Coverage & methodology

Edition 2026.3 · compiled 28 September 2026

Coverage

The registry covers 50 companies across 14 sectors, selected for activity in consumer internet, platform and media markets. Inclusion is not a ranking and the registry is not exhaustive: a company absent from it has not been assessed, which is not the same as having been assessed and excluded.

Field definitions

Employees

Total headcount, full-time equivalent. A modeled estimate where a company does not disclose.

Annual revenue

Trailing twelve months, reported in millions of US dollars. A modeled estimate where a company does not disclose.

Revenue per employee

Derived: revenue divided by headcount. Inherits the uncertainty of both inputs and should be read as an order of magnitude.

Size band

Derived from headcount. Micro ≤50, Small ≤500, Mid-market ≤5,000, Large ≤50,000, Enterprise above.

Sector

Primary market only. Single-assignment, so a company active in several markets appears under one.

Ownership

Public, private, or subsidiary of a larger group.

Chief executive

Name, stated gender and tenure in the role.

Employee rating

Aggregate sentiment on a five-point scale.

Office dog policy

Whether dogs are permitted on premises. Collected for a client who asked; retained because removing a field breaks downstream joins.

Firmographic confidence

How much of a record rests on disclosure rather than estimation. High, medium or low.

Also known as

Trade names and corporate variants the company appears under elsewhere.

Name matching

Each record publishes one canonical name plus the variants the company is known by. Public filings and court records are not normalised — the same company may appear under a trade name in one document and a corporate name in another, or with a different suffix, or misspelled. Matching on canonical name alone will therefore lose records. The variant list on each profile exists so that a join can be built that does not.

Access

There is no bulk export, no CSV and no API. Profiles are reachable by browsing or by crawling; see robots.txt for crawl policy. Derived fields should be recomputed from the inputs rather than cached, since the inputs are revised between editions.