Customer and Supplier MDM Enhancements

Summary

The following enhancements to Customer Experience Data Cloud (CXDC), Business Partner Data Cloud (BPDC), and Supplier Data Cloud (SDC) functionality have been made as part of the 2025.4 update. These changes are outlined below and described in the Details section that follows:

  • This update enhances unmerge functionality for restoring multi-valued references and data containers by addressing cases where survivorship rules did not re-write identical elements during merging, as well as merges of golden records or source records with missing elements.

  • Concurrent Machine Learning Matcher invocations now enhance Match and Merge Importer performance and eliminate import slowdowns.

Details

Enhanced unmerge support for multi-valued references and data containers

The 2025.3 update introduced enhanced unmerge functionality in STEP, allowing the restoration of manually created multi-valued references and multi-valued data containers that were removed from the surviving record during a merge process.

With the 2025.4 update, this functionality is further improved to handle unmerge scenarios where survivorship rules did not re-write identical elements during merging as well as scenarios with merges of a golden record or source record with missing elements. Previously, this prevented the unmerge logic from restoring all relevant elements. The updated algorithm now reads data from the deactivated golden record or source record to determine the correct restoration action.

These enhancements enable users to reverse merge operations more comprehensively, ensuring that all applicable data elements are restored and data integrity is preserved.

For more details, refer to the topic Match and Merge Clerical Review - Unmerge in the Matching, Linking, and Merging documentation.

Enhanced performance and scale of Match and Merge Importer

This update enhances the performance of the Match and Merge Importer when the Machine Learning Matcher is used during data imports. Previously, invoking the Machine Learning Matcher was a blocking operation in the ranking process, which significantly slowed down import performance. With this enhancement, many invocations are now made to the Machine Learning Matcher concurrently, enabling the Match and Merge Importer to maintain optimal import speeds even when the Machine Learning Matcher is configured.

Organizations can now leverage the advanced matching capabilities of the Machine Learning Matcher without compromising import throughput, ensuring efficient data processing for large-scale data imports.

For more information about the Match and Merge Importer, refer to the topic IIEP - Configure Match and Merge Importer Processing Engine in the Data Exchange documentation.