MES/QMS/LIMS Boundaries — Avoid Creating 'Three Sets of Truths'

Summary: For the same batch of nonconforming products, MES shows it has been released, QMS has an open CAPA, and LIMS inspection reports are still in the 'Inspection Pending' status — the auditor asks, 'Is it conforming or not?' Each system gives a different answer. System integration failures often stem from unclear boundaries: which data is the master data, which events trigger actions, and which decisions depend on the outcomes of other systems. This article provides a three-layer system responsibility division, interface list, and release logic calculation to help companies build a truly traceable digital quality chain.

1. Case Study: Mismatch Across Three Systems

A chemical company, after implementing a digital system, experienced increased chaos:

• MES: Records batches, material input, and process parameters
• QMS: NCR, CAPA, and change management
• LIMS: Raw material and finished product inspection

During a customer audit, batch B20260701 was randomly selected:

The auditor immediately identified a nonconformity: 'What is the basis for your release? If the LIMS report is not yet approved, how can MES release the batch?'

Root Cause:

• MES release business rules do not verify the final LIMS judgment — MES only checks if the current workstation is complete, without confirming whether the quality inspection has been signed off.
• LIMS detects OOS (Out of Specification) but does not automatically trigger QMS NCRs; QE relies on emails to learn about abnormalities, which are often overlooked.
• The batch number coding rules are inconsistent across the three systems — MES uses B20260701, LIMS uses B2026-0701, and QMS uses BATCH20260701, leading to manual reconciliation issues.

This case is not unique. In the digital transformation of manufacturing, over 60% of QMS projects encounter 'multiple sets of truths' within the first year of implementation. The root of the problem is not the software functionality but the lack of thorough boundary definition before system deployment.

2. System Positioning (In One Sentence)

Core Principle: Inspection conclusions are based on LIMS; release decisions are executed by MES + QMS rule verification; QMS records nonconformities and improvements but does not replace MES for storing process parameters. In short, MES is the 'executor', LIMS is the 'referee', and QMS is the 'prosecutor' — each role has its own responsibilities and should not be conflated.

If one system attempts to take over the functions of another (such as storing process parameters in QMS or managing CAPAs in MES), it may seem to reduce integration workload in the short term, but it will lead to three long-term issues: data inconsistency due to duplicate entry, performance degradation due to unbalanced system loads, and limited scalability when business volume increases.

3. Boundary Matrix

The core design concept of this matrix is Single Source of Truth (SSOT) — each type of data is created and updated by only one system, and other systems obtain read-only copies through interfaces. The benefit of this approach is that regardless of which system is used for querying, the data seen is consistent.

Each line in the matrix must clearly define three elements: Data Structure (field definitions and formats), Synchronization Trigger Conditions (real-time/batch/event-driven), and Exception Handling (fallback plan when the interface fails). Many projects fail not because the technology is unachievable, but because these details are not sufficiently discussed during the architecture design phase.

4. Calculation Example: Release Interface Logic

Release Business Rules:

1. LIMS Conclusion = 'Conforming' and Report Status = 'Approved' (not 'Pending Review')
2. QMS has no open NCRs for the batch (no open nonconformities)
3. MES key process steps are 100% complete (no skipped or missed steps)

Pseudo Code:

IF lims.result==PASS AND lims.report_approved
   AND qms.open_ncr(batch)==0
   AND mes.steps_complete(batch)==100%
THEN mes.allow_ship(batch)=TRUE
ELSE mes.allow_ship(batch)=FALSE

Before Implementation: MES only checks if the workstation is complete → Misrelease Rate averages 3 batches per month. Two of these 3 batches are physically released before LIMS reports are approved, requiring recall and handling, with each recall costing about 20,000 yuan.

After Implementation: 0 misreleases; average release delay +4 hours (waiting for LIMS review completion, especially for night shifts which need to wait until the next day's morning shift for review). This delay is something management must accept — trading time for correctness is a non-negotiable aspect of quality decision-making.

From a system implementation perspective, this release rule should not be hard-coded into MES business logic but managed through an independent Rule Engine. This way, when business rules change (e.g., a customer requests an additional release condition), only the rule engine's strategy needs to be modified, without altering MES code or redeploying the system.

