Entity-based SEO means optimising for the things a search engine understands rather than the exact words people type. An entity is a distinct thing: a company, a person, a product, a place, a concept. Google stores these and the relationships between them in its Knowledge Graph, which is how it knows Apple the company is not apple the fruit.
Why it matters more now
Keyword matching has been in decline for a decade. BERT and MUM taught Google to read meaning. AI Overviews and chat assistants now assemble answers by pulling facts about entities from several sources at once, rather than handing back one page with the right phrase. If a model doesn’t recognise your business as an entity, it cannot cite you as one.
The practical consequence is that being known as the commercial roofing firm in Leeds beats ranking for a single phrase, because that recognition applies across thousands of queries you never targeted.
What makes an entity recognisable
- Consistency. Name, address and description identical everywhere: Companies House, LinkedIn, your Google Business Profile, industry directories, your own site.
- Confirmation from elsewhere. Google trusts what other sources say about you more than what you say about yourself. Press coverage, listings and reviews carry the weight here.
- Explicit markup. Organisation schema with sameAs links to your verified profiles, Person schema for named authors, Product and Service markup where they apply.
- Coverage of the surrounding topics. Being connected to the concepts around your subject, not only to the subject itself.
Applying it to your pages
Stop writing to a keyword and start writing to a topic with its related entities named inside it. A page about commercial solar should mention inverter types, MCS certification, the Smart Export Guarantee and realistic payback periods, because those are the entities attached to the subject. Covering them tells Google you understand the field, and gives a model the specifics it needs to quote you.
Do the same with authorship. A named author with a real bio, stated credentials and links to their work elsewhere is an entity Google can verify. An anonymous admin byline is not, and on anything touching money or health, that gap costs you.
Checking whether Google knows you
Search your brand name and see whether a knowledge panel appears. Then ask two or three AI assistants what your company does, and read the answers carefully. Wrong is far more common than empty, and it usually traces back to one stale directory listing that everything else copied. Cleaning that up is the first job in any AI search engagement we take on.
Common questions
Do I need a Wikipedia page?
No, and most businesses would not qualify for one. Wikipedia helps because it is well structured and widely trusted, but Google builds entity understanding from many sources. Consistent listings, press coverage and clean schema get you most of the way there.
Is this just schema markup?
Schema helps, but it’s a shortcut, not the whole job. Markup tells search engines what you claim to be. Recognition comes from independent sources saying the same thing. Markup without corroboration is an assertion nobody has confirmed.
Does entity SEO replace keyword research?
No. You still need to know what people search for and what they want when they search it. What changes is the output. Instead of one page per keyword variation, you build proper coverage of a topic and the entities sitting inside it.
How long does it take to establish an entity?
Months rather than weeks. Fixing listings and adding schema is quick work. Getting independent sources to describe you consistently takes a PR cycle or two, and knowledge panels appear on Google’s schedule rather than yours.
See our glossary entry for entity-based SEO.
