Responding to change over following a plan
Truth in advertising, I am a certified Project Management Professional (PMP). I’m proud to be a PM. If you meet me on the street, I probably have my member ID card from the Project Management Institute in my wallet. As a PM, I love to make a plan. I will talk about this in depth in a later chapter, but right now we have a tendency to make the wrong kind of plan. We set the goal posts for our plan, in most cases, a year or more out. If you think about the duration of the acquisition process, most government agencies are really setting their plans a year and a half out. That gives you three to four months to get the contract awarded (that is somewhat optimistic BTW) and then a year to execute. This is so fricking stupid, I literally have to stop writing this for a couple minutes because it makes me too mad.

We absolutely have to stop shooting ourselves in the foot. But we do it all the time, and we will likely do it a thousand more times before people start waking up. But shooting yourself in the foot is a good place to start. If you have a gun in your hand and you aim at your foot, I bet that most people would be able to actually put a bullet into their foot. It is like, 3 or 4 feet away, right? My point is that we can regularly hit a target that is 3 or 4 feet away. Now imagine that you could take your foot off of your leg and move it 78 feet away from yourself. Shooting your foot becomes a lot more difficult, right? That is my point, and what the manifesto people were thinking about.

When the target is so far away, a lot more of the risks and externalities come into play. At 78 feet you need to consider wind speed, air temperature, humidity, and the movement of the target. When it is 4 feet away, you don’t need to think as much about those variables because the bullet will travel for such a short time. That is what we in the government need to think about. When we have a time horizon that is so far away, a lot more stuff can happen that will affect our project. We will go crazy trying to account for all of it and there is so much that we could never foresee the further out you go.

Cone of Uncertainty

This concept is what Steve McConnell talks about in his seminal book, Rapid Development3 . What he points out is that for all projects there is a cone of uncertainty. At the beginning of a project you generate estimates about how long it will take, how much it will cost, how much effort you will need, how many risks you will need to deal with, how many defects you are likely to encounter, etc. The further out that horizon is, the wider your estimate will need to be to accommodate all of the variables.

variability chp 3 For argument’s sake, let’s agree that a project is a $1 million project. That would be the nominal cost for the project. If the duration of the project is a year and a half (78 weeks), then your estimate should not be $1 million on day 1. The reason for this is because the No. 1 rule for estimating is that we never give people a single number. An estimate is always a range of numbers until you get to the last day of the project and are delivering. As such, gun to your head, or to build upon my analogy, gun to your foot, on day 1 of a 78-week project that you think will be $1 million, how much will it cost and how long will it take? If you say $1 million and 78 weeks, you will likely lose a foot because it is extremely unlikely that it will cost exactly $1 million and take exactly 78 weeks. You should use a range based on the cone of uncertainty. At the beginning of the project the variance should be very wide. In day 1 of that project I would say that I have confidence that it will cost between $500,000 and $2 million and that it will take between 52 and 144 weeks. This is my rough order of magnitude (ROM) estimate, which is -50% +100%.

This is why what we do is so stupid. We try to plan for something that is probably $1 million, but we need $2 million because it is such a long duration. We think it will need 78 weeks, but we need to budget for up to 144 weeks. Do you think most PMs do that? We have an IT Dashboard that is filled with two types of development projects; those that are over budget and over schedule, and those that are lying about the performance. Except for the DoD, no other agency builds their initial projects with the budget and schedule to accommodate all the stuff that will happen in a 78-week project. DoD definitely understands this concept.

Read More About
Recent
More Topics
About
Demosthenes
Demosthenes
Demosthenes is a pseudonym for a senior Federal IT official.
Tags