Demand Planning That Reads Your ERP and MES

Dashboards describe what already happened. This brief shows how demand forecasting becomes an executable, explainable plan when it reads live ERP and MES — orders, inventory, open POs, BOM, routing, capacity, and downtime — through Analytics, Prediction, and Planning, with Optimization and Decision as the closed-loop handoff.

Audience and purpose

This brief is for manufacturing and supply-chain planners who still rebuild the executable plan in a parallel spreadsheet, and for ops and IT leaders evaluating whether a planning stack can sit on live ERP and MES without inventing a second source of truth.

What this paper does. It frames demand inside a continuous pipeline rather than as a standalone forecast tool.

What this paper does not do. It does not claim quantified forecast accuracy, ROI, or automation percentages as proven outcomes. It does not name customers. Inventory depth belongs to the companion inventory brief.

The demand–plan gap

Manufacturing demand planning breaks when the forecast is not the same object as the executable plan. Demand is predicted in one place, constrained planning happens in another workbook, and the action that should close the loop is still a meeting decision.

• Signal lag. By the time a weekly extract shows demand shift or capacity shortfall, the window inside lead time has narrowed.
• Constraint blindness. A forecast that ignores live BOM, routing, and work-center capacity is a wish list.
• Broken handoff. Prediction does not reliably become a constrained plan, and the plan does not become an explainable next action.

What ERP/MES-honest requires

Continuous live feeds — not a one-shot extract:

• Live demand, order, and inventory data
• BOM, routing, and work-center masters
• Supplier lead-time and performance history
• MES machine, shift, and downtime feeds
• Procurement receipts and open POs

For a discovery workshop, “ERP/MES-honest” means mapping which of those objects actually feed the prospect’s demand path.

Demand inside the pipeline

Five engines run as one continuous path: Analytics → Prediction → Planning → Optimization → Decision. Each level hands the next something executable.

• Analytics — continuous demand KPIs, classifications, and trends from the same live feeds.
• Prediction — demand forecasts plus lead-time / delay-risk and capacity-shortfall scoring from ERP/MES history.
• Planning — demand (and adjacent procurement and production) plans on real BOM, routing, and capacity.
• Optimization / Decision — feasible allocation and an explainable next action a planner can approve.

Capability surface

Qualitative practice language for Demand Forecasting & Planning:

• ML-based demand forecasting
• Supply chain risk modeling
• Seasonal and trend analysis
• Handoff into inventory optimization (owned in depth by the inventory brief)

How to use this in discovery

Bring planners and IT into the same session. Map live objects to pipeline levels. Pick whether demand or master data proves the gap fastest. Leave quantified ROI claims for engagement scoping — not for this paper.