Contact management API duplicate detection runs slowly when processing bulk imports

CRM Oracle CX Cloud
, , , , , , , ,
1

We’re importing 50K+ contacts daily via the REST API and duplicate detection is taking 4-6 hours. Our batch processing calls the /contacts endpoint with 500 records per request. The duplicate matching runs synchronously which blocks our import pipeline.

We’ve tried adjusting batch sizes (200-1000 records) but performance doesn’t improve much. The API response shows it’s running fuzzy matching on name/email/phone for every contact. Is there a way to optimize duplicate detection or use async patterns?

Current approach:


POST /crmRestApi/resources/11.13.18.05/contacts
Payload: [{"FirstName":"John","LastName":"Smith","EmailAddress":"john@example.com"},...]
Header: DuplicateDetection: enabled

Our indexes on Contact objects look standard. Any recommendations for batch processing optimization or tuning duplicate detection rules?

2

For async processing, use the bulk import API instead of the standard REST endpoint. Submit your batch to /crmRestApi/resources/11.13.18.05/contacts/bulk and you’ll get a job ID back immediately. Then poll /jobs/{jobId} for status. This runs duplicate detection in the background. Also, make sure your Contact object has indexes on EmailAddress, LastName, and Phone fields - these are critical for duplicate matching performance.

3

We had similar issues last year. The synchronous duplicate detection is the bottleneck. Have you looked at the matching rules configuration? We reduced our rule complexity and saw 40% improvement. Also check if you really need fuzzy matching on all three fields - exact matching is much faster.