"Here's a good question. How long does it take to catch a fish? Ridiculous, you say. One can't estimate the time to catch a fish. It could be just after you cast a line in the water. It might be never. Or anywhere in between. As ridiculous as this sounds, that is just the feeling that goes through our minds when we are asked to estimate the duration for a task. Our first thought is How the h*** should I know? But we can't get away with this. So we dig in and take a scientific stab at the task duration.

First, we come up with a most likely estimate of the duration. This is the time that we feel it would take about 50 percent of the times that we were to execute the task. But we're not comfortable with a 50 percent confidence factor. So we add some time that we feel we could support about 90 percent of the time.

Next, we think about what we will need to start the task, including what kinds of conditions are required. If we are concerned that we will not have everything that we need to start the task, we add some more time to the task estimate (even though these issues do not impact upon the actual time to execute the task itself).

Then there is the collection factor. When a group of tasks come together, we tend to add some more safety margin, to allow for one of the tasks to slip. Similarly, we note that there is a tendency to lose time between tasks. I call this the 5 + 5 = 13 rule. Two tasks, each estimated at 5 days, performed in series, will take 13 days because we lose 3 days between the completion of the first task and the start of the second task.

So what do we do? We compensate for all these factors that are external to the immediate task, by adding time to the task estimate, itself.

Finally, everyone knows that the total duration will not be accepted. They expect to be pushed for a 20 percent reduction, so they add 25 percent to the already inflated estimate."[2]

But Harvey is not yet finished. Estimating the task duration is only the start. What happens when it comes to doing the task? How long does it take? On this part, Harvey has this to say ...

The Psychology of Task Durations

"There is a self-fulfilling prophecy regarding performance of tasks within planned durations. A task is hardly ever completed ahead of schedule. There are several reasons for this. We can demonstrate these using an illustration of a task that has a 50-50 chance of being completed in 5 days, but has been scheduled for 10 days to allow for uncertainty, risk, emergency diversions, and so on.

First, there is Parkinson's Law. Work expands to fill the time available for the work. Work on the task has commenced on schedule and is essentially completed within the first 5 days. But, because 10 days have been allocated for the task, the performer spends the next 5 days fine-tuning the deliverable. This is a natural work ethic of most people. We reach 98 percent completion on our task and, if additional time is available, we attempt to refine it until a delivery deadline is reached.

Second, is procrastination. We are able to start the task as scheduled. But, because there are 10 days allocated, and we know that we need only 5 days, we wait a week to start the task. Now, of course, the contingency has been exhausted before the task has been started, and the potential for a schedule overrun has been increased. But, even if there are no problems, the 5-day task has taken 10 days.

Less obvious are the subtle motivators to avoid early completion of tasks. If we estimated 10 days and complete the task in 5 days, we might be criticized for padding the estimate, even though the extra 5 days was a legitimate allowance for uncertainty. Or we might be under increased pressure to shorten duration estimates in the future. There rarely is a reward for finishing tasks early - only demerits for running over. So where is the motivation to do the task in 5 days?"[3]

So, Harvey offers this tip to project managers:

"The time to complete a task will almost always take a minimum of the allocated time, and probably more. If pressure is to be maintained to minimize the time spent on tasks, it is advantageous to move most contingency out of the individual tasks and allow for contingency in other ways."[4]

Amen to that! And there is a whole heap more good advice in this fascinating book of Harvey's practical experience.