I love things that make my job simpler. Although a statistical model for estimating is not something I would normally put in the “make my job easier” box, I may have just found one that works.
William W. Davis MSPM, PMP, and Project Management SuperheroI met William W. Davis MSPM, PMP, and Project Management Superhero. He has taken the PERT (Project Evaluation and Review Technique), estimating approach to the next level, allowing you to add a bit of professional judgement with the numbers.
I began by asking him why he felt this was necessary.
William, why isn’t normal PERT enough?
The PERT formula calculates an expected value for uncertainty with bell-shaped properties. However, expected values are only 50% reliable. What if you need an estimate that is, for example, 75% reliable? Or even 90% reliable? These types of estimates can’t be provided by PERT.
PERT cannot rationally adjust estimates to include an estimator’s intuition and knowledge about uncertainty. PERT is all brain, but no heart. However, decision-making requires both intellect and emotions. This is true when we make estimates about project uncertainty.
OK, I get it. What does the S have to offer?
All project managers have the power of the S in SPERT.
That sounds great!
Statistics is the first superpower. According to the American Statistics Association, statistics can be described as “the science of learning using data and of measuring and controlling uncertainty.” Statistics is my way of learning, measuring and controlling uncertainty and communicating it to my project stakeholders.
Statistics is a scientific superpower. Project managers who can harness this power can address the two main reasons that anyone estimates anything.
Please remind me of them again.
We estimate to:
Harmonize expectations across many stakeholders so that everyone knows what to expect from future uncertainties.
Take better and more informed decisions regarding future uncertainties.
Ah, yes. Did you mention more than one superpower?
Sensing is the second superpower. Statistical PERT allows estimators to rationally adjust their estimates based on their sense of the most likely outcome.
SPERT uses an estimator’s perception of the most likely outcome to adjust SPERT estimations.
Can you please explain it?
Statistical PERT is a five step process. However, a SPERT template only requires three steps. These are the five steps:
For some uncertainty, identify the minimum, most likely, and maximum outcome
Use the PERT formula to calculate the expected value
Make a subjective judgement about the most likely outcome.
Calculate a standard deviation
Select any probabilistic estimate that best suits your risk level
SPERT templates take care of steps 2 and 4, and Excel’s statistical functions make it easy.
You have a SPERT Template that you are offering for free. What’s the deal?
All Statistical PERT template workbooks and examples are available for free to anyone.
Visit William’s website to download a free SPERT Template. I want to remove any barriers that prevent people from discovering their statistical superpowers.
I encourage all businesspeople, especially project managers, to use statistics to quickly align stakeholder expectations to improve executive decision-making.
What was it that inspired you to create it?
Two years ago, I interviewed peer project managers and asked them the following question: “How confident are you when estimating your projects?”
Their anonymous responses varied from 50% to 100%. However, none of these project managers had calculated their confidence levels and their sponsors didn’t know how much risk they were taking by approving their schedules and budgets.
I realized that project managers need a way to communicate their confidence and risk to others. I couldn’t find a good way to do this so I invented one.
It’s great, we all now benefit