Data sources / Commercial

Commercial company databases

Vendor-maintained company profiles, lists and search products. Coverage and field definitions depend on the provider; reported values, modeled estimates and imported public records may coexist within a single company profile.

Last reviewed 14 October 2026 · report an error

Publisher: Various
Access: Mixed
Coverage: Vendor-maintained company profiles, lists and search products. Coverage and field definitions depend on the provider; reported values, modeled estimates and imported public records may coexist within a single company profile.
Identifiers: Vendor record identifier; Domain; Legal-entity identifiers where supplied
Freshness: varies
Official URL: Varies by publisher or jurisdiction; confirm the originating source for each record.

What it can establish

Use Commercial company databases for a defined acquisition-research question. Match each record to the right entity using its identifiers (Vendor record identifier, Domain, Legal-entity identifiers where supplied) before accepting its contents, and keep Various's own scope and attribution on the finding.

That a provider publishes a particular company record or field, subject to its stated definitions, provenance and available observation dates.
A candidate-generation path for research questions that still need identity resolution and field-level supporting evidence.

What it cannot establish

A complete acquisition universe, verified target revenue, independent ownership or present operating status merely because a filter returns a match.
That estimated values are reported financials or that multiple vendors constitute independent confirmation when they share an upstream source.

A search that returns nothing in Commercial company databases describes the search performed. Coverage here is vendor-maintained company profiles, lists and search products, so a missing record becomes a negative finding only where that coverage reaches the company in question and the evidence standard supports the inference. Otherwise the question stays open.

How Docket uses it

Your acquisition criteria become the agents' questions. Commercial company databases can support some of them and leave others unanswered; its freshness (varies) is weighed answer by answer and it receives no blanket reliability score.

Triage. Resolves a candidate profile to the correct operating business and checks duplicates before detailed research.
Scout. Reads public pages without logging in, retains attributed fields and seeks primary support for consequential screening answers.
Audit. Checks retained provenance and estimate labels, striking unsupported precision and circular corroboration.

Scout records typed answers from a fixed list with verbatim quotations from retained pages. Audit reviews only the retained material, with every fetch tool denied. It settles supported contradictions, strikes unquotable answers and leaves gaps visible. See the research-agent method and screening criteria guide for the surrounding workflow.

Pitfalls and review questions

Provider terms govern access and reuse. Public visibility alone is not a blanket permission for automated collection; restricted interfaces are not bypassed.
Coverage can vary by size and geography. Treat an absent company as a coverage gap to investigate, and keep estimated fields visibly distinct from disclosed facts.

Terms of service and applicable privacy obligations govern collection and reuse. Docket reads only public pages without logging in; public visibility does not override a publisher's restrictions. Review our trust and data practices. Research availability also does not authorize contact: outreach is optional and requires client approval of the messaging and each target before contact.

What vendors compile

Commercial company databases combine registry data, filings, web data, surveys and modelled estimates to offer records on large numbers of private companies: names, locations, industry codes, estimated employees and revenue, ownership indicators and contacts. They save search time and help build first-pass universes. Some are built around a company identifier that the vendor assigns, such as a nine-digit number, and describe their own identifier on their site.

Their accuracy has been tested. A Federal Reserve study compared a widely used establishment database with government data and found that it does not cover the whole employer universe and that the largest differences were among small establishments, where imputation is prevalent, while concluding that with restrictions the data can mimic official sources with reasonable precision. An academic paper on constructing firm-level samples from a commercial database describes the work needed to make such data representative. The lesson: databases are valuable and need validation.

Using them honestly

Record the vendor, the extract date and whether each field is observed, reported or modelled. Estimated revenue and headcount are modelled for many private firms and should not be presented as facts. Where a field drives a decision, verify it from a primary source or label it as a vendor estimate. Research reports from data providers describe market-level trends and are also the vendors' own analyses.

Check licence terms, which often restrict redistribution and storage, and respect them. Overlap among vendors is partial, so using two databases can raise coverage, and comparing them is a way to estimate completeness. Where they disagree on a field, record the disagreement and investigate; do not average them.

Illustration (fictional)

A researcher pulls 600 records from a vendor for a trade in one region. After entity resolution against registries, 540 are distinct operating companies; 45 are branches or duplicates, and 15 are dormant. The vendor's revenue estimates are labelled as modelled. The researcher compares the count of firms by size class with government statistics, finds the vendor over-represents larger firms, and states that limit in the coverage note.

Filings and registries

State business registries

Legal existence, registered name, entity identifier and the particular status or filing information made public by that jurisdiction.

Company-published

Company websites

What the identified company publicly says about its services, footprint, customers or ownership on a retained page.

Company-published

Professional profiles

That a publicly visible profile describes a role, employer relationship, career history or claimed expertise at the observed time.

Sources and further reading

An Assessment of the National Establishment Time Series (NETS) Database, Board of Governors of the Federal Reserve System (2017). An independent test of a widely used commercial establishment database against government data, with its accuracy findings.
How to Construct Nationally Representative Firm Level Data from the Orbis Global Database, National Bureau of Economic Research (2015). A careful account of what commercial company databases are missing and how their coverage skews by size and country. Read before trusting any vendor's universe count.
D-U-N-S Number, Dun & Bradstreet. The vendor's own description of its company identifier.
PitchBook research and reports, PitchBook Data. Quarterly US PE breakdowns, including add-on share of deal count.
Statistics of U.S. Businesses (SUSB), U.S. Census Bureau. Firm counts and employment by enterprise size, which is how you size the population of buyable companies.

These links identify research contexts to investigate, rather than conclusions about sector attractiveness. Ask how Commercial company databases would be used on your list or return to the complete source directory.

Send us ten targets. We research them for free.

A target list and your acquisition criteria are enough. You get back company records with scores, the evidence behind every field, and the questions the public record could not answer.

Start an enquiry →Open the platform