What Is a Knowledge Graph and Why Your Business Isn't In One

Let me tell you what Google's Knowledge Graph actually is, because most of what you've read about it is either wrong or incomplete.

It is not a reward for writing good content. It is not a prize for having lots of backlinks. It is not something your SEO agency can "submit you to" for a fee, though many will pretend otherwise.

The Knowledge Graph is a database. Specifically, it's a structured database of verified entities: people, places, organizations, and things that Google has determined to actually exist in the world. The key word is verified. Not ranked. Verified.

This distinction changes everything.

What an Entity Actually Means

In Google's framework, an entity is anything that has a distinct, independent existence that can be unambiguously identified. Ibrahim Anwar the person is an entity. Witanabe, the company I founded, is an entity. The pump brand ALBIN, whose distribution rights I hold for Indonesia starting January 2026, is an entity.

A Facebook page for a business called "Toko Sembako Pak Budi" is not an entity in this sense. Not because it doesn't exist. It does. But Google cannot verify it independently, cannot confirm it's the same "Toko Sembako Pak Budi" referenced elsewhere, and cannot distinguish it from the seventeen other businesses with similar names.

An entity requires consistent, machine-readable identity across multiple independently maintained sources. Without that, you're just another string of characters on a webpage.

The Graph Part

The "graph" in Knowledge Graph refers to the relationships between entities. Not just "Ibrahim Anwar exists" but "Ibrahim Anwar is Director of Witanabe, which is an Organization, located in Bogor, which is a city in West Java, which is a province in Indonesia."

These relationships are what make the Knowledge Graph useful. When someone searches for "pump supplier West Java" and you're a verified entity in the graph with a confirmed relationship to that geography and industry, you surface. When you're not in the graph at all, you don't surface regardless of how many keywords you've stuffed into your landing page.

This is why traditional SEO thinking breaks down here. You cannot keyword your way into the Knowledge Graph. You have to build verifiable entity relationships.

How Entities Get Into the Graph

Google's process for adding entities is not transparent and not entirely predictable. But the signals are well-documented enough to work with.

First, consistent structured data on your own domain. JSON-LD schema, specifically the type declarations that tell Google "this page is about a Person" or "this page is about an Organization." With properties filled in correctly: name, URL, founding date, address, description. And critically, a sameAs array that links to your external profiles.

Second, external profile verification. Your LinkedIn, your Google Business Profile, your ORCID if you're an individual professional, your trademark registration, your mentions in institutional contexts. Each of these is a node. The links between them, pointing consistently to the same identity, form the graph.

Third, third-party corroboration. This is the hardest to manufacture and the most important. When EFEO Paris references your work independently, when KPK documentation includes your name, when a government program like Wall of Talents from West Java publicly recognizes you, that's corroboration you didn't create yourself. Google weights this heavily precisely because it can't be faked easily.

Why Indonesian Businesses Are Almost Entirely Absent

I've spent significant time looking at this problem specifically for Indonesian businesses, because I run three of them and I needed to solve it for myself before I could offer it to anyone else.

The Indonesian business landscape has a structural problem. Most businesses built their digital presence inside closed platforms: Tokopedia, Shopee, Instagram, WhatsApp Business. These platforms are effective for transactional commerce. They're useless for entity verification.

When I'm in Tokopedia, I'm "Toko Witanabe" or some variation. When I'm on Instagram, I'm an account handle. When I'm on WhatsApp Business, I'm a phone number. None of these create the kind of open, machine-readable, linkable identity that Google can incorporate into its Knowledge Graph.

Meanwhile, my competitors in Europe and North America are building on their own domains, filing for ORCID identifiers, registering their works with ISBN, creating structured presence that accumulates verification signals over time.

This is not a technology gap. Indonesia has the developers, the designers, the infrastructure. It's a strategy gap. Most Indonesian businesses don't know the game exists, let alone that they're losing it.

What It Looks Like When It Works

My own situation is instructive. I have a trademark registration (IDM001337019). I have an ORCID. I have a publishing catalog of 558 titles across 5 languages. I have institutional client relationships that are documented in public records. I have a domain with proper JSON-LD schema. I have three companies, each with their own verified presence, all linking to a consistent identity.

The result is that "Ibrahim Anwar" as an entity is becoming increasingly verifiable. Not famous. Not viral. Verifiable. There's a difference, and the difference matters for the specific goal of appearing in AI search results and eventually the Knowledge Panel.

For Arsindo, the industrial services company, I have procurement records. Business permits. A registered address. A founding date. All of these feed into the entity graph. The company exists not just as a claim on a website but as a verifiable fact in multiple independent systems.

For Witanabe, I have 60+ documented projects. A blog with six pages of case studies. Real photos from real sites. Clients who can be found and asked. This kind of documented work history is exactly what the Knowledge Graph uses to characterize what an organization actually does.

The Practical Implication

Most articles about the Knowledge Graph end with a vague suggestion to "optimize your Google Business Profile" and call it a day. That's not wrong, but it's about 10% of the answer.

The full picture requires building what I call entity infrastructure: a system where your identity is declared, verified, and cross-referenced across enough independent sources that a machine can confirm you exist, do what you say you do, and are who you say you are.

The businesses that build this infrastructure now are establishing a compound advantage. Every new piece of verified content, every new cross-reference, every new institutional mention adds to the graph. The graph is sticky. Once you're in it, it's hard to fall out. And once you're not in it, it's increasingly hard to compete with those who are.

The Indonesian businesses doing enterprise-level work, the ones capable of serving multinational clients, the ones with the talent and the track record, deserve to be visible at that level. Most of them aren't, not because of capability, but because of infrastructure.

That's the gap. That's what's worth closing.