Wikidata is the underrated building block in entity SEO. While everyone talks about Wikipedia, Wikidata supplies the structured side: machine-readable statements reused by Wikipedia, search engines and knowledge services. Its link to Google is documented historically: from 2014, Google handed the content of its Freebase database over to Wikidata (Pellissier Tanon et al., 2016). According to Google's help center, though, Wikidata is not a required source for Knowledge Panels. A maintained Wikidata item is often more valuable for entity SEO than a weak Wikipedia stub that Google does not weight as an authority signal.

What is a Wikidata item, and when may you create one?

A Wikidata item is a record in the free knowledge base Wikidata, identified by a unique Q-ID such as Q2283 for Microsoft or Q42 for Douglas Adams. It holds labels in many languages, descriptions, aliases and statements made of property, value and reference. You should only create one if the entity meets Wikidata's notability rules.

Under Wikidata:Notability, one of three criteria is enough: a valid sitelink to a Wikimedia project such as Wikipedia, a clearly identifiable entity that can be described using serious and publicly available references, or a structural need within Wikidata. The community decides disputed cases, and Wikidata says the rules are intentionally left a bit vague. The essay Wikidata:Self-promotion strongly discourages creating items about yourself or your own organization and suggests three independent, reputable sources as orientation. Items without that evidence end up in the requests for deletion. Paid contributions must be disclosed under the Wikimedia Terms of Use.

The Q-ID is the reference that sameAs links and @id graphs point to. A statement consists of a property (P-code) + value + references. Example: P31 (instance of) + Q4830453 (business) + reference to a commercial register entry.

The quality of a Wikidata item is directly visible in the density and quality of its statements. An item with five properties and no references is weak. An item with thirty properties, all referenced to independent sources, is strong. In our assessment, which Google does not document in this form, that density supports confidence in the Knowledge Graph and entity resolution in LLMs.

Q-ID

stable identifier — foundation for the @id graph and sameAs

20+

properties as the target for a strong item (rule of thumb from our practice)

References

required per statement — without refs, a weak signal

The property set for companies

Wikidata does not mandate any properties; "required" here means our minimum set. Required properties: P31 (instance of → business / company / startup), P17 (country), P571 (inception date), P856 (official website), P1454 (legal form).

Strong properties: P112 (founded by — reference to a founder item), P169 (chief executive officer), P452 (industry), P361 (part of — for group structures), P159 (headquarters location), P1128 (employees).

Identifiers: P1320 (OpenCorporates), P1278 (LEI code), P2088 (Crunchbase organization ID), P4264 (LinkedIn company ID), P2013 (Facebook username), P2002 (X username), P2003 (Instagram), P2397 (YouTube channel ID). In our experience, this identifier chain is one of the strongest levers for entity resolution in LLMs.

The property set for people

Required: P31 (human), P106 (occupation), P569 (date of birth), P27 (country of citizenship), P21 (sex or gender).

Strong properties: P108 (employer), P69 (educated at), P463 (member of), P1416 (affiliation), P800 (notable work), P166 (awards).

Identifiers: P496 (ORCID iD), P6634 (LinkedIn personal profile ID), P214 (VIAF), P227 (GND), P244 (LoC ID), P2037 (GitHub), P2003 (Instagram), P2002 (X).

Wikidata property matrix for companies and people (property IDs checked in September 2026; priorities are our recommendation)
CategoryCompany (property)Person (property)Priority
Instance typeP31 instance of → businessP31 instance of → humanRequired
Location / originP17 country, P159 HQP27 citizenship, P19 place of birthRequired
Occupation / industryP452 industryP106 occupationRequired
DateP571 inceptionP569 date of birthRequired
LeadershipP112 founded by, P169 CEOP108 employerRecommended
Official websiteP856 official websiteP856 official websiteRequired
LinkedInP4264 company IDP6634 personal profile IDRecommended
Academia / ORCID—P496 ORCID iDYMYL-required
Identifier codeP1278 LEI, P2088 CrunchbaseP214 VIAFStrengthening
Education—P69 educated atRecommended
Mid-read · Wikidata status

Is your Wikidata item strong enough?

30 minutes of live analysis: property count, reference density, identifier-chain quality. A clear priority list for what to add within the next 30 days.

Wikidata check →

Which sources does Wikidata accept as references?

Wikidata requires a source for most statements. According to Help:Sources, the two main reference properties are stated in (P248) for publications and databases and reference URL (P854) for websites. Exceptions apply to common knowledge and to statements that can be verified through external identifiers.

A statement supported only by imported from Wikimedia project (P143) is explicitly not considered sourced. Primary, secondary and tertiary sources are all allowed in principle. Your own website therefore works for self-descriptions such as the official website. For facts like revenue, headcount or founding date, independent sources should dominate, because they also support notability.

Maintaining Wikidata in steady state

Wikidata items are never "finished". A quarterly review across four dimensions: (1) add new publications as references, (2) reflect personnel and organizational changes, (3) update external identifiers (LinkedIn URL changes, new Crunchbase profiles), (4) drift control — identify and correct incorrect edits by third parties. Tool recommendation: a watchlist plus Wikimedia Toolforge utilities for structured reviews. If you have a personal interest in the item, propose disputed changes on the talk page instead of editing back and forth yourself.

Conclusion: Wikidata is the lever beneath Wikipedia

For many brands, Wikidata is a more sensible structured step than Wikipedia, provided independent sources document notability. Without that evidence, deletion looms, and a deleted item helps no one. Brands that maintain a legitimate item with 20+ cleanly referenced properties hold a knowledge-graph asset that supports all three entity-SEO dimensions: identifiability, consistency and corroboration.

Sources