About the Enrichment and Data Quality Services

The B2B customers with D&B data product includes D-U-N-S match enrichment and two D&B data blocks (Company Information L2 and Hierarchies and Connections L1) that enrich your data with corporate hierarchy and other referential data. This enrichment can help solve problems such as:

• Missing D-U-N-S numbers.
• Missing corporate hierarchy information.
• Out-of-date information, such as company names that have changed due to acquisition.
• Missing details about an organization, such as number of employees, revenue, SIC codes, national identifier, and so on.
• Misidentifying different companies that reside in the same building as the same company.

See Dun & Bradstreet Enrichment for complete details.

This data product also includes data quality services for the following values:

• Country code
• Company name

These data quality services supplement your source data with standardized, validated values for these essential attributes. Tamr Cloud examines values for the template's attributes, and adds any resulting validated, standardized values to each record in new enrichment-specific attributes. The original values mapped from your source datasets remain present and unchanged.

About the Clustering Model

The B2B customers with D&B clustering model identifies records that represent the same entity by first considering the D-U-N-S number. If two records have been enriched with the same D-U-N-S number, they are clustered together, provided that the D-U-N-S match is above the Confidence Code threshold or matches any specified Match Grade Patterns specified in the Set Confidence Code Thresholds and Match Grade Patterns mastering flow step.

Note: If you have mapped D-U-N-S numbers from your source datasets to the unified source_duns_number attribute, those D-U-N-S numbers are considered high confidence and are assigned a Confidence Code of 10.

Then, the clustering model considers the similarity of values for the following fields, and uses decision-tree logic to accurately identify records that refer to the same entity:

• Company name variations
• All address fields
• Phone number
• Website

For low-confidence (unaccepted) D-U-N-S matches, the above fields only contain information from the input fields. For high-confidence (accepted) D-U-N-S matches, the above fields contain information from both the input fields and the returned fields from the D-U-N-S Match mastering flow step.

Additionally, the model will not cluster together different high-confidence D-U-N-S numbers, except in rare edge cases.

Clustering Examples

Examples of Source Records Clustered Together

These records are clustered together because they have the same enriched D-U-N-S number and the same or similar values for company name and full address, indicating that they most likely represent the same company site.

Examples of Source Records Not Clustered Together

These records are not clustered together because they have different enriched D-U-N-S numbers and missing or dissimilar address information, indicating that they most likely represent different company sites.