Fight the Power — Why Technical Performance is a Power Law

Just returning from a month of travel (and some summer bug fighting).  Had the opportunity to attend a large portion of EVM World 2015 in New Orleans the last week of May and, as usual, came out with some food for thought.  First is that my colleague Glen Alleman is also working on the proposition for a general theory of project management.  We shared some mutual insights on the topic and the ramifications for acquisition policy, particularly in regard to high tech.

In thinking about my previous studies on technical achievement–and the ways of tracking that achievement tied to performance management–it has become clear that my paper, published in 1997 for the Defense Acquisition University Acquisition Symposium, requires some updates, given subsequent advances in our knowledge of complex adaptive systems.  It is now clear to me that the basis for determining technical achievement against a plan is not a normal distribution, but a function of a power law.

This revision in insight comes from the developing science of complex adaptive systems–which is proper categorization of a project management organization.  Dave Gordon presents some skepticism in the comments section as response to my previous musings on a general theory of project management, but I think that he falls into the fallacy of special pleading.  I often encountered such skepticism when approaching people in commercial project management in their rejection of methods used in public project management.

There are a couple of problems with this line of thinking:  first, it is based on ideology most often, and not facts.  Project management is project management, whether you are building an aircraft, a ship, a satellite, a building, a dam, or software.  There are common principles that apply to organizational dynamics, structure, and behavior regardless of the end item being developed.  One noteworthy example to prove the fallacy is the example of earned value management (EVM).  Today, especially in approaching complex projects, earned value management is considered a best practice, whether applied to public works or to private development.  But this was not always the case.  Back in the 1990s I often heard the refrain that “we don’t do that guvmint stuff” when I would suggest the use of EVM.  The overwhelming evidence is that using EVM in tracking program performance is an essential tool (though not the be-all) in project management.  No longer do we hear the refrain “that guvmint stuff.”  Second, we always tend to believe psychologically that our situation is special.  This is a normal human reaction, especially among successful people.  But, once again, our experience with using larger repositories of data, applying stochastic methods, networking analysis, and recent insights into complex adaptive systems have changed our perceptions of the discipline.

For example, our understanding of self-organized complexity has greatly influenced economics over the last decade.  Recent studies here and here (as well as elsewhere in too many places to link) show a consistent rate of failure in projects that threaten the existence of the organization or enterprise–so called, Black Swans.  This has the ring of self-organized complexity that is yet to be fully identified.

Some final thoughts.  There has been some pushback from some, but not all, people involved in applying performance management to more closely associate the results to technical achievement.  The pushback may be somewhat defensible only because it is not fully understood.  I have heard many misattributions in project management meetings regarding the concept.  Aside from my own paper, I think the community would find it interesting that Martin Marietta applied the concept for many years.  In addition, SEI at Carnegie Mellon has been applying the model for quite some time.  A Google search will reveal many other proven methods.

When we take a step back and look at our processes, what good is the end item if it does not stay within the technical framing assumptions that initiated the requirement?  The challenge is always going to be the find the best way to do this on a regular basis and ensure fidelity and credibility in our performance reporting systems.  I believe that the first company that commits to this concept and does it on a consistent basis in an integrated fashion, much as Norm Augustine committed his own company to it back in the 1990s (who endorsed technical risk management in the cited book), will have a significant competitive advantage.