013-Don’t guess! Plan your project with SCIENCE using Monte Carlo analysis

013-Don’t guess! Plan your project with SCIENCE using Monte Carlo analysis

Although touched on in the PMBOK, most of us have little exposure to Monte Carlo analysis – which is a shame! Monte Carlo analysis can add a new level of credibility to our planning when the stakes are high by backing it with real science and statistical models. In this episode, we demystify this tool, discuss its benefits, when it is appropriate, and why math is our friend J From mining to professional services, contact centers and orbital battle stations, we discuss how Monte Carlo can help you run your business.

We also discuss photon versus proton torpedoes.

Our special guest is Luc Vandamme, who has been in the mining industry for more than 25 years and is now a Senior Director in his organization’s Enterprise Project Management Office.  Luc’s career has taken him from mining technology research and development to business optimization, organizational development, and now to project management.  Luc has a Civil and Geological Engineering degree from the University of Louvain in Belgium (he is NOT related to Jean-Claude!) and a Ph D in rock mechanics from the university of Toronto. And of course he knows Monte Carlo simulations inside and out!

GIVE MONTE CARLO A TRY!

Here are links to a couple Excel based tools that our host mentioned. At least one has a free trial. They aren’t cheap – but then neither is project failure!

Oracle’s Crystal Ball:

https://www.oracle.com/applications/crystalball/index.html

At-Risk

http://www.palisade.com/risk/

JOIN TODAY, AND GET IMMEDIATE ACCESS TO TAKE THE QUIZ!

Members get unlimited access to take the quizzes for our podcasts. Not only, that, but you also get a certificate, we’ll upload your PDU’s to PMI, and you get lots of other great benefits – click here for details on becoming a member.  

JOIN US FOR HAPPY HOUR!

Come by and say “hey!” in our PM Happy Hour FB Page at https://www.facebook.com/pmhappyhour/

Show Notes

Introduction – Get To Know Luc Vandamme [3:52]

Luc Vandamme has been in the mining industry for more than 25 years. He started his career with Noranda and Falconbridge, and is now a Senior Director in Barrick’s Enterprise Project Management Office.

Luc has a Civil and Geological Engineering degree from the University of Louvain in Belgium and a PhD in rock mechanics from the University of Toronto. His career has taken him from mining technological research to business optimization to organizational development, and now to project management.

Luc sharing his journey to Project Management [5:15]

Luc was hired at Barrick back in 2010 to optimize the company’s capital allocation process. This is a process by which funding is requested and approved (or not) for capital projects.

After implementing a stage-gate process for project approval, he quickly realized that it’s one thing to approve projects, while it is quite another to execute them successfully. This led to the idea of creating a PMO at the enterprise level.

Luc’s favorite thing about being in the field? [6:18]

Luc is a very visual person – he love to see the results of his work. Through his work, he has been privileged to go to places which not too many people have the chance to visit. To him, mining projects tend to be quite spectacular, and he is literally down to earth. 

Luc on the most valuable business lesson he learned the hard way [7:05]

When something takes 1 month to complete, don’t give yourself or anyone else 2 months- if you give 2 months, it will take 2 months!

You have to keep schedules tight, but don’t forget to include an empty task at the end of your schedule; this is for you to absorb unforeseen delays that might come up.

Luc discussing Monte Carlo Simulation [7:46]

Monte Carlo Simulation is a tool that has existed for a long time, but is just now slowly being adopted by industries such as project management.

What is the Monte Carlo Simulation? [8:53]

Monte Carlo simulation is a relatively simple technique whereby you quantify uncertainty, variability, and risk. A very simple example: if a factory produces 5 widgets of type A and 7 widgets of type B in a day, how many widgets does it produce in total?  The answer is obviously 12. In reality, however, the answer may range from 6 (or even zero) on those days where nothing seems to go right, then maybe 15 when all the stars are aligned! The calculated number 12 is just an average. But there is a lot of uncertainty or variability or risk associated with that average.  So how do we quantify this to the point where we can start answering questions such as: What is the likelihood I will produce at least 10 widgets on any given day? How many times can I expect not to meet my quota of 12? Those are the quantitative questions that can be answered using Monte-Carlo.

Where does the name “Monte-Carlo” come from? [10:34]

Every iteration that is required in this method makes use of random numbers.  This reminded early users of games of luck. Hence, they decided to name this method after the most famous casino in the world – The Monte-Carlo.

