An end-to-end causal modeling framework for advanced attribution in supply chain operations

Effective attribution of causes to outcomes is crucial for optimizing complex supply chain operations. Traditional methods, often relying on waterfall logic or correlational analysis, frequently fall short in identifying the true drivers of performance issues. This paper proposes a comprehensive framework leveraging data-driven causal discovery to construct and validate Structural Causal Models (SCMs). We contrast this approach with baseline models derived from existing business definitions or metric-guided Large Language Models (LLMs). The core methodology involves (1) discovering a Directed Acyclic Graph (DAG) from observational data using the PC (Peter-Clark) algorithm, (2) comparing it to a baseline DAG, (3) building SCMs from these DAGs using DoWhy’s GCM module, (4) rigorously validating both DAGs (via falsification tests) and SCMs (via mechanism and model fit evaluations), and (5) utilizing the validated SCM to perform advanced causal queries—including root cause attribution, intervention analysis, and counterfactual reasoning. We illustrate the framework’s superiority over traditional methods through its application to a supply chain KPI, demonstrating how it provides deeper, actionable insights. Results suggest that data-driven SCMs, when properly validated, offer more robust and nuanced attribution than simpler rule-based or purely qualitative models. Our results maintain analytical accuracy while utilizing representative metrics instead of proprietary organizational data.