Company
Date Published
Author
Matt Sornson
Word count
300
Language
English
Hacker News points
None

Summary

Following the launch of company categories, which initially covered about 30% of companies, a new machine learning categorization system has been developed to significantly enhance accuracy and coverage. This system, which includes approximately 140 unique categories, allows the automatic application of relevant sectors, industry groups, industries, and sub-industries to any website by analyzing its text. The tags generated are then aligned with a standardized industry hierarchy akin to the GICS framework. Previously, many self-reported data sources were found to be inaccurate and lacking in coverage, but the new system has improved the categorization of private companies with fewer than 50 employees from 15% to 95%. This advancement has greatly upgraded the Enrichment API, and users can access these categories via the Discovery API to create highly targeted company lists.