Simulating Work Processes and Organizations to “engineer” Strategic Projects

Experience shows that companies consistently underestimate the time and effort to develop complex products quickly. Managers of strategic projects can often anticipate effort to do direct work, but they routinely underestimate effort to coordinate and do rework. Neither managers’ intuitions nor standard project management tools can effectively predict the consequences of complex and dynamic interactions that emerge when a project team executes highly interdependent activities concurrently. This paper introduces VitéProject, a software tool and set of methods to model, simulate, analyze and optimize the design of strategic projects and their implementing organizations. We introduce a test case of developing an ethylene plant. Over the past 10 years, we have demonstrated that these software tools and methods can help managers of strategic projects to document critical aspects of their project design and manage the complex interactions between the variables that influence project execution, product and project objectives, activity schedules, direct work, coordination and rework, and project team organization. INTRODUCTION Based on our research at Stanford University (Kunz et al 98, Levitt et al 94), we developed tools and methods that allows a product or project manager to model and estimate the relative schedule, cost and quality risks of both activities and participants in a single, fast-paced project. In our Stanford teaching and research and in subsequent commercial work through Vité, we have now used these tools and methods for a number of clients in the oil and gas, software, high-tech and construction industries. These tools and methods represent the first significant step in our vision that a manager should design an organization in the same way that an engineer designs a bridge: first build computational models of the design, then optimize the model, and finally implement the design. It allows pre-project “sanity checks” and scenario analyses to optimize goal tradeoffs at the early stages of configuring a product development team. As the project proceeds, it helps a manager to guide interventions, set specific measurable goals and track performance against goals. ViteProject provides computational tools and methods to analyze decision-making and communication behavior and thereby to support true organizational (re) engineering. This article introduces the underlying theory, the implementation of the theory as a computational model, and results from industrial test cases. A model can be “run” by a discrete event simulation. Model output behaviors include the predicted time to complete a project, the total effort to do the project, a measure of process quality, and several charts that identify specific activities and actors at risk for increasing project duration. An additional model output is the predicted time-varying backlog of individual actors. The results are detailed and specific, so they can guide specific managerial interventions in a project team and can support sensitivity studies of the relative impact of different organizational changes.