Reltio's Master Data Management (MDM) platform enhances the enterprise's ability to organize and analyze data. In the first two parts of this series, we covered some of the ways that Reltio's MDM approach generates value for enterprises that need to consolidate records and establish a single point of truth for customer data.
There are a number of important challenges that Reltio technology needs to address in order to achieve this goal. One of the most important parts of this solution concerns how the system recognizes similarities in profiles, and how users can consolidate similar data in ways that generate value.
One of the fundamental things Reltio does is identify when two records contain similar data. Those records can represent individuals, organizations, products, services, employees, or anything else. Reltio looks deeper into the data the records contain and looks for correlations between them.
It does this through matching logic. Defining matching logic in a simple but robust way is an important part of how Reltio's MDM solution is different.
There are multiple ways to configure matching logic. The first and most common is called the rules-based approach. This framework relies on specific policies that reach towards a configurable threshold. Once two records have enough data in common to pass this threshold, they are correlated and merged together. This can happen automatically or with a manual review process.
Reltio also provides users with an additional option called Match IQ. This product uses machine learning to analyze data, look for correlations, and identify records that have some degree of similarity between them.
Match IQ is a premium service from Reltio that works best when given access to a sufficiently large volume of data to learn from. We recommend you start building out the model in a test environment with around 10,000 records in order to give the machine learning algorithm a decent starting point.
Whether using Match IQ or a traditional rules-based approach, you can establish different levels of review for records with different similarity thresholds and keep track of which profiles consolidated based on their similarity. These methods go hand-in-hand to deliver a golden record which is the result of these merge decisions.
You may find duplicate records within a single system, like SAP or Salesforce, or you may find similar records distributed across different systems that need to correlate and consolidate into a single golden record. Once consolidated, the best version can then be published to accurately reflect the profile in each originating system. Reltio achieves these tasks using customer controlled logic.
Reltio allows users to combine rule-based matching and Match IQ machine learning for matching records that are not loaded into our repository against records which are in the repository. It is possible to deploy both, compare the recommendations each process makes independently, and process records based on the best result. This external match service allows users to streamline operational processes that span multiple systems - effectively allowing you to search Reltio before you create new records.
One of the useful aspects of this approach is that you can gain the benefits of rules-based and machine learning data consolidation without ever having to load data into the system. You can establish diverse match rules and individually select which logic will be applied in each matching iteration based on your configuration settings and the results you're looking for.
While looking at a single profile, Reltio will suggest pending potential matches within the system. You can use different matching types to find records in close proximity with the one you're currently looking at. Reltio will provide a list of records that may be potential matches and allows you to drill into each of them and verify whether they really are.
Reltio supports side-by-side profile comparison, and even allows users to add additional fields when conducting analysis. Since the system tracks this lineage, you can easily see combined records and conduct audits to make sure information has not been lost.
When merging two or more records together, the surviving entity ID is usually determined by the oldest entity ID created in the merge set. Reltio does this automatically in order to establish a clear and coherent set of identifying metadata for your records.
Customers who want to have their own control over how Reltio creates and assigns universal identifiers can define their own IDs. The system does not force users to adopt Reltio's auto-generated IDs, which makes it easy for organizations that need to maintain an existing structure when assigning identifiable tags to data. You can define your own sequence identifiers and allow Reltio to follow the pattern your organization needs.
Related Content: Data Governance with Reltio: How to Improve the Value of Data as an AssetData Governance Part II: Role-Based Workflow and Collaboration Tools