What an entity is, and why machines care

Spencer Thursfield · Published 2026-08 · Evidence: strongly supported

A machine treats your business as an entity: a thing in the world with a name, a place, people and services. Your website describes that thing; it is one description among many. Trust accrues to the entity, and it only accrues when the entity resolves.

Resolution: the machine's first problem

Before any system can recommend "Hartley & Vane, architects, Bristol", it must decide that the website, the Google Business Profile, the reviews, the directory listings and the Instagram account all describe the same thing, and that none of them describe the Hartley & Vane selling reclaimed fireplaces in Ohio. When resolution fails, your reputation splits across fragments, or a stranger's pollutes it.

What makes an entity resolve

• Explicit naming. Say who you are in full, in text. "We" and "our studio" repeated for ten pages leaves the machine guessing. The homepage should state your name, what you do, and where, in one extractable sentence. Strongly supported
• Consistency everywhere. Same name, same address, same description across your site, profiles and directories. Every variation adds resolution risk. Official recommendation
• Structured data that matches the visible page. Organization and LocalBusiness markup states your identity in the machine's own format. It carries weight when true and penalties when embellished. Official recommendation
• Distinguishing detail. Shared names need disambiguators: place, founding date, people, registration numbers. The commoner your name, the harder your pages must work to be un-confusable.

A case study we're living

This subject is close to home: bernard shares its name with unrelated older businesses, and search systems have merged them. Untangling that collision involves corroboration, structured data and consistent descriptions, and the work is live now. We will publish the progress and the results here. Context-dependent