Sensitivity analysis measures how much a model's outcome changes in response to changes in each of its inputs, identifying which assumptions have the greatest influence on the result, and therefore which are most worth validating.
Not all assumptions matter equally. Sensitivity analysis ranks a model's inputs by how much each moves the outcome, turning a long list of uncertainties into a short list of the ones that decide success or failure. In an evidence-driven approach this is the bridge from modeling to action: the most influential, least-supported assumptions are exactly where limited experimentation and evidence-gathering budget should go first. For how BRI carries uncertainty through its models, see the Modeling Uncertainty page at /supporting/uncertainty. Related Terminology Index entries: Uncertainty Assumptions; Monte Carlo Simulation; Evidence Level.
Sensitivity analysis is a long-standing tool of decision and risk analysis; its modern global form is codified by Andrea Saltelli and colleagues in Global Sensitivity Analysis: The Primer (2008), and its decision-analytic use traces to Ronald Howard's foundational work on decision analysis. BRI's refinement is to connect it to assumption management: Growth Forge® Software pairs sensitivity output with the Assumptions Manager so the highest-impact assumptions are automatically prioritized for evidence gathering, and confidence is scaled with investment. This keeps the analysis from being an academic exercise and makes it a direct input to Continue, Pivot, Pause, or Stop decisions.
