Paper: Stop Predicting, Start Forecasting
Abstract: Project managers and weather forecasters both make projections about uncertain, future events. Weather forecasters primarily focus their projections on temperature, precipitation, air pressure, and wind. Unlike weather forecasters, project managers primarily make predictions (not forecasts), about project duration and cost, and the likelihood of risk occurrence and their associated impacts.
To create their project predictions, most project managers use deterministic (single-value) estimates about project uncertainties, like schedule and budget. Usually these estimates are the most likely outcome, but that ignores many other possible outcomes, both probable and improbable.
Weather forecasters know, however, that predicting the weather won’t help people make informed decisions involving the weather. Instead of predicting, weather forecasters make stochastic (probabilistic) estimates that account for many outcomes, both probable and improbable. They know that deterministic estimates neither align stakeholder expectations, nor do they enable informed decision-making by their many stakeholders. They also know that predictions do not convey any sense of risk to decision-makers, and decision-makers cannot make informed decisions on the strength of a single, possible outcome of the predicted future.
In this presentation, we will examine the limited usefulness and hazards of making deterministic project estimates, and contrast that with making stochastic forecasts.
Stochastic estimation has two chief advantages over deterministic estimation. Firstly, stochastic estimation better aligns stakeholder expectations. Stakeholders know the specific probability of every project estimate. Secondly, stochastic estimation fosters better, more informed, decision-making. Using forecasts, decision-makers evaluate risk/reward trade-offs by evaluating both probable and improbable outcomes (which can be favorable or unfavorable).
Biography: William W. Davis, MSPM, PMP works as a senior project manager at NCCI Holdings, Inc., in Boca Raton; he also is an author of new courseware for Pluralsight, the world’s largest, curated, e-learning library for technology and creative professionals. William enjoys sharing innovative solutions with managers and leaders in the project management industry. 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 PMP credential. He holds graduate degrees from Nova Southeastern University (M.S. in Leadership) and The George Washington University in Washington, DC (M.S. in Project Management). William created Statistical PERT, an Excel-based, easy-to-use, probabilistic technique for estimating project uncertainties like task duration, work effort, or risk impact.
