Abstract for: Agile SD: Fast, Effective, Reliable

Recent concern about the progress and impact of system dynamics calls into question the means by which the method is deployed. Books, courses and published cases suggest we start by defining how the issue of concern is changing over time, then build qualitative causal-loop diagrams with stakeholders. The resulting shared mental model is taken to encompass the scope of the issue and represent the causal mechanisms involved. Stock-and-flow structures are then added and data sought with which to populate and formulate those structures, to create a working mathematical model. The process is difficult, time-consuming and unreliable. An alternative process moves directly from the performance behavior to a quantified mapping of stocks and flows. From there, interdependencies are traced and significant feedback mechanisms identified. Models are easier and faster to build, and incorporate quality from the start. They also deliver insights throughout the process – reminiscent of the “agile” approach in software development. The approach is also consistent with a complementary approach, common among leading practitioners, of leveraging proven structures repeatedly across similar cases.