PERT Variance

PERT Variance measures how much uncertainty, or "spread," there is around your estimate for how long a single task will take. PERT stands for Program Evaluation and Review Technique, a method that uses three guesses for each task: the Optimistic time (O, the fastest it could go if everything goes right), the Pessimistic time (P, the slowest it could go if things go wrong), and the Most Likely time. Variance is a number that tells you how confident you can be — a small variance means the task is fairly predictable, while a large variance means the outcome is very uncertain.

To calculate it, you take the Pessimistic time, subtract the Optimistic time, divide that difference by 6, and then square the result (multiply it by itself). The "divide by 6" comes from statistics, and squaring turns the standard deviation (a measure of spread) into variance. In fact, variance is simply the square of the standard deviation.

How do you read the result? A bigger variance number means the task carries more risk and is harder to predict; a smaller number means you can plan around it with more confidence. The real power comes when you add up the variances of every task along a path through your project — variances add together nicely, which lets you understand the total uncertainty of an entire chain of work and set realistic buffers.