Based on managing several part management migrations with complex ERP and MES integrations, here are the key challenges and lessons learned:
Attribute Mapping Challenges
The most common and impactful challenges:
Data Type Mismatches:
Your legacy system might store manufacturing lead time as text (“5 days”), but SAP expects an integer (5) with a separate unit of measure field. During migration, you must transform these attributes to match downstream system requirements.
Example transformation needed:
- Legacy: LeadTime = “5 days”
- Agile: LeadTimeDays = 5, LeadTimeUOM = “Days”
- SAP expects: LEAD_TIME (integer), UOM (code)
Create a comprehensive attribute mapping matrix:
Source Attribute → Agile Attribute → ERP Field → MES Field → Transformation Rule
LegacyPartNum → PartNumber → MATNR → PartID → Direct mapping
MfgLeadTime → LeadTimeDays → LEAD_TIME → ProdLeadTime → Extract integer
WorkCenter → MfgLocation → ARBPL → WorkCenterCode → Lookup table conversion
Missing Required Attributes:
ERP and MES systems often have mandatory fields that weren’t required in your legacy PLM. During migration, you must either:
- Populate with default values (document the business rule)
- Derive from existing data (e.g., calculate unit cost from total cost and quantity)
- Obtain from supplementary data sources
We encountered this with SAP’s Material Type field, which was mandatory but didn’t exist in the source PLM. We created a classification rule based on part number prefixes to assign appropriate material types.
Multi-System Attribute Dependencies:
Some attributes feed multiple downstream systems with different requirements. Manufacturing routing information might need to be in Agile for MES consumption but also summarized differently for ERP costing. This requires careful design of Agile’s attribute structure to satisfy both integrations.
Custom Attributes and Extensions:
If your integrations rely on custom Agile attributes, ensure these are recreated in the new environment before migration. We’ve seen cases where custom attributes weren’t migrated, breaking integration mappings that referenced them.
Integration Cutover Planning
A structured cutover approach minimizes disruption:
Phase 1: Pre-Cutover Preparation (4-6 weeks before)
- Freeze integration configuration changes
- Create integration test environment mirroring production
- Migrate sample part data to test Agile instance
- Validate test integrations with sample data
- Document all integration endpoints and data flows
- Identify critical parts that cannot have data gaps (A-items, production-active parts)
Phase 2: Parallel Run (2-3 weeks)
Run both old and new PLM systems simultaneously:
- Configure dual integration feeds (legacy PLM → ERP/MES and Agile → ERP/MES)
- Implement data reconciliation checks to compare outputs
- Monitor for discrepancies in real-time
- Maintain change synchronization between systems
This approach requires additional infrastructure but provides safety net. For your 35,000 parts, focus parallel run on:
- 100% of active production parts (A and B items)
- 50% sample of C items
- All parts with recent engineering changes
Phase 3: Cutover Weekend
Execute final migration and integration switch:
Friday Evening:
- Complete final part data migration to Agile
- Run comprehensive validation queries
- Execute integration smoke tests
- Verify critical part attributes in downstream systems
Saturday:
- Disable legacy PLM integrations
- Enable production Agile integrations
- Monitor integration logs continuously
- Execute test transactions in ERP and MES
- Validate BOM explosions and work order creation
Sunday:
- Run production validation scenarios
- Test end-to-end flows (engineering change → ERP update → MES work instruction)
- Prepare rollback plan if critical issues found
Monday Morning:
- Hypercare support mode
- Real-time monitoring of integration queues
- Rapid response team for integration issues
Lessons from Real Projects
Lesson 1: Integration Testing Must Include Business Scenarios
Technical integration tests (“Does the API call succeed?”) aren’t sufficient. Test complete business processes:
- Create engineering change in Agile → Verify change propagates to ERP → Confirm MES receives updated work instructions
- Release new part in Agile → Verify part appears in SAP → Confirm procurement can create purchase requisition
We discovered integration issues only during business scenario testing that passed technical API tests.
Lesson 2: Attribute Validation Before Migration
Validate part attributes against downstream system rules before importing to Agile:
-- Example: Validate part numbers meet SAP format requirements
SELECT part_number, 'Invalid SAP format' as issue
FROM staging_parts
WHERE LENGTH(part_number) > 18 -- SAP MATNR limit
OR part_number NOT REGEXP '^[A-Z0-9-]+$'; -- SAP allowed characters
Run these validation queries against staging data, fix issues before Agile import. This prevented 2,500 parts from failing ERP integration in our project.
Lesson 3: MES Integration Requires Manufacturing Input
Manufacturing engineers must validate work instruction attributes. We had technically correct migrations where attribute values were preserved, but manufacturing couldn’t use the parts because work instruction formatting was lost or attribute semantics changed.
Conduct manufacturing validation sessions:
- Show sample parts in new Agile system
- Have engineers verify work instructions are usable
- Test actual MES consumption of part data
- Confirm shop floor displays show correct information
Lesson 4: Supplier Integration Often Overlooked
If suppliers access Agile for part specifications or have integration feeds, coordinate with them:
- Provide advance notice of migration and any part number changes
- Share updated part cross-reference tables
- Test supplier portal access post-migration
- Verify supplier integration endpoints still function
We had a supplier integration failure that delayed production because supplier systems couldn’t find parts under new Agile part numbers.
Lesson 5: Phased Cutover by Product Line
For very large migrations, consider phased cutover by product family:
- Week 1: Migrate and cutover Product Line A parts
- Week 2: Migrate and cutover Product Line B parts
- Week 3: Migrate and cutover remaining parts
This limits blast radius if issues occur and allows learning from early phases. We used this for a 50,000+ part migration across multiple manufacturing sites.
Practical Recommendations for Your Project
Given your scope (35,000 parts, SAP ERP, Siemens MES):
- Create comprehensive attribute mapping matrix covering Agile → SAP and Agile → MES
- Build validation queries that check part data against both SAP and MES requirements
- Plan 2-week parallel run focusing on active production parts
- Schedule cutover during planned production downtime if possible
- Establish 24/7 support coverage for first week post-cutover
- Prepare rollback plan with detailed steps and decision criteria
- Conduct manufacturing validation sessions before cutover
- Test complete business scenarios, not just technical integration calls
The integration complexity typically doubles the overall migration effort. Budget accordingly and don’t underestimate testing requirements. Our most successful migrations allocated 40% of project time to integration validation and cutover activities.