Enigma DocsAPI ReferenceAttribute DictionaryRelease Notes
Back to All

Release Notes 11/20/2023

5 months ago by Charles Zhu
  • Growth use cases:
    • Enigma descriptions: Enigma Descriptions consist of human-readable descriptions of the industry the business operates in and a level of granularity typically beyond the NAICS 6-digit classification.
    • Industries: Alongside the creation of Enigma Descriptions, more than 55% of all operating businesses now have a 6-digit NAICS code as of November, up dfrom 37% in July.
    • Operating status: Both new businesses and refreshing existing businesses's operating status have been updated, with an automated pipeline to refresh businesses every ~6 months. In November, we removed ~2M closed businesses and added about 3M newly identified operating businesses, and increased the precision of operating businesses from 70 to 80%.
  • KYB use cases
    • High risk activities: As part of the KYB launch in October, KYB customers will be able to see high risk activities like cannabis, tobacco and vaping, gambling and sports betting, crypto.
    • KYB endpoint: As part of the KYB launch in October, we created a dedicated KYB endpoint that enables matching and retrieval of key attributes with one query.
    • Improved registration filings: KYB customers can now retrieve many more corporate registrations and state records. During testing, registration fill rate on the new KYB endpoint compared to the old increased ~1.7x, from 40% to 70%.
    • OFAC screening: KYB customers can screen inputted Ultimate Beneficial Owners (UBOs) as well as any people retrieved from registration filings against the OFAC sanctions lists.
  • Entity resolution and data asset
    • Third party services removal: The HQs of third party platforms (e.g. Doordash) have been removed as operating location addresses for individual businesses.
    • Mailing address linking: We have fixed a bug in which many distinct operating locations that share a mailing address on a corporate registration record are incorrectly linked together.
    • Projection factors: We have corrected for geographic, card type, industry, income, and size of purchases bias in our panel every individual business location.
    • Brands model and heuristics v0: We have a new entity resolution model for the Businesses/Brands data asset - while previously, Businesses/Brands relied only on a common URL to group different locations together, the new model takes names, addresses, websites, and all firmographic information into account.