Houston, Texas (PRWEB) May 18, 2013
Traditional project schedules are based on single-point estimates of task durations, i.e. on the assumption that task durations are known precisely in advance. Real life is not like that, with the result that deterministic project schedules are bound to be wrong.
The solution is project risk analysis, aided by software such as “Full Monte.” Taking into account the inevitable uncertainty inherent in any project is the path to more realistic and achievable project plans.
“The development of project simulation software is essential to the analysis of project risk,” says Tony Welsh, CEO of Barbecana and developer of the “Full Monte” software that adds Monte Carlo simulation to Microsoft Project. “By using project risk analysis software, a project manager will better understand the risks involved in projects and the impact they have.”
What’s the Chance of Finishing on Time?
The deterministic estimate of the project finish date is unreliable, because it’s based on single-point estimates of durations some of which are almost bound to be wrong. Even if the expected finish date right is correct, this still means there is roughly a 50% chance of finishing on time. “If you really need to finish the project by a particular date,” comments Welsh, “it’s not prudent to use a plan that gives you only a 50% chance of doing so. What you really need is a plan which gives you maybe a 90% or 95% chance of finishing on time, depending upon how critical it is. Many use 80% as a rule of thumb, sometimes called the “P80” point.”
But for most project schedules the chance of finishing by the deterministically determined end date is not even 50% because of a pernicious built-in optimistic bias called “merge bias.” Whenever parallel paths merge the net result is to add an extra delay because we have to wait for the longer path to complete. Some may take less time than expected, while others take longer. But there is no benefit from the shorter path, and the project gets a hit from the late one. A gain on the swings is lost on the roundabouts.
This means that quite often the deterministic date derived from a traditional critical path analysis is not only uncertain, it’s highly biased. Sometimes the probability of finishing the project by that date is not even close to 50%. The bias builds up at successive merge points as the project progresses, so while we may meet some milestone dates early on it becomes harder and harder to do so.
These two reasons mean that many projects don’t finish on the projected time schedule. “People may think this means they’re not executing the project properly,” claims Welsh. “However, the reason for not finishing on time may be because what they mean by ‘on time’ is ‘according to the schedule.’ If schedule itself was never realistic, it’s the plan that’s wrong, not the execution. People tend to have expectations that can’t be met.”
But why Monte Carlo?
Attempts at analytical methods like PERT and the Method of Moments simply cannot handle the complexity of a project network and are therefore forced to make unjustified assumptions in order to make them tractable. PERT in particular completely ignores merge bias. Monte Carlo simulation is the only known way to accurately analyze a project network subject to uncertainty.
How Monte Carlo Simulation Works
Project risk analysis using Monte Carlo simulation takes all these factors into account. There is a probability distribution for each of the durations in the network, samples are drawn from these, and calculations are performed as though they were the actual durations.
After performing this exercise thousands of times with different samples, a picture is accumulated of the relative probabilities of the possible outcomes, and in particular of the project finish date as well as the dates of important milestones and costs.
“You end up with an analysis that determines a date by which you have say a 90% chance of finishing,” says Welsh. If that isn’t acceptable, then the schedule can be adjusted either by increasing resources or changing the order of things — whatever makes it acceptable. “The point is that by performing the Monte Carlo simulation, you’ll know that you’re unlikely to finish by the day you need using the original plan. Often the schedule was never realistically achievable in the first place.”
Full Monte, with its extended list of features, uses Monte Carlo simulation, allowing the user to better understand the risks involved in every project and to adjust his plan accordingly. Full Monte is risk analysis software for project management that integrates seamlessly with Microsoft Project 2007 and above.
It’s a requirement for most government projects, and especially those for the Department of Defense, to perform risk analysis. “That’s why most of our clients have been in the aerospace and defense industry,” states Welsh, “but risk analysis software should be used by everyone, regardless of the type of project they’re doing. And with the Full Monte 30-day free trial download offer, why wouldn’t they?”
For more information about Full Monte and Barbecana, visit us online at http://www.Barbecana.com or call today at (888) 706-8789.