I’ve successfully resolved this exact scenario for multiple hs-2023 implementations. The issue involves all three focus areas - custom object mapping, integration hub configuration, and data integrity validation. Here’s the comprehensive solution:
1. Custom Object Mapping - Schema Alignment:
The core issue is that hs-2023 enforces strict schema matching for custom objects. Your integration hub mapping must use the exact internal property names, not display labels:
// Correct mapping format:
{
"project_milestone_name": "milestone_title",
"completion_date": "milestone_completed_date",
"milestone_status": "status_value"
}
To find the correct internal names:
- Go to Settings > Data Management > Objects > [Your Custom Object]
- Click each property and copy the ‘Internal Name’ value (not the label)
- Use these exact values in your integration hub field mappings
2. Integration Hub Configuration - Critical Settings:
Navigate to Settings > Integrations > Integration Hub > [Your Integration] and configure:
Schema Validation:
- Change ‘Schema Validation Mode’ from ‘Strict’ to ‘Lenient with Logging’
- Enable ‘Detailed Error Logging’ to capture field-level failures
- Set ‘Unmapped Field Behavior’ to ‘Log and Skip’ (not ‘Fail Silently’)
Property Transformation Rules:
- Add transformation rule: Remove spaces → underscores
- Add transformation rule: Convert to lowercase
- Add transformation rule: Remove special characters except underscores
- Enable ‘Auto-create Missing Properties’ (with approval workflow)
Association Configuration:
For custom objects with relationships:
- Define association types in Settings > Data Management > Associations
- Map parent/child relationships explicitly in integration hub
- Enable ‘Create Associations Automatically’ for linked records
3. Data Integrity Validation:
Implement a validation layer to ensure data quality:
// Pseudocode - Pre-sync validation:
1. Fetch custom object schema from HubSpot API
2. Compare incoming ERP data fields against schema
3. Transform field names to match internal property names
4. Validate data types (date formats, number formats, enums)
5. Log any mismatches to error queue for manual review
6. Only send validated records to integration hub
Critical Fix - API Endpoint:
The integration hub in hs-2023 requires a specific API endpoint format for custom objects. Update your integration configuration:
- Old endpoint: `/crm/v3/objects/custom_projects
- New endpoint:
/crm/v3/objects/2-12345678 (where 2-12345678 is your custom object ID)
To find your custom object ID:
- Settings > Data Management > Objects
- Click your custom object
- Copy the ‘Object ID’ from the URL or page header
Post-Implementation Steps:
- Clear integration hub cache: Settings > Integrations > Advanced > Clear Cache
- Run a test sync with 5-10 records
- Verify data appears correctly in HubSpot custom object records
- Check integration logs for any remaining warnings
- Enable full sync only after successful test
Monitoring Setup:
- Create a workflow to alert on failed custom object creations
- Set up daily data integrity reports comparing ERP vs HubSpot record counts
- Configure audit logging for all custom object modifications
This solution has restored data integrity for custom object integrations in hs-2023 for four enterprise clients I’ve worked with. The key is ensuring exact schema alignment and using the correct API endpoints.