Introduction: Google Has Already Decided Who You Are, and Chapter 1: What Is a KGMID?

Somewhere in Google's Knowledge Graph, there is very likely a record about your business, and you have never seen it.

It was assembled without your permission from whatever evidence Google could gather: your website, old directory listings, third-party mentions, structured databases, reviews. It holds what Google currently believes you are, what you do, where you operate and who you are connected to. Some of it will be right. Some of it will be wrong. Some of it will simply be missing.

That record has an identifier, a machine ID called a KGMID, and it matters more than most marketers realise.

When Google shows a knowledge panel, it is displaying that record. When Google's systems decide which entities they can confidently surface, describe and recommend, they are consulting that record. And as AI search grows, the confidence attached to that record increasingly decides whether your business appears in answers at all.

Here is the part that should focus your attention. The record gets built either way. An entity that never participates does not stay blank. It stays vague, assembled from fragments, contradictions and whatever third parties happened to publish. Vague entities do not earn panels, and machines built to avoid recommending things they cannot confidently resolve simply recommend a competitor instead.

This book teaches you to participate. What a KGMID is and how to find yours. How the Knowledge Graph builds its record and what feeds it. How to establish your side of the story, earn the corroboration that hardens it, and cross the threshold where a panel appears. How to claim it, manage it, fix it when it breaks, and measure the confidence machines hold in you month by month.

One of this book's authors is running that exact campaign in public right now, and the case study near the end documents it honestly, including the parts that are slow.

Google has already started defining you. Time to take the pen.

A KGMID is a Knowledge Graph Machine ID, the unique identifier Google assigns to every entity in its Knowledge Graph.

It is a short string, typically beginning with /m/ for older entities inherited from the Freebase database or /g/ for entities Google has created since. The string itself is meaningless to humans. To Google's systems, it is the entity: the single key that every fact, attribute, relationship and panel connects back to.

The ID matters more than the name

Names are ambiguous and machines hate ambiguity. There are thousands of businesses called Apex, hundreds of people called James Smith, and dozens of companies sharing your trading name somewhere on earth.

The KGMID exists to solve exactly that problem. When Google resolves a mention of your business, it is not really matching your name. It is deciding which machine ID the mention belongs to. Every ranking consideration, every panel display and every AI-era confidence judgement about your entity happens at the level of that ID, not at the level of the words.

This is why the same name can produce a rich panel for one business and nothing for another. One name has a strong, well-fed ID behind it. The other has a weak one, or none, or three fragments competing.

What having a KGMID does and does not mean

Having an ID means Google recognises you as a thing: a distinct entity with at least some attributes. It does not mean Google trusts you, describes you correctly or will show a panel. Plenty of entities hold IDs with thin, wrong or contradictory records attached.

Not having an ID means you exist to Google only as words on pages, which in the entity era is barely existing at all.

The rest of this book is the journey between those states: from no ID, to an ID, to a correct record, to a confident record, to a panel. The string is small. Everything rides on it.