The Decision Policy is one configurable object

The Decision Policy is a first-class object, not a collection of scattered settings. Every behavioral choice the system makes lives in this one place. Seven sub-policies cover the surface area: Demand (pricing, promo, substitution, channel mix), Supply (sourcing, lot sizing, BOM, transport, safety stock, pegging, ATP, production scheduling — sixteen capabilities total), Optimization (objective weights, risk posture, named risk events), Autonomy (auto-commit thresholds, escalation rules), Publishing (what flows to which downstream systems), Scenario (sandbox configuration), and Governance (audit, versioning, approvals).

Each sub-policy is a configurable surface, edited through the Policy canvas, reviewed by the people in the organization who actually own that lever. Marketing edits Demand's promo section. Operations edits Supply's production scheduling rules. The CSCO edits Optimization's risk posture. Finance edits Governance's approval thresholds. The policy becomes a shared artifact rather than a buried IT setting.

The lifecycle is Draft → Active → Suspended → Retired with versioning at each step. A customer can A/B compare two policies via PVA before promoting either; can roll back a policy version; can audit when a posture changed and who changed it. Posture changes become defensible to the audit committee because every change carries its author, its timestamp, its before-and-after Delta PVA estimate, and the approval trail.

Chapter 5 dollarizes the policy via PVA. Chapter 4.4 develops the meta-optimization that finds optimal policy postures across the candidate space. Chapter 7.1 is where executives use the policy as their primary governing artifact.