%
100% correlation using Risk Driver approach
Risk No 2 P=40% Factors
.95, 1.1, 1.15
Duration Activity A
Duration Activity B
Presence of non common risks reduces Correlation
between durations of Activity A and B to 48%
Risk No 1 P=25% Factors
.8, .95, 1.05
Risk No 3 P=45% Factors
1.0, 1.1, 1.2
Less than 100% correlation using Risk Driver approach
Mechanics of the Risk Driver Approach
• Step 1 is to specify the Risks that are 100% likely to occur
– Typically this would relate to a background risk around estimating accuracy based on maturity of the project, for example could be a simple -5% best case and +5% worst case
– Or could be other risks with 100% certainty
• Step 2 specify the Risks that are less than 100% likely to occur
In the schedule example above, the less than 100% Risk could be
• Regulatory approvals
• Long Lead Time Procurement Supplier Delays
• Design Productivity
• EME HR shortages
• Construction Logistics
• Government Infrastructure Delivery
Mechanics of the Risk Driver Approach Step 3: Assign risks to activities
Regulatory LLT SuppliersDesign
Productivity
EME HR
shortages
Construction
Logisitics
Government
Infrastructure
Delivery
Approvals
Design phase
LLT Procurement
Construction
Commissioning
Activities
Risks Assigned
Regulatory LLT SuppliersDesign
Productivity
EME HR
shortages
Construction
Logisitics
Government
Infrastructure
Delivery
Approvals
Design phase
LLT Procurement
Construction
Commissioning
Activities
Risks Assigned
Regulatory LLT SuppliersDesign
Productivity
EME HR
shortages
Construction
Logisitics
Government
Infrastructure
Delivery
Approvals
Design phase
LLT Procurement
Construction
Commissioning
Activities
Risks Assigned
Regulatory LLT SuppliersDesign
Productivity
EME HR
shortages
Construction
Logisitics
Government
Infrastructure
Delivery
Approvals
Design phase
LLT Procurement
Construction
Commissioning
Activities
Risks Assigned
0.95 1.05 1.0
P=100% Background risk, estimating accuracy
Regulatory LLT SuppliersDesign
Productivity
EME HR
shortages
Construction
Logisitics
Government
Infrastructure
Delivery
Approvals
Design phase
LLT Procurement
Construction
Commissioning
Activities
Risks Assigned
0.9 1.2 1.0
0.95 1.05 1.0
P=30%
P=100%
1.0 1.3 1.1
0.8 1.5 1.0
1.2 1.75 1.5
1.0 1.3 1.1
1.2 1.75 1.5
P=15%
P=35%
P=45% P=25%
P=40%
Background risk, estimating accuracy
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
0 X <=31/05/2012
95%
0
0.2
0.4
0.6
0.8
1
22/03/12 09/04/12 27/04/12 15/05/12 02/06/12 21/06/12
Distribution for Complete/Finish
Prioritising Schedule Risks using Risk Driver Approach Deterministic Date: 19-Jan-12
P80 Impact on P80 Date
Background plus All Risks 21-Jun-12 Days % of Total
Risks removed in Prioirty order
LLT Procurement 14-May-12 38 25%
Approvals 10-Apr-12 34 22%
Construction 13-Mar-12 28 18%
Design Phase 22-Feb-12 20 13%
Construction 6-Feb-12 16 10%
Commissioning 25-Jan-12 12 8%
Background Risk 19-Jan-12 6 4%
Total 154 100%
Prioritising Schedule Risks using Risk Driver Approach
Wrap up
• Very quick introduction to Risk Driver Approach to Schedule Risk Analysis
• Hopefully demonstrates the benefits of applying a Risk Driver Approach to Schedule Risk Analysis
• If it is to be adopted needs to flow through to how Schedule Risk Workshops are conducted to focus on key risk drivers and follow up risk interviews
• TBH Capital Advisers is developing a 2 day, hands on workshop on the approach for anyone interested in more detail