Do you feel confident Building an integrated Monte
Do you feel confident? Building an integrated Monte Carlo risk register into your project plan Project Challenge - September 2011 Presented by Ian Wallace
Typical project risk register
Palisade Corporation » Decision-making software • Decision. Tools Suite (Excel add-ins) • – @RISK – Top Rank – Precision. Tree – Optimizer, Neural. Tools, Evolver, Stat. Tools – Developer’s Kit @RISK for Project » Established 1984, HO in Ithaca, New York state. » European HQ in London » Over 150, 000 users worldwide » Taught at top 60 MBA programs globally » Used by a huge variety of companies » www. palisade. com
Introductions - Ian Wallace » Accountant in Industry - ACMA » KPMG Management Consulting • 15 years in Project Management » Palisade - specialise in Decision Analysis using the Decision. Tools Suite: • @RISK for Excel • @RISK for Project • Top Rank • Precision. Tree • RISKOptimizer, Neural. Tools, etc
Agenda » Introduction » What is Monte Carlo risk analysis? » Why bother? » Example Monte Carlo risk register model using @RISK » Conclusion
History
Probability distributions Example - The Triangular distribution - spreads the probability across a simple 3 point estimate - quickly captures the 90% confidence range (or lack thereof!)
Monte Carlo sampling
Monte Carlo distribution Expected Value/Mean ML Min 0 10 Max 20 30 40 50 60 70 90% chance ‘Stresses’ the model with probability
Probabilistic sensitivity analysis Key risk drivers Systems design acceptance Commissioning Data migration Test acceptance Installation Development Task 1 Task 2 Helps prioritise the mitigation effort
Agenda » Introduction » What is Monte Carlo risk analysis? » Why bother? » Example Monte Carlo risk register model using @RISK » Conclusion
Decisions, decisions…… What is the margin of error? Have we got enough time? What are the ‘odds’? Is this a ‘quality’ plan? Have we got enough budget? What can we do to reduce uncertainty? What price/date shall we quote? Is this the most risk efficient approach? What is the residual risk – do we need to off-lay more risk? ?
Typical risk approaches » Risk registers » » 3 -point estimates – minimum, most likely, worse cases Scoring methods What-ifs Just plain “gut feel” Not focused enough for difficult decisions
They also lead to a single number » “The Number, ” once written, becomes set in stone » “The Number” is disseminated » “The Number’s” underlying assumptions – and errors - are forgotten » “The Number” becomes the basis for big decisions » Most likely case in 3 -point estimate = “The Number” » Scoring methods produce a final single “score” = “The Number” » Running lots of subjective What-ifs forces managers to pick “most likely” point assumptions to get “The Number” » Gut feel simply guesses “The Number” (well, we all do it!) The 90% confidence range is more useful
Monte Carlo measures the confidence level Expected Value/Mean ML Min 0 10 Max 20 30 40 90% confidence range Is this acceptable? 50 60 70
Probabilistic sensitivity analysis Key risk drivers Systems design acceptance Commissioning Data migration Test acceptance Installation Development Task 1 Task 2 Prioritises mitigation based on probable effect on the target outturn
Compare confidence levels
So - why bother? » Monte Carlo analysis provides the ‘ammunition’ for making difficult decisions, e. g. • changing the plan • supplier selection • contingencies – time and money • go or no- go (Stopping!) » Other benefits: • corporate benefits – improved risk management = competitive edge – helps avoid over-commitments to customers – good corporate governance practice - improved transparency • personal benefits – looks professional – shows deep understanding of the problem
Agenda » Introduction » What is Monte Carlo risk analysis? » Why bother? » Example Monte Carlo risk register model using @RISK - Building an integrated Monte Carlo risk » register into your project plan » Conclusion
Cost risk analysis using @RISK for Excel Recommended cost contingency after probability
Schedule risk analysis - @RISK for Project
And soon, @RISK v 6 » Imports a MS Project plan into Excel for Monte Carlo simulation » Provides a dynamic link between MS Project and MS Excel • means that all the Monte Carlo analysis is in the same place • provides access to Palisade’s Decision. Tools Suite • includes new risk perspectives – e. g. probabilistic project cashflows » See demonstration
Agenda » Introduction » What is Monte Carlo risk analysis? (probabilistic sampling) » Why bother? (‘ammunition’ for difficult decisions) » Example Monte Carlo risk register model using @RISK » Conclusion
Conclusion » Confidence and trust are vital in project management: • nobody wants poor quality plans and estimates • adding a probabilistic risk register to a cost estimate or project plan: – measures the confidence level and the key risk drivers – provides the ‘ammunition’ to make difficult decisions and build consensus on the way forward » Not difficult: • just add probability distributions to existing spreadsheets and project plans • works in familiar technology
Remember - ask yourself this question “I know what you’re thinking: 'Did he fire six shots, or only five? ' Well, to tell you the truth, in all this excitement, I’ve kinda lost track myself. But being this is a. 44 Magnum, the most powerful handgun in the world, and would blow your head clean off, you’ve got to ask yourself one question: 'Do you feel lucky? '
Then, ask yourself again Well do ya, punk? ”
Questions? Good ‘Luck’! iwallace@palisade. com
- Slides: 27