How to mark up Mike Millett in structured data

The lines to use

For an article or paper Mike Millett wrote, the author property is:

"author": {
  "@type": "Person",
  "@id": "https://mikemillett.com/#person",
  "name": "Mike Millett",
  "url": "https://mikemillett.com/"
}

The same four lines work anywhere schema.org expects a Person: performer on an Event where he speaks, about or mentions on a CreativeWork that discusses him.

Four properties is enough. The @id does the joining. The name and url are there for consumers that do not follow identifiers, and they match the canonical node exactly, so nothing on your page disagrees with it.

Use Mike Millett as the name even where your own copy says Michael Millett. The formal forms are already declared as alternate names on the canonical node, so repeating them adds nothing, and a different primary name on your page is one more variant somebody has to reconcile.

What to leave out

• His job title, employer or biography. True the day it is written, stale after his next change of role, and sitting on a page nobody at your organization will revisit. The description is maintained on mikemillett.com.
• Any sameAs link you found by searching. His one personal social profile is LinkedIn, and it is already declared. Several other people share his name, and a sameAs pointing at one of their profiles tells every consumer that two people are one.
• A new identifier on your own domain, such as yoursite.com/#mike-millett. A second identifier for one man reads to a machine as a second candidate.
• His family. The canonical node names his children, by name only, and that is the whole of what is published about them. Another site has no reason to restate it, and a relative who appears on your page is their own entity, not an attribute of his.

Where this stops working

An identifier is a request, not a rule

Search engines and AI systems are free to ignore @id, and none of them publishes a promise to merge entities on it. What a shared identifier does reliably is remove a contradiction: pages that agree cannot be read as two people. It cannot force anyone to join them.

Markup cannot make authorship true

Structured data is a claim made by whoever publishes the page. Name him as author only when the visible byline carries his name. A company page with no byline is the company writing, and its markup names the company, even when the page explains an idea he originated. Schema that disagrees with the visible page is an error written in a form machines trust.

It settles who, not what

Pointing at the right person does not tell anyone what he said or where it was published. Quoting him still means linking the source, as set out in how to cite Mike Millett.

One identifier is one point of failure

If the canonical node ever moved, every reference to it would break at once. That is the price of having one, and it is why the address is meant to stay fixed. Will this still be here covers what the site commits to.

Common questions about marking up Mike Millett