I was reading through an old text

Project Management With Cpm, Pert and Precedence Diagramming (Hardcover) isbn 0442254156

book of mine, just the other day, and two things really struck me, from reading this.

Firstly, that again, here is evidence that the industry (in this case the Project Management pool of “experts”) has lost the plot - there are a broad range of concepts, and approaches here in a 45 year old text, that your average Project Manager of today, wouldn’t even consider let alone understand. I have bemoned this before, and will explore in more detail again, but there is a definite failing in our ability to learn from those who have gone before us, learning from them, more of the “why” than the “how”.

Secondly, one of the specifics that I took from reading this, is that with all of the “Project Management” in this day and age, we haven’t learnt how to apply the core science. This text, touches on the concept that the predicted time to completion of a particular task in a particular project is subject to variation, under standard probabilistic models.

What don’t we see today. Let me insert this into an example. How many companies have done, say a Microsoft “Active Directory” rollout (let alone, how much has Microsoft had the opportunity to thoroughly qualify and quantify the process of doing so).

So why don’t we see a greater collection of “standard” and tested Project Plans for such things.

A project plan is firstly the collection of steps necessary - that’s easy, then the efforts and skills required to complete these, as well as the relationships between all factors.

Surely we should be heading towards the situation whereby say, such a task/project is something that we can pull from a previously defined (and tested) repository of project plans. Part of the definition is the “variance” factors, such as the number of clients (or the size of the house, you are building), and more importantly, out of this comes a feedback mechanism, that a) tracks the conformity to the original “model”, b) documents necessary modifications or improvements to the model, and c) most importantly, validates the original model. Mathematically, if the standard deviation(s) of the factors of all tasks (and the total project) are found to be consistent over 50 or 500 or 5000 samples, then the “quality” of the core model improves…

Is it just me, or isn’t this where the whole Science of Project Management should really have been heading…