Statistical PERT® is a freely licensed, probabilistic estimation technique for use with Microsoft Excel® (2010 / 2013 / 2016 / 2019 / Microsoft 365).
Nearly every professional project manager is familiar with the Program Evaluation and Review Technique (PERT) to create a risk-adjusted estimate using a minimum, maximum and most likely outcome for any bell-shaped uncertainty. The PERT formula to estimate the mean (that is, the average or expected result) is:
(Minimum + 4(Most Likely) + Maximum) / 6
Credentialed project managers from the Project Management Institute memorize this formula in preparation for the Project Management Professional (PMP®) examination.
Statistical Program Evaluation and Review Technique (Statistical PERT® or, SPERT®) starts with this PERT formula to estimate the mean for a normal, bell-shaped probability distribution. Then, Statistical PERT uses the estimator’s subjective judgment about how likely the most likely outcome really is to adjust the standard deviation for the normal (or beta) curve. By doing this, Statistical PERT will create greater certainty around the mean when an estimator indicates high confidence in the most likely outcome. Conversely, Statistical PERT will create greater uncertainty away from the mean when an estimator indicates low confidence in the most likely outcome (this situation occurs when, for example, an estimator has little knowledge about a particular uncertainty, so the most likely outcome is more of a best-guess about that uncertainty).
Statistical PERT uses the built-in statistical functions of Microsoft Excel. There are no Excel add-in programs to install, no macros to worry about, nothing to buy, and nothing to register. The free, downloadable Statistical PERT templates and example workbooks can be modified and redistributed according to the GNU General Public License by the Free Software Foundation.
Statistical PERT uses Excel functions that come with Microsoft Excel (2010 / 2013 / 2016 / 2019 / Microsoft 365). There are now three editions of Statistical PERT: Normal Edition, Beta Edition, and the Bootstrap Edition (the Lognormal Edition is coming soon). The Normal Edition uses the STDEV.P, NORM.DIST and NORM.INV functions. The Beta Edition uses BETA.DIST and BETA.INV (along with the NORM.DIST and NORM.INV functions). The Bootstrap Edition uses PERCENTRANK, PERCENTILE, along with several other advanced Excel functions. Older versions of Excel are compatible with Statistical PERT concepts, but not the free, downloadable Excel templates.
Statistical PERT® Normal Edition uses ratio scale modifiers to modify the standard deviation of a normal curve. This lets Statistical PERT work with any bell-shaped uncertainty that can be estimated using a 3-point estimate for minimum, maximum and most likely outcomes.
Statistical PERT® Beta Edition uses ratio scaling to determine the mean and standard deviation, but it also performs analysis on the implied shape of the bell-shaped curve to select the two shape parameters for use with the beta distribution functions of Excel.
Statistical PERT® Bootstrap Edition uses statistical bootstrapping, which is a simulation technique, to create a probabilistic, delivery date forecast for agile teams using Scrum or Kanban.
All three editions of Statistical PERT are extremely easy to use. For instance, for the Normal Edition, simply create a 3-point estimate for any bell-shaped uncertainty, then render a subjective judgment about how likely the most likely outcome really is. The Normal Edition template comes pre-calibrated so stating Medium Confidence in the most likely outcome will result in probabilities that are similar to a Monte Carlo simulation using the RiskPERT function in Palisade’s @Risk Excel add-in program. (Click here to see samples of this comparison). Choosing other opinions will increase or decrease the calculated standard deviation for an uncertainty, and the probabilities will re-calculate accordingly.