Wednesday, February 3, 2010

But We Still Get to Work.........

I recently did a speech in Austin, TX.  One of the individuals attending asked a great question and I have thought about it ever since.

For those of you that have seen me, I often use an example of variance in estimates by asking someone how long it takes them to get to work.  Usually, it is between 300-500% variance.  I then explain that it is something that we do every day, but we have a wide variance.  If we have a variance on a known activity, how can we possibly select a date in the future on something that we have never done and be accurate? 

Someone in the audience raised their hand and said, "Yet, we still get to work on time.  If not, we would be fired!"  A great observation.  My response to it was that there is a wide acceptance of being late due to unforeseen circumstances.  However, I have thought about the question further and wanted to expand the answer.

First, there are several adjustments when we start a new job.  We may leave really early to ensure that we are at work on time and then slowly leave later and later until we settle on the right time.  There are also times where we can't leave any earlier due to having to leave a child at daycare or other circumstances.  In almost all cases, we come to an agreement with ourselves and our employer about what is and is not acceptable.

Second, there is acceptance of things beyond our control.  If there is a major accident on the highway and someone is not to work on time, the assumption is that they are caught in traffic.  In fact, many people will defend the missing person with this excuse without truly knowing the cause.  If the daycare opened late, the employee simply apologizes when they do get to work.

Third, the number of times early and late generally fall within the probability distribution that people estimate.  When you ask them how long it takes to get to work, their first response will follow the most likely distribution plus or minus a standard deviation.

Therefore, the case still stands.  If we can't guarantee how long it takes us to get to work, even though some of us have done it thousands of times because there are just things beyond our control, then how can we ask several people to perform tasks they may have never done before and be able to guarantee a date and time of being finished?

The point of all of this is that project manager's dates of completion are only our best guess or estimation.  We have a tremendous toolset available to help select the date, yet we are still predicting the unknown.  If we could guarantee the date, then we are in the wrong business.  Maybe we should have been stockbrokers!  Just a thought.....

No Day But Today,

Rick

No comments: