Towards Improved Project Management Practice

Abstract

Projects are important to industry, but project performance continually disappoints stakeholder expectations. Organizations react to this performance problem in many ways, and purchase consultancy, training, methods and tools as possible solutions.

There is no published evidence that any of these solutions are consistently successful in improving project performance. This thesis answers the question, "What can be done to improve project management practices, and thus project performance?" by demonstrating that a novel form of continuous action research can contribute such evidence.

Firstly a community of practice was formed from practitioners with major corporations interested in answering the question, and commercially motivated to implement changes. A programme was developed that centred around project management, but linked to project and corporate performance and success.

A well-resourced support structure was established to administer the programme, facilitate dialogue, hold confidential data securely, and provide data analysis. Members provided data for the anonymous databases about their practices and about specific project results, and first-hand case studies for discussion at workshops. They discovered, shared and created both tacit and explicit knowledge through the formal programme and through informal contact.

Secondly, the thinking of practitioners, theorists and researchers was challenged. The literature on project management was found to reveal an unbalanced worldview that lacked coherent underlying theory. The literature on theory was found not to distinguish adequately between one-off "discrete" projects and the ongoing continuous operations of an organization. The academy's "paradigm wars" were found to have discouraged the creation of an appropriate research methodology.

Thirdly, different pieces of research using the community's data showed that some practices (notably aspects of risk management) lead to superior performance independently of context, while others appear to be context-dependent. No companies were found to have all the answers, and each member of the community has been able to learn from others.