Entity Relationships (Relationship Mapping)

Entity relationships are the labelled connections between entities: Jane Smith works for Acme, a product’s brand is Acme, an article is about a topic. Relationship mapping is the work of identifying those connections, by extracting them from text or declaring them in structured data, so that search engines and AI systems see a connected graph instead of isolated facts. MLforSEO calls it the most important concept in entity development, and the one most teams skip.

How it works

Every relationship can be written as a triple: subject, predicate, object. MLforSEO’s EAV model guide uses “Apple Inc. founded 1976” and “Apple Inc. CEO Tim Cook” as examples; a knowledge graph is built from triples like these connected through shared entities (see also tuples and RDF). Wikidata stores the same idea as item, property and value statements.

In SEO work, relationships come from two directions:

• Extracted from text. NLP tools can detect relations between entities in content. MLforSEO’s API comparison notes that IBM Watson NLU has a dedicated relations module, and LLMs can be prompted to do the same, with human review.
• Declared in markup. Schema.org properties state relationships explicitly in JSON-LD. A stable @id gives each entity one identifier, so every reference resolves to the same node.

How it’s used in SEO and marketing

MLforSEO teaches six core relationship patterns that form the backbone of a brand’s entity network:

• Person to Organisation: worksFor, founder
• Product to Organisation: brand, manufacturer
• Article to Person: author
• Article to Concept: about, mentions
• Organisation to Location: location, address
• Organisation to Audience: audience

Beyond markup, relationship maps guide content planning (which people, products and topics a hub should connect), semantic keyword research (relationships shape how topics are grouped), query fan-out coverage (a product page that declares its brand, category and audience can match more sub-queries) and audits (checking that your site, LinkedIn and Wikidata all agree on who leads the company).

Example

Acme publishes an Organization node with the @id https://acme.com/#organization. Jane Smith’s Person markup uses worksFor pointing to that @id; Acme’s Organization markup lists her as founder; every article she writes uses author pointing to her @id; and each product uses brand pointing back to Acme. Instead of fifty unconnected mentions, a crawler finds one company, one founder and a set of products and articles tied to both. MLforSEO recommends declaring important relationships from both sides like this, so they can be verified in either direction.

Limitations and things to watch

• Declared relationships need to match reality and what external sources say; a contradiction lowers confidence instead of raising it.
• Automated relation extraction is noisy, so review its output before acting on it.
• Search engines and AI providers do not publish how they weigh individual schema relationships. Bidirectional linking and consistent @ids are sound practice for clarity and verification, with no guaranteed ranking effect.
• A relationship graph decays: new leadership, products or locations need updating everywhere.

Learn it on MLforSEO

• Why Entity Relationship Mapping Is Critical for Brand Visibility
• Entity Relationship Planning Workbook
• Entity Analysis & Relationship Mapping with IBM Watson NLU in Google Colab
• What Schema.org @id Does and Why Your Entity Graph Falls Apart Without It
• AI Search & LLMs: Entity SEO and Knowledge Graph Strategies for Brands (course)