Overview:

When running a source subscription, some of the records are not matched as expected.

Root Cause:

 The most common reasons for unexpected matching results in source subscriptions are related to the configuration of data grouping and matching rules.

1) Data Group (Blocker) Configuration:

Incorrectly Defined Data Groups:** If data groups are not properly configured, records that *should* match might be placed in separate groups and never compared. This can happen due to:

• • Insufficient Criteria: Data groups might not include enough fields to effectively group related records.
  • Overly Restrictive Criteria: Data groups might be too specific, separating records that have minor variations.
  • Missing Data Groups: Necessary data groups might be absent, leading to incomplete comparisons.
  • Data Group Size: Large data groups (e.g., over 1000 records) can impact performance. While not strictly a "matching failure" cause, they can slow down processing and make troubleshooting more difficult.

The job details in the UI may display warnings if data group sizes are excessive.

2) Matching Rule Configuration:

• Insufficient Match Rules: If the matching rules don't cover the variations in your data, records that should match might not meet the criteria.
• Incorrect Match Rule Logic: Errors in the logic of match rules (e.g., incorrect operators, missing conditions) can lead to inaccurate matching.
• Confidence Levels: If confidence thresholds are set too high, legitimate matches might be rejected.
  * Conversely, if thresholds are too low, inaccurate matches can occur.

3) Data Quality Issues:

While not a configuration issue, inconsistencies in the source data (e.g., variations in spelling, formatting) can prevent matches, regardless of configuration.

Solution:

• Review Job Details:
  • Check the job details in the Network UI for any error messages or warnings related to data group size or other processing issues.
    • If the data group size is too large (over 1000) a warning appears in the error log.
    • If no errors are returned, verify that the data groups are configured to include the data from the unmatched records.
• Verify Data Group Configuration:
  • • • Carefully examine the data group configuration to ensure it includes the appropriate fields and criteria to group related records.
      • Adjust the data group criteria to be more or less restrictive as needed.
      • If the data groups are configured correctly but the records are not matched as expected, check the matching confidence levels (in Advanced mode under Match Rules/classifiers) 
• Validate Matching Rules:
  • Review the matching rules to confirm they are correctly defined and cover the expected variations in your data.
  • Pay close attention to match rule logic and confidence level settings.
• Address Data Quality:
  • Implement data cleansing processes to standardize the source data before it is loaded into Network.

• Grouping data
• Creating data groups
• Match, merge, and survivorship