What are some benefits of using a Monte Carlo simulation? [11:57]

    • An appreciation of the amount of variability that is present in whatever you do.  This is applicable whenever you try to capture a real phenomenon in a model. This means it’s useful in fields as diverse as medicine, meteorology, investment banking, engineering, and, of course, project management.  
    • One example from the project management world is project estimates. Be it of cost or time durations, project estimates are never exactly right.  When a project manager tells that his project is going to cost $2,537,443 and it will take 143 days, the one certainty is that he will be wrong.

It would be an incredible stroke of luck for the project to unfold exactly as predicted.  Instead, the project managers should give ranges: for example, it is going to cost anywhere between 2 and 3 million dollars and it will take between 4 and 5 months- this, at least, has a greater chance of being right. 

This raises a crucial question from given ranges: What is the range based on? Can you check whether the assumptions make sense or not? How sure are you? And how can you check whether they’re 50% or 95% sure?

What are the typical uses of Monte Carlo in project management? [13:54]

1. Calculation of contingencies  

The key question here is how much the project budget should carry as a contingency. Contingency is basically a fund set aside to cover minor mistakes or omissions in the cost estimate.  

How much should the contingency be? Contingency is calculated by making assumptions about the ranges around every major cost element that make up the estimate. This includes everything from the cost of individual pieces of equipment, to labor, to materials.

Given these assumptions, the Monte-Carlo software will calculate a range of the total project cost.  Once you have the complete distribution of what the total project cost might come to, you can choose a project budget that makes sense. 

2. Quantifying risk 

This means event-related risks- if there are good assumptions on the likelihood and the potential cost of such events, the Monte Carlo can calculate a range of risk-related cost outcomes.

3. Application to the project economic model

The Monte Carlo process can be used to calculate potential ranges on performance metrics (NPV, IRR, Payback periods). In this case, Monte-Carlo software can be used to find the ranges on NPV or IRR.

4. Identifying Economic Factors

It will also help you find out what the main economic drivers are for any particular project. In other words, out of all the potential sources of variability in a project, which are the ones that affect the project’s economics most?

Are there times when Monte Carlo is not appropriate or not worth the effort? [20:50]

Projects that do not carry any risk would not benefit much from a Monte Carlo analysis.  Also, applying Monte-Carlo to very small or simple projects may not justify the cost of preparing for and running the simulation.

What planning or preparation is required to make sure we are getting valuable output? [21:36]

  • The principle of garbage-in, garbage-out still holds.  The key to ensuring valuable results is to make sure that the assumptions are valid.  As much as possible, the ranges on the input variables need to be based on data, not just wild guesses.
  • There are companies out there that calibrate estimators by making them aware of their degree of conservatism. They do so by asking you a bunch of “trivia questions”, but you must answer with a range.
  • Actual distributions for the input are not as important.  The software typically allows you to pick from many distributions, such as normal, beta, gamma, Weibull, lognormal, etc.  In most cases, however, the only thing that really matters is whether you are picking a discrete or a continuous distribution.

Sample tools for Monte Carlo Simulation [38:57]

There are two software products that work as Excel add-ons: Crystal Ball and At-Risk. The Risk module of Primavera, for instance, includes an integrated Monte-Carlo simulator for both cost and schedule.

Where can listeners go to find out more about Monte Carlo simulation? [39:46]

The literature on Monte-Carlo is exploding, and so are the self-help videos online. Just type “Monte-Carlo simulation” on a Google search and you’ll find thousands of entries.

You might be in trouble if on Monte Carlo Simulation… [42:21]

  • You ignore variables which can significantly impact your project or business case (currency exchange rates, etc.)
  • You do not question the assumptions that go into the simulation- remember, garbage-in, garbage-out. The first thing to do when you review someone else’s simulation is to ask what the ranges are on the input variables.
  • You overdo it and put too many ranges on too many variables – this tends to introduce a lot of unnecessary noise and adds to the run time.
  • Finally, not inputing the correlation of variables. This includes cases where when one variable tends to be high, the other one is also tends to be high, and vice versa.  Forgetting to include those correlations may lead to skewed results.

Home Forums 013-Don’t guess! Plan your project with SCIENCE using Monte Carlo analysis

Viewing 1 post (of 1 total)
  • Author
    Posts
  • #4670
    Kim Essendrup
    Keymaster

    I'm sorry, you are not authorized to view this page. Why not sign up for a membership so you can get access to this and even more content!

Viewing 1 post (of 1 total)
  • You must be logged in to reply to this topic.