Tuesday, October 27, 2009

Common Project Risks and Agile Mitigations, Part 8, Epilogue

Part 7 closed out the discussion of issues that the various sources identified as contributing to project failure. As promised at the end of that Part, this Part looks back on the first 7 parts and offers some summary comments. I’ve also added some further bibliography entries that I encountered them while doing this series. While they were not used to compile the list of failure issues, they do discuss project failure issues.

As implied in Part 1, the series was inspired by my having reviewed a number of surveys and articles on why projects failures occur. It seemed to me that the Agile Values and Principles, if adopted using a number of generally accepted Agile practices and techniques, could mitigate many of these issues. While an Agile approach cannot pretend to solve all the identified problems, its communication and feedback model for team-based work is far more likely to surface and deal directly with such problems when they are usually less severe and less costly to address.

However, these project failure risks have existed, in some cases, for decades. People have been dealing with them long before the Agile Manifesto was created. Over the years, the failure rate has been declining, as some of the source surveys have noted. The recommended solutions have been codified in various bodies of knowledge (e.g., PMIBoK, SWEBoK, CSQE BoK), models (e.g., CMM®/CMMI®), and standards (e.g., ISO 9001, IEEE S2ESC, ISO/IEC JTC1/SC7). These approaches to the problems have been (often unfairly) characterized by voluminous documentation, significant cost, and bureaucracy. Agile methods have (equally as unfairly) been characterized by applicability only to small, low-risk efforts. The BoK, models, and standards approach is a reflection of what people believe to be currently understood “best” practices derived from those used in other (non-software) engineering and project domains. Agile methods reflect the view that a less complex, less prescriptive approach is possible which relies less on detailed process guidance and more on process transparency and person-to-person communication.

What is interesting, however, is the position often taken by both approaches with regard to why projects employing them continue to have problems. From a phase-based, sequential project perspective, the view is that such problems can be avoided if only project personnel would follow the existing, accepted, and well-defined project management and software engineering BoKs, models, and standards. That is, projects fail because people do not follow known methods and procedures effectively, i.e., they do not carry out such guidance “right.” Interestingly, when an Agile project is not successful, we often hear the same things, i.e., it failed because people did not do Agile (or a specific method) “right.”

So the question, in either case, is: can it merely be a matter of doing what is predetermined to be “right,” whether traditional or Agile? And, why is it so hard to do what is “right” to achieve the results expected from doing so in either case? Proponents of an Agile approach say the BoK/model/standards approach suffers from too much overhead, inapplicable detail, and inflexibility causing an inability to respond promptly to the circumstances and changes that lead to failures. Proponents of the BoK/model/standards approach say Agile methods suffer from ad hoc behavior, short-term thinking, and transfer of significant risk responsibility to the business and customers.

What may be most true of either approach is that a formulaic application of their practices and techniques, with insufficient knowledge and understanding of their basic principles can produce an inflexibility (or inability) to respond when circumstances require. It may be that the “best” approach would be to ensure that the guidance in the BoKs, models, and standards is understood by those working on projects, but that the Agile Values and Principles are used as the structure within which to apply that knowledge.

Related Bibliographical Items

No comments:

Post a Comment