Entities and semantics
Knowledge graph
A knowledge graph is a structured store of entities and the relationships between them, used to answer questions about things rather than to match documents.
What is knowledge graph?
A knowledge graph is a structured store of entities and the relationships between them, used to answer questions about things rather than to match documents. The knowledge panel shown beside brand searches is one visible output of it.
What turns on it
Presence in it is what lets a system state facts about a brand confidently rather than guessing from page text.
How it gets implemented
Built from corroborating sources, so influence comes from making those sources exist and agree: your own markup, the sector’s authoritative databases, and the third-party records your category relies on.
The common misunderstanding
A knowledge panel cannot be requested or created directly. It appears when corroboration is sufficient, and the feedback mechanism on it is slower and less reliable than correcting the sources.
Worth knowing because it comes up in almost every conversation about this.
Where this comes up in the work
It is part of Agentic Commerce, and more specifically of Product feed readiness. Across the rest of the work it shows up in:
- Digital PR and Link Building, which cannot be done properly without it
- Google Ads Management, where getting it wrong shows up in reading the search terms weekly rather than quarterly
- Paid Social and Meta Ads: it is the reason designing for a muted feed, where most impressions happen is scoped the way it is
- Conversion Rate Optimization, where it is the difference between fixing the step where intent was highest and the loss is largest working and appearing to
What to check
Two questions that establish whether this is actually in order on a given site:
- Whether it is consistent across templates, or correct only on the page somebody checked
- Whether the measurement that would expose a problem is even being collected
More in entities and semantics
Common questions
How do you prove knowledge graph is worth fixing?
By measuring it rather than by argument. Built from corroborating sources, so influence comes from making those sources exist and agree: your own markup, the sector’s authoritative databases, and the third-party records your category relies on.
Is knowledge graph just another name for something?
A knowledge panel cannot be requested or created directly. It appears when corroboration is sufficient, and the feedback mechanism on it is slower and less reliable than correcting the sources. Worth clarifying before agreeing to work on it.
Is knowledge graph a ranking factor?
Presence in it is what lets a system state facts about a brand confidently rather than guessing from page text. The honest framing is that it removes a constraint rather than adding a boost.
What tells you whether knowledge graph is in order?
Built from corroborating sources, so influence comes from making those sources exist and agree: your own markup, the sector’s authoritative databases, and the third-party records your category relies on. The check is cheap; the fix is where the work sits. If the brief is still open, Google Play optimization and iOS App Store optimization are the two that change it most often.
Who should be responsible for knowledge graph?
Usually whoever owns AI platform optimization, because getting the pricing where an assistant can state it rather than hedge depends on it. In practice it falls between teams often enough that naming an owner is half the fix. When this comes up in a pitch, Abu Dhabi and Ahmedabad are the pages that settle it.
What does fixing knowledge graph involve?
The diagnosis is cheap. The fix ranges from an afternoon to a project depending on how deep it sits in the build, and we scope the second after doing the first. The audit version of this is Home and trade services. The ongoing version is Legal and professional services.
Get a straight answer before you commit
Tell us what you sell and what is not working. We will come back with the two or three things we would look at before anything else.