What is Industry Prediction, and what does it return?

Industry Prediction classifies a business into an industry code. It is an add-on to a business search or web presence request, or a standalone request, and it returns one prediction per business.

What it uses. The prediction draws on multiple sources, including the business’s website, public web search results, the business name, address, and registration details, and online reviews and social profiles where available. Because it does not depend only on a website, it can classify businesses with little or no web presence, though with lower confidence.

What you get back.

• A single six-digit NAICS code and its title. We return the one highest-confidence code, never a list of candidates.

• accuracy: our confidence in the prediction, from 0 to 1.

• mcc_codes and sic_codes: the merchant category codes and SIC codes that map to the predicted NAICS code, for teams that work in those systems.

• keywords: the business activities detected, for example "recruitment" or "nail salon", which are useful when the code alone is ambiguous.

• risk_level: Baselayer’s assessment of the inherent risk of that industry, as low, medium, high, or no risk.

• reasoning: a short explanation of why that code was chosen.

Why six digits. NAICS is a hierarchy: about 20 two-digit sectors, hundreds of four-digit industry groups, and over a thousand six-digit industries. Six digits is the most specific level and the hardest to predict, which is why many providers stop at two or four. Baselayer always returns the six-digit code, and you can truncate it to a coarser level yourself if your policy works at four digits.

Baselayer returns 2017 NAICS codes.

These lists are publicly available:

• NAICS: https://www.census.gov/naics/?58967?yearbck=2017

• SIC: https://www.sec.gov/search-filings/standard-industrial-classification-sic-code-list

• MCC (Mastercard): https://www.mastercard.us/content/dam/public/mastercardcom/na/global-site/documents/quick-reference-booklet-merchant.pdf