Planning workflows¶

A planning engine produces a result; a workflow turns that result into an organizational decision. The workflow defines cadence, inputs, owners, versions, approvals, exceptions, and the handoff to the next planning horizon.

Connected horizons¶

The horizons should share governed master data and explicit version handoffs. They do not need to run in one monolithic calculation.

A practical monthly cycle¶

1. Close actuals. Reconcile sales, inventory, orders, production, and deliveries through the agreed cutoff.
2. Refresh demand. Generate the statistical baseline, collect collaboration, approve consensus, and record the released version.
3. Create unconstrained supply. Quantify the production, purchase, transfer, and capacity needed for full service.
4. Create constrained supply. Apply capacity, lead time, inventory, eligibility, and policy.
5. Resolve exceptions. Identify affected sales and decide capacity, timing, sourcing, priority, or demand actions.
6. Review finance. Compare service, inventory, cost, margin, and risk.
7. Approve the Working Plan. Freeze the plan or time fence required by execution.
8. Operate through S&OE. Revisit short-term exceptions without silently rewriting the monthly assumptions.

Versions and traceability¶

Every decision cycle should identify:

• input cutoff and source versions;
• Demand Plan and Released Demand;
• Supply Plan, network version, profile, and scenario;
• Unconstrained, Constrained, and Working Plan roles;
• collaboration changes, author, reason, and approval;
• economic assumptions and comparison baseline;
• operating release date and frozen horizon.

A Planning Book is a decision workspace, not a replacement for version governance. Approved edits must remain attributable to the person, stage, and plan that received them.

Collaboration and value contribution¶

Demand collaboration can be configured as layers for Marketing, Engineering, Sales, Operations, or other roles. Accuracy by stage helps determine whether a contribution improves the baseline. The same principle applies to supply exceptions: an override should state what restriction it resolves, which demand it protects, and which risk it introduces.

Good governance does not prohibit manual judgment. It makes judgment visible and measurable.

Exception management¶

An exception is actionable when it contains:

• material, location, customer, and period affected;
• comparison baseline and measured gap;
• physical or policy cause;
• sales, service, inventory, or financial impact;
• decision owner and due date;
• proposed alternatives and required capacity or lead-time change;
• final decision and plan version.

Avoid maintaining parallel spreadsheets whose numbers cannot be reconciled to the versioned plan. Export can support analysis; the authoritative decision must return to the governed workflow.

Minimum operating controls¶

• one owner for each stage and data cutoff;
• explicit approval and release criteria;
• a documented calendar and time zone;
• role-based access to scope and editable measures;
• failure and late-data handling;
• reconciliation before and after each execution;
• accuracy and adherence metrics reviewed at the next cycle.

Edition comparison¶

Use the data catalog to define each handoff and the technical integration guide to make source-to-plan reconciliation repeatable.