According to the Gartner IT Glossary, entity resolution (also known as matching) is “the capability to resolve multiple labels, products, or other noun classes of data into a single resolved entity.” Creating this resolved entity across disparate sets of data is an extremely difficult process, especially when faced with thousands to millions of records, but is at the core of creating reliable, useful master data.
Entity resolution is especially important for companies seeking to create a master data program in their organization, as it can be used to create a single version of the truth for any given business entity (i.e. customer, supplier, prospect) within an organization’s operational systems. This single version of truth allows the data to be used across multiple teams and to accomplish multiple goals from growth to risk mitigation. Some companies accomplish this using Master Data Management (MDM) software, while others have their own homegrown methods.
Leveraging Dun & Bradstreet as a strategic partner within your master data program can simplify and improve entity resolution for your company. Because Dun & Bradstreet can be used as a reliable third-party source and offers matching capabilities that span all current and past iterations of companies within its global database, the organization can help significantly optimize your ability to create a single version of truth, which proves crucial to any program’s success.
Match(ing) Made in Heaven
All MDM software solutions provide the ability to decide with varying degrees of confidence when separate records should be consolidated into one master record. It’s a simple concept, but very hard to do. However, without effective entity resolution, MDM software would be little more than a proprietary database with workflow and Extract, Transform, and Load (ETL) tools. So, if most MDM software already supports entity resolution, why would you need a reliable third-party to complement your system?
- MDM Matching
Most MDM software solutions rely on clustering and inference algorithms to produce ‘fuzzy’ matches. Producing these potential matches requires all possible combinations of companies/records in your database to be matched to each other. As you add more companies to your data set, the combinations and the computing power needed to calculate them will grow exponentially. Even if you successfully match two or more records into a single record, your MDM software will be unable to tell you if that unified record represents an actual operating business entity.
- Dun & Bradstreet Reference Data
Dun & Bradstreet’s pre-mastered commercial data is coded, semantically stable, and relatively static, and can be trusted as an accurate, complete, and consistent reference data set. You can match your entities against a single candidate in a known universe of more than 300 million companies, all of which are known to be unique and are identified by the Dun & Bradstreet D-U-N-S® Number. When dealing with thousands to millions of records, the ability to match your data against a known reference drastically increases the level of reliability of the output and dramatically reduces the number of instances that must be manually vetted by your data governance team.
Once your business entity records are resolved, the final record becomes your master and your teams can rely on it to create data hierarches, make data-driven decisions about growth and risk, and look for efficiencies within customer and supplier relationships. To learn more about entity resolution and how it can benefit your organization, download the whitepaper by our Master Data Distinguished Architect, Malcolm Hawker: “Fundamentals of Entity Resolution.”