In regulated manufacturing, integration between a Manufacturing Execution System (MES) and an Enterprise Resource Planning (ERP) system means a defined, validated exchange of master data and transactional data across a clear boundary: ERP plans and accounts for the work, MES controls and records execution.
Typical data flows from ERP to MES
While details vary by vendor and plant, ERP usually sends:
- Planned orders / production orders: item, quantity, required dates, plant, and sometimes priority or customer references.
- Bills of material (BOMs): required components, revisions, quantities, alternates, and sometimes effectivity dates. In many environments the master source may actually be PLM, with ERP passing a derived production BOM.
- Routings / work centers: operation sequence, standard times, work centers, and sometimes resources or tools. Again, source may be ERP or an external process planning system.
- Material master and basic item data: item codes, descriptions, units of measure, procurement type, and sometimes quality status.
- Customer / sales order references: where lot- or unit-level traceability to a customer order or contract is required.
This information lets MES create and manage executable work (e.g., shop orders or operations) with the correct structure and planning context.
Typical data flows from MES to ERP
MES returns actual execution data to keep ERP planning, inventory, and costing credible. Common flows include:
- Production confirmations: quantities produced, scrapped, or reworked against an order and operation, with timestamps.
- Material consumption: which components, lots, and quantities were consumed, including substitutions and over/under-issues.
- Inventory movements: finished goods and intermediates moved into stock, WIP transfers, and sometimes location-level updates (if ERP owns inventory).
- Labor and machine time: operation actuals for costing, capacity analysis, and OEE-related metrics, where ERP is the system of record for cost.
- Quality results and holds (high level): pass/fail, usage decisions, or holds that affect availability or release of stock, often summarized rather than detailed test data.
The granularity of feedback depends on what ERP can handle and what has been validated across systems. Many plants keep detailed, record-by-record traceability in MES or a specialized QMS and send only the minimum necessary aggregates to ERP.
Division of responsibilities
MES–ERP integration works best when each system has a clearly defined role:
- ERP: demand management, MRP, purchasing, high-level capacity planning, inventory valuation, and financial accounting.
- MES: detailed scheduling and dispatching, work-in-process control, detailed traceability, enforcement of work instructions and recipes, in-process quality data capture, and operator guidance.
In many brownfield plants, some production and quality steps still occur on paper, spreadsheets, or legacy terminals alongside MES and ERP. Integration needs to account for these gaps to avoid double-booking production or creating untraceable material movements.
Integration patterns and technical options
Common patterns include:
- Batch file exchange: scheduled CSV/XML/IDoc transfers, often easier to validate but less real-time. Common in heavily regulated, change-controlled environments.
- Web services / APIs: more synchronous and near real-time but require tighter change control, versioning, and monitoring to avoid unplanned outages.
- Message queues / middleware: decouples systems via an integration layer (ESB, iPaaS, or custom broker). Adds robustness and observability but increases architectural complexity.
The choice affects latency, recoverability, and validation effort. For example, moving from nightly batch to real-time APIs may improve planning accuracy but requires more sophisticated failure handling, rollback strategies, and test coverage.
Key constraints in regulated, long-lifecycle environments
MES–ERP integration in regulated or aerospace-grade contexts is constrained by:
- Validation and traceability: every automated data exchange that affects product, process, or quality records usually needs documented requirements, testing, and change control. Interface logic can be audited.
- Long equipment and system lifecycles: MES and ERP may be on different upgrade cycles, with decade-old interfaces that cannot be easily reworked without requalification and downtime.
- Downtime risk: if integration breaks, ERP may plan based on stale data or MES may receive invalid orders. Plants often design manual fallback procedures, which must be realistic and trained.
- Data ownership and mastership: conflicts arise when BOMs, routings, and quality dispositions exist in multiple systems. Integration must reflect a clear system of record for each object and attribute.
Because of these constraints, attempts to fully replace either ERP or MES with the other usually fail in regulated environments. ERP modules marketed as “light MES” typically do not handle detailed traceability, complex routing logic, or shop-floor usability at the level needed; conversely, MES rarely replaces the full financial and supply chain scope of ERP. Integration, not replacement, is the realistic long-term strategy.
Practical scope definition for MES–ERP integration
When scoping integration, plants with existing systems should make a few decisions explicitly:
- What is the order of record? Decide whether ERP or MES owns the authoritative production order and status. Avoid dual maintenance of order structures.
- Who owns WIP and inventory? In some plants, ERP tracks only inventory at the warehouse level, while MES manages WIP and sub-operations. In others, ERP expects operation-level confirmations for costing.
- What level of traceability is required where? Decide which details stay in MES/QMS (e.g., per-measurement data, signatures) and what summary data must be visible in ERP to support planning and customer commitments.
- How are exceptions handled? Define how rework, nonconformances, deviations, and unplanned operations are reconciled between systems so that orders and costs stay consistent.
In a brownfield environment, it is common to integrate incrementally, starting with core flows (orders down, confirmations and material consumption up) and only later adding finer-grained integration, once data quality, processes, and validation practices are stable.