It’s natural for managers to want to know when a project will end.

It’s also an exercise in fiction.

Here’s why.

Every estimate is just that: an estimate.

The time it takes to perform a task equals the estimate plus or minus some variability. One measure of variability is the variance. Never mind what the variance means exactly. Just remember that it measures the uncertainty about how long it’ll take to complete a task.

Imagine a project comprised of three tasks: A, B, and C.

Say the estimated duration of each task is 5 days and the variance of each duration is 1 day. These are realistic numbers.

It’s natural for a manager to add the durations of the tasks and conclude that the project will end in 5 + 5 + 5 = 15 days.

But Statistics says that the variance of the sum equals the sum of the variances.

Huh?

That means that the variance of the project duration is 1 + 1 + 1 = 3 days. So the uncertainty about the duration of the project is 3 times the uncertainty of each task.

That’s for 3 tasks. Now imagine projects with 15 tasks or more.

The uncertainty of the duration of a project is so great that the schedule becomes a work of fiction.

So why do some projects end on time?

If a task is late, and the pressure to end on time is strong enough, the person who’s performing the task will cut corners. Quality will suffer.

If a task is early, the person performing the task will slow down to make the duration fit the deadline. That way she can have some respite from the daily grind.

No judgment. That’s just human nature.