Emsi Burning Glass Company Names & Standardization

In addition to our implementation of a new in-house dataset, we will also be deploying a new classifier to better allow us to properly identify and classify companies. Our company names will be more unified, focusing on what businesses are known as to the general public, rather than their legal names. For example, in the previous system, the company name was Wal-Mart, Inc., whereas in the new system, this company would be Walmart.

With this new classifier we also gain the ability to quickly correct mis-codes such as NAICS mapping, staffing/job board classification, or address misspellings, add missing companies, or standardize duplicate companies within our data.  This classifier is also global in scale allowing us to be able to increase the precision of company classification by up to 15% from the previous dataset, bringing the accuracy to between 95 and 99% depending on the geography.

Customers can expect to see some changes within the product, however in most cases the changes will be minimal such as the example listed above (Wal-Mart, Inc becomes Walmart). In a small number of cases there may be more significant change due to previous mis-classifications now being correctly parsed to the appropriate company name.  Company names will display the same way in both postings and profiles.  

In the case that a company which was previously in the dataset is no longer appearing, a customer or employee can provide feedback so that our team can review adding the company back in.  This can take 1 to 4 weeks to appear in the data depending on when the feedback is received in our release cycle. Once the new dataset is deployed the previous company name version will no longer be visible in the product.