5. Calculation Example: LIMS OOS Automatically Triggers NCR

Physical and chemical OOS (Out of Specification, out-of-spec results):

• LIMS detects OOS → automatically pushes to QMS: Pre-filled NCR (batch number, test item, actual value, standard limit, tester)
• QMS automatically assigns QE based on predefined rules and triggers a response plan (isolation instructions written back to MES)
• MES locks the inventory location of the batch upon receiving the instruction, prohibiting material withdrawal and shipment

Closure Time:

The key to this improvement is the 'pre-filled' mechanism. Originally, when manually opening an NCR, QE had to first check the results in LIMS, copy and paste them into an email, and then manually create the NCR in QMS — each OOS event took an average of 40 minutes of manual work. Automatic pre-filling compresses this time to zero and avoids human errors during copy-paste (such as incorrect batch numbers or mixed units).

Another important role of automatic OOS triggering is preventing omissions. In the original process, if the testing engineer was too busy or forgot to send an email, the OOS could be 'overlooked'. The automatic trigger mechanism eliminates this risk point, ensuring that 'every abnormality must be responded to'.

6. Master Data Governance

The foundation for integrating the three systems is unified master data, at a minimum, the following must be aligned:

• Material codes and descriptions
• Supplier codes and names
• Measurement units and conversion relationships
• Inspection item codes and standard values

Recommendation: Master data for materials should be authoritative from ERP, with MES, LIMS, and QMS synchronizing read-only copies. No system should be allowed to modify master data locally. The synchronization mechanism should use a message queue (such as Kafka or RabbitMQ) to ensure that master data changes are transmitted in real-time to all subsystems.

Master data governance is the 'foundation' of a digital quality system. Without a solid foundation, even the most beautiful integration will result in data inconsistency. It is recommended to conduct a comprehensive master data cleanup before system deployment, correcting duplicate, missing, or erroneous records in existing data, and then establishing a continuous master data maintenance process.

7. Implementation Roadmap

Phase 1 — Draw end-to-end swimlane diagrams (from material requisition to shipping, covering the complete business process across the three systems)

Phase 2 — Define boundary matrix + interface tables (field-level definitions, interface documentation between each pair of systems)

Phase 3 — Pilot 1 product line with 2 interfaces (release verification interface + OOS automatic trigger interface, to validate integration effectiveness)

Phase 4 — Full-scale deployment + audit trails and permissions (establish operation logs between systems as per 12.1.2 requirements)

Not Recommended: A full-scale rollout all at once. Choose one product line for a pilot to validate the correctness of interface logic and system stability, collect operational feedback, and then roll out to other lines. Experience shows that system integration issues during the pilot phase almost always involve corrections to interface documentation — it is much faster to make changes on one line than on ten.

8. Common Pitfalls

9. Conclusion

MES/QMS/LIMS are not three separate software systems, but three roles on a quality chain. Only through clear boundary definitions and reliable interface integration can a complete management loop be formed. The success or failure of integrating the three systems is 70% dependent on boundary division during the architecture design phase and 30% on interface debugging during the implementation phase. Design is more important than coding.

It is worth noting that one should not be misled by the 'all-in-one' claims of software vendors. While there are indeed integrated solutions on the market that claim to cover 'MES+QMS+LIMS', for most medium-sized manufacturing enterprises, choosing specialized systems in each domain and integrating them through interfaces is more advantageous in terms of flexibility and replaceability. Specialized systems have faster iteration speeds and deeper industry expertise, and when you need to replace one, you don't have to overhaul the entire architecture.

Human factors are equally important. System integration is not just a technical project but also a management change project. MES, LIMS, and QMS often belong to different departments (Production, Quality Inspection, Quality Management System). When implementing an integration project, a cross-functional joint project team should be established to clearly define the interface responsibilities and decision-making authority of each department. Without organizational support, technical solutions are unlikely to receive operational cooperation during the implementation phase.

Suggestion: This week, randomly select 20 batches from the last month's shipped batches and reconcile them across the three systems — any inconsistencies found are potential seeds for future audit issues. Address data consistency issues before discussing process optimization.

Three systems, three sets of truths, is the biggest illusion in digitalization

Knowledge Number: 12.3.1

Version: v20260711

Author: Quality Excellence Think Tank Quality Excellence Think Tank is dedicated to providing systematic professional knowledge, methodologies, and practical tools for quality management practitioners, helping companies continuously improve their quality capabilities.