Abstract for: System Dynamics-based Lean Startup Integrated Framework (SyD-LSIF)

Startups operate under conditions of extreme uncertainty, characterized by limited resources, lack of operational history, and unpredictable market responses. While Lean Startup provides an experimental approach to reduce uncertainty, it remains limited by its reliance on local validation and short feedback loops, often neglecting underlying causal structures and dynamic effects. This paper introduces the System Dynamics-based Lean Startup Integrated Framework (SyD-LSIF), a hybrid methodological framework developed within the Design Science Research paradigm. It combines empirical experimentation (Build–Measure–Learn) with structural modeling (Model–Simulate–Calibrate), enabling the transformation of business hypotheses into dynamic hypotheses and supporting targeted experimentation through causal analysis and simulations. A first illustrative case on development capacity shows how the framework supports decision-making through model-based and empirical learning. Initial simulations guide a targeted experiment, and data from four sprints confirm that the backlog can be delivered within the expected timeframe. A second case on legal uncertainty illustrates Problem-Structure Solution-Structure Fit (PSSSF), where AI and human escalation modify the underlying causal structure. The framework contributes at both qualitative and quantitative levels. At initialization, it supports the emergence of critical hypotheses, key variables, and more precise questions, enabling a targeted build. More broadly, it shifts analysis from static metrics to dynamic trajectories, allowing the identification of thresholds, trade-offs, and system equilibria. However, quantitative modeling requires non-negligible effort and should be applied when the expected analytical and decision-making benefits justify the investment. AI used for translation from French and minor wording improvements