Getting a Wikidata Entry Created: What's Legitimate - and What's Not

What Wikidata is

Wikidata is a free, structured knowledge base run by the Wikimedia Foundation. Instead of prose, it contains items - each with a unique Q-number - and statements in the form of property/value, such as "occupation: author" or "inception: 2014."

Because this data is machine-readable, openly licensed, and easy to query, it gets reused by countless systems: from search engines to library catalogs to AI applications.

Why it's the anchor in the Knowledge Graph

The Google Knowledge Graph isn't identical to Wikidata and isn't copied from it either. But Wikidata provides exactly what machine entity resolution needs: a unique identifier, linked external identifiers, and documented statements in a clear structure.

External identifiers are especially valuable - ISNI, ORCID, GND, VIAF, trade register IDs. They connect your item to other authoritative databases and turn an isolated dataset into a node in a network. We describe this under Wikidata & Schema.org.

The notability rules

Wikidata is more accessible than Wikipedia, but not rule-free. An item is admissible if one of three conditions is met: there is a corresponding article on a Wikimedia project; or the entity is a clearly identifiable subject that can be substantiated by at least one serious, publicly accessible reference; or the item serves a structural function, such as enabling other statements to make sense.

In practice that means: it doesn't need encyclopedic significance, but it needs evidence beyond your own website. An item whose only source is self-description is a deletion candidate.

Why items get deleted

• No external references - only self-reported information or no source at all
• Promotional framing: statements that read more like marketing than fact
• Mass-created items from an account with no other contribution history
• Invented or unverifiable claims, such as awards that don't exist
• Clear conflicts of interest without disclosure that the community notices

Duplicates: the most common panel killer

When two items describe the same person - one from an old database import, one newly created - the evidence gets split across two datasets. Neither reaches the necessary density, and Google receives contradictory signals about who is meant.

The result is incomplete panels, incorrect attributions, or no panel at all. The solution is a merge following community rules - not deleting and recreating, since that loses references and links. That's why, before creating any new item, we always research existing ones first.

How to recognize a clean item

• Every non-trivial statement has a reference to an independent, permanently accessible source
• External identifiers are set where they exist (ISNI, ORCID, GND, register IDs)
• The official website field points to your Entity Home, not to a landing page
• Labels and descriptions are neutral, short, and factual - no superlatives
• Exactly one item exists for this entity, checked across name variants
• Links to related items (employer, works, places of education) are set and also documented

What legitimate service looks like - and what doesn't

Legitimate: researching existing items before creating new ones, backing every statement with evidence, disclosing conflicts of interest per Wikimedia rules, avoiding claims that can't be substantiated, and honestly declining when no external source exists.

Not legitimate: mass-creating items without evidence, selling an item as a "Knowledge Panel," promising permanent control over the content - Wikidata is an open project, anyone can change statements - and concealing paid work.

For panel impact: a single item rarely carries enough weight on its own. It needs to be combined with Entity Home, specialist databases, and press, as described in Knowledge Panel Without Wikipedia.

How we work

We first check whether verifiable external sources exist. If not, we decline - an unsourced item does more harm than good. If they do exist, we research existing items, clean up duplicates, and build the statements with proper documentation.

Wikidata is part of the Complete Package at Brendit (4,900 EUR total, in two success-based installments of 1,500 EUR and 3,400 EUR) and part of the Basic Panel (1,500 EUR). Here too, payment is due only once results are visible.