Project managers forecast the future all the time: we forward-schedule tasks, plan how much money we’ll need in Quarter 4 and book resources for activities that haven’t happened yet.
But how much do you really know about the business of predicting the future? I caught up with forecasting superhero William M. Davis for a quick primer in forward planning.
William, what’s the difference between predicting and forecasting the future?
Think about weather forecasting. Your favorite meteorologist (or your favorite smartphone weather app) makes both weather predictions and forecasts.
A prediction is a single-value, deterministic estimate of a future uncertainty. A forecast is a probabilistic estimate about the possibilities, both probable and improbable, of a future uncertainty.
I don’t get it!
Weather forecasters predict high and low temperatures for the day. For example, in South Florida tomorrow, the daytime high temperature prediction is 90 degrees Fahrenheit.
I know that it might not be exactly 90 degrees tomorrow, but I have an intuitive tolerance for error for predicted temperatures. Tomorrow’s temperature might rise to only 88 degrees, or it might rise to 92 degrees. Either way, it’s going to be hot!
But when it comes to estimating the chance for rain tomorrow, meteorologists forecast the chance of rain.
Tomorrow’s rain forecast is 50% where I live. Rain tomorrow is both possible and somewhat probable.
Got it! Which one is better?
Predictions and forecasts are both useful!
When we need quick estimates about the future, predictions are easier to create and share. But if there isn’t a common understanding of how much uncertainty surrounds our predictions, then offering predictions can be hazardous.
To convey our sense of uncertainty, we should forecast, not predict.
That’s OK for the weather, but how does this have anything to do with work? How does forecasting allow me to better manage stakeholder expectations?
Have you ever offered a project prediction for cost or schedule early in a project’s lifecycle, but then the sponsor regards those early predictions as firm commitments?
When you need to convey a sense of uncertainty about your estimate, it’s much better to create a forecast. A forecast often uses a bell-shaped probability curve to convey the probabilities of the many possibilities of an uncertain future.
When you forecast, you can make statements like these:
- “With 80% confidence, we can complete that project for $300,000 or less.”
- “If we want to be 95% confident, we’ll need a budget of $340,000.”
- “Yes, we can cut the budget by $30,000. But then I’m only 60% confident that the project won’t exceed its budget—are you willing to accept that risk?”
What about helping my sponsor or organization make better decisions?
Suppose you’re my project sponsor, and you ask me for an early, high-level cost estimate for the project we’re about to start. And suppose you’re willing to invest up to $1M in this project.
If I offered you a project cost prediction of $900,000, which is my “most likely” outcome for your project, would you fund the project?
You probably would. My cost prediction is less than your spending threshold. But my cost prediction doesn’t convey any sense of uncertainty that I have about the project’s “most likely” outcome.
Now let’s suppose I created a project cost forecast. It might look like this:
You see right away the cost uncertainty that exists for your project. Even though the “most likely” outcome is $900,000, there is a risk this project will exceed your $1M funding threshold.
But how much risk is there? Let’s look at the same forecast, but in another way:
Yikes! My forecast shows that there’s a 29% chance (the red pie slice) that the project will cost more than your funding threshold of $1M! And there is a 73% chance that the “most likely” outcome of $900,000 will be exceeded (the blue and red pie slices together).
Maybe this project doesn’t look quite so good!
Forecasting gave you, the project sponsor, important information about the project proposal. You didn’t learn the riskiness of this project when I shared a predicted cost estimate of $900,000.
This all sounds pretty complicated. How do I make forecasting easier?
Forecasting can be easy! Or hard. It depends on the forecasting model you’re using.
To make forecasting easy, I created a freely-licensed template for Microsoft Excel called Statistical PERT. Statistical PERT creates simple forecasts by using 3-point estimates (minimum, most likely, maximum) and the estimator’s subjective opinion about how likely the most likely outcome really is.
Statistical PERT uses Excel’s built-in, statistical functions to create the forecasts you saw in this article.
OK, I’m sold. How do I get started?
Try creating your own project forecasts! You can create forecasts for virtually any project uncertainty involving numbers, and at any time during the project lifecycle. The Statistical PERT website has free whitepapers that explain how to forecast.
Oh, and one more thing.
Don’t forget your umbrella when you leave for work tomorrow!
Your next steps
So what can you do with this information? Here are some next steps to consider for managing your project.
- Forecasts and predictions normally require an understanding of what you think is going to happen, so read this article on project assumptions so you have the full picture.
- Take a look at these 14 or so common project risks that might affect your project, so you can factor them into your planning.
- Book a meeting with the team to discuss how you are going to do project estimating and come up with these forecasts.
- Talk to you sponsor about their understanding of risk levels with the numbers you have presented to them.
- If you’re starting a new project, look at my detailed article on how to create a project budget for some tips on putting the numbers together.
About my interviewee: William W. Davis, MSPM,
He has 30 years’ experience working as a software developer, Oracle ERP implementer, and IT project manager. William has been a PMI member since 2005 when he earned his
William created Statistical PERT, an Excel-based, easy-to-use, probabilistic technique for estimating project uncertainties like task duration, work effort,