Different external and internal data providers usually use different master data attributes, like keys, number ranges or descriptions. To ensure that you have a consistent set of master data for consuming applications and for analytics and reporting across different data providers, the data must be harmonized.
Example
A certain product is produced by the manufacturer “Taste” is called “Chili Chocolate” in-house at the manufacturer, the different retailers and syndicated data providers might have other names, such as “CHILI_CHOC_1254”, “Chocolate_Taste_Chilli”, or “03-TAS-chil”.
You use data harmonization to link all data that refers to the same object (for example, the same product or the same retailer store) to one single data record which exclusively identifies the object, no matter in which data source it appeared and independent from its description or ID that was provided. This harmonized object record contains consistent harmonized attributes. It also contains a link to all data records that were taken into account for harmonization of this record.
Data Harmonization Process
The process of data harmonization can be segmented into the following phases:
The administrator checks and adapts the Customizing settings for automatic harmonization.
The administrator defines allowed attribute values for the harmonization of attribute values on the Web user interface.
The data is uploaded. During the upload process the system maps and harmonizes the uploaded data for all object types for which automatic harmonization is turned on.
For any issues that are not resolved automatically, the system creates work items for manual processing.
The harmonization user performs the following manual steps:
Reviewing worklists and processing work items
Defining and editing mapping of source objects to harmonized objects
Editing harmonized objects in single editing or mass change mode
Defining unit conversions for source products
Mapping of source attribute values to allowed attribute values