Skip to content

Cookbook

Entity canonicalization: make machines resolve you to one thing

A repeatable procedure to make an organization or person unambiguous to search and AI — so your namesakes stop absorbing your signal.

Published ·

Most visibility problems for a new or common-named brand aren't ranking problems — they're identity problems. A model or a retriever sees your name resolve to several different referents and can't tell which one you are, so your signal scatters across your namesakes. Canonicalization fixes that: it gives machines one resolvable entity, wired consistently across every surface. Here is the procedure we run, in order.

The procedure

  1. Fix one canonical name and its disambiguators

    Choose a single canonical form, plus the attributes that separate you from namesakes — domain, city, category, founder. Write them down; they become the signature you repeat everywhere.

  2. Claim authoritative entity records

    Create a Wikidata item backed by genuine references so it isn't rejected, and claim any authoritative registry profiles relevant to your category. These are the anchors machines trust.

  3. Wire sameAs to every authoritative profile

    In your Organization and Person schema, link the Wikidata item, official domain, and verified founder profiles with sameAs, so a machine collapses them into one identity instead of several.

  4. Enforce a first-mention convention on every page

    The first time your name appears, pair it with the disambiguators — for example, "Brand — category, city, domain". Keep the name and core facts identical across pages and off-site.

  5. Single-source the schema

    Define one Organization node and one Person node, referenced by @id across every page inside a single @graph. Duplicate nodes fragment the very signal you're consolidating.

  6. Verify there's no drift

    Visible text must match the schema — same name, same description. Validate the JSON-LD; a mismatch teaches machines to distrust the page.

  7. Corroborate off-site

    Get third-party sources to repeat the same name, disambiguators, and identifiers. Third-party mentions correlate with AI visibility roughly three times more strongly than backlinks (Agile Brand Guide, 2026-05), and they're what move the training side over time.