After the client’s files have been profiled both at the data source level and at the record level, we are now in position to start to make some data determinations for the Master Data Management system. Data determinations for a Master Data Management application typically fall under the following categories:
1. Data Reformulation
2. Native Intelligence
3. Non-reliable Data
4. Missing Data
5. Accuracy
6. SingleVision Functionality
7. Operational Data
Data Reformulation
Sometimes, we need to change the way the client has the data for matching purposes in order to get the best possible Master Data Management match. For example, the client may have Department data in an Address field. Depending on the client goals, it may be best to extract the Department Data and put that data into a Department field that we created strictly for matching purposes.
Native Intelligence
By examining the Data Profiles discussed in the previous Discovery Posts, we can determine if there is any Native Intelligence that can be used to improve the Master Data Management match results. Native Intelligence is data that is already in the client’s files that indicates that one record is somehow related to another record. Sometimes the client has created the Native Intelligence through their own manual research or other method. Other times, the Native Intelligence has been provided by purchasing or otherwise accessing information from a third-party resource.
Non-Reliable Data
Sometimes, after examining the Data Profiles, it may be determined that certain data elements are just not reliable enough and cannot be used for the Master Data Management solution. However, in some cases, it may be determined that a particular data element could be useful after some clean-up. Other times, we can use the data element if it is used in conjunction with some other criteria that restricts its use so that it is reliable for matching purposes.
Missing Data
After reviewing the Data Profiles, it may be discovered that certain data elements that we would like to use for the Master Data Management match either do not exist in the client’s files or only exist in a limited fashion. At minimum, this can prompt a discussion with the client to double check to see if they do actually have this data element. Depending on the missing data element, we could look to an outside source to provide the missing data. However, this is typically not necessary as it is our general practice to use whatever data the client has.
Accuracy
In addition to determining how to use the client’s data, we also check the data in order to determine if it’s accurate. For example, the client may have manually researched certain relationships and provided Native Intelligence indicating that certain records have a relationship. However, after reviewing the data, it may be determined that there were some mistakes made. Depending on the severity of issues with accuracy, it may be determined to remove the identified mistaken data, to revise the mistaken data, or to add additional criteria to the mistaken data so that it is accurate for matching purposes within the client’s Master Data Management solution.
SingleVision Functionality
Within our Master Data Management system, SingleVision, there is certain functionality to process data for matching purposes. The review of the data helps us determine which data processing features within SingleVision should be used and how they should be configured.
Operational Data
While we are generally looking to review the Data Profiles for matching within the Master Data Management system, sometimes we review the results for purposes beyond matching. For example, it may be discovered that there are many contact names in an address field. The client may want to pull out these contact names and then populate the operational system with the contact data in the appropriate fields.
Comments