Open topic with navigation

Matching, Linking, and Merging Configuration

Considering the robust nature of the matching, linking, and merging functionality, configuration can be complex. Before beginning the configuration it is necessary to develop a complete deduplication strategy. Forming a strategy requires an intimate knowledge of the data in question. Data profiles are a great way to examine your data and can aid in developing your strategy. Ultimately, your strategy could be:

• Identify duplicates in order to address them (via the Matching component)
• Keep duplicates but identify them by linking them to a golden record, for example, the same product from different suppliers (via the Match and Link component)
• Eliminate duplicates but keep a golden record of the 'best' data - for example, the same person but with different email addresses (via the Match and Merge component)

For more information about deduplication strategies, see the Configuration Example - Basic (here) and Configuration Example - Advanced (here) sections of the Matching, Linking, and Merging documentation.

For more information about data profiles, see the Data Profiling documentation here.

Once a strategy has been developed, users will have to configure underlying attributes for use in the configuration, as well as potentially writing JavaScript for dictating the logic of the matching algorithm. Typical configuration steps include:

• Configuring the Component Model(s)
• Configuring the Setup Group
• Configuring the Match Codes
• Configuring the Matching Algorithm
• May include configuring Golden Records
• Configuring the Inbound Integration Endpoint
• Configuring the Event Processor
• Including Matching Algorithm / Match Code maintenance tasks

2019, Stibo Systems – Confidential