Not too long ago, I applied for a Head of Marketing role at a local Guelph company, SchemaApp.
I didn’t land it, but the process turned into a real crash course in something I had thought was mostly “technical SEO hygiene”: structured data and knowledge graphs.
The more I dug in, the more it felt like it wasn’t about chasing rich results, but rather a much bigger shift: As marketers we have to realize that machines are becoming our first audience (or at least the first filter), and most B2B websites (and stacks) are not ready for it.
The shift: discovery is moving from “pages” to “meaning”
Classic SEO trained a lot of us to think in documents: pages, keywords, backlinks, and rankings.
But modern search systems (and answer engines) increasingly try to understand entities (things) and relationships (how things connect). That’s the backbone of semantic/entity-first thinking in SEO. (IPullRank)
Forrester has been blunt about the practical implication: if answer engines’ crawlers are trying to interpret content quickly (and often without heavy JavaScript rendering), structured data like schema markup is necessary to help them interpret and contextualize what they find. (Forrester)
And from the “big strategy” side, McKinsey keeps circling the same core truth in their genAI coverage: what’s possible with AI is constrained by your data foundations, not just your models. (McKinsey)
So, when people ask, “Is schema just an SEO thing?” I’m starting to think the better question is:
How do we make our website legible to machines that summarize, recommend, compare, and cite?
Structured data: the instruction manual you attach to your website
Google’s framing is refreshingly straightforward: structured data is a way to describe your content in a standardized format (largely using Schema.org vocabulary), and Google provides specific guidance for what it supports and expects. (Google)
Schema.org itself describes what it’s for in one sentence: a vocabulary that covers entities, relationships between entities, and actions. (Schema.org)
That’s the whole game.
Instead of hoping a crawler “reads” your Services page and guesses correctly, structured data lets you say:
- this is an Organization
- this is a Service
- this is an Article
- this person is the author
- this page is about that topic/entity
It’s not romantic. It’s not mysterious.
It’s just: make the meaning explicit.
Knowledge graphs: when you stop treating content like a pile of posts
A knowledge graph is basically what happens when you take that “explicit meaning” idea seriously across the whole site.
SchemaApp calls this a Content Knowledge Graph: A structured layer of entities and relationships derived from your content. (SchemaApp)
But they’re not alone in the broader narrative.
Forrester has been making the case that knowledge graphs help manage knowledge (not just data), and that they can complement AI by providing a more reliable (i.e. less hallucinatory) layer for discovery and decision-making. (Forrester)
While at Gartner, knowledge graphs keep showing up as an enabling technology for AI architectures. Gartner’s 2024 Hype Cycle for AI notes knowledge graphs moving along the maturity curve and call out the real work: getting from prototype to production, and maintaining quality at scale. (Readwise, Neo4j)
So the idea isn’t “everyone needs a fancy graph database.”
It’s: stop publishing content like it’s disconnected. Start publishing it like it’s a model of your business.
Why this hit me: Our internal systems are structured… our external footprint often isn’t
Here’s an incongruency I can’t unsee:
Inside the company, we obsess over structure:
- CRM objects, lifecycle stages, lead routing rules
- attribution models
- governance (“what’s the source of truth?”)
Outside the company, we often publish a fog:
- inconsistent service naming
- proof scattered across random pages
- “we do everything” messaging
- content islands that don’t reinforce each other
This gap begins to matter a lot more as discovery becomes AI-mediated.
McKinsey’s “data dividend” point maps cleanly here: if your foundations aren’t solid, AI outcomes aren’t either. (McKinsey)
Forrester makes a similar argument in their AI-readiness framing: modern data platforms + governance determine trustworthy AI performance, and knowledge graphs are explicitly part of that “AI-ready” toolkit. (Forrester)
This is where structured data stops feeling like “SEO markup” and starts feeling like brand infrastructure.
What this changes for B2B marketing (in plain English)
A few practical implications that I would take to the bank:
1) Your positioning needs to be machine-readable
If your site is vague, machines will still summarize you, but they’ll do it poorly. Structured data is one way to reduce ambiguity. (Forrester, Google)
2) Proof needs to behave like a connected system
Case studies, integrations, partner pages, industries served are key credibility signals. Knowledge-graph thinking pushes you to connect them intentionally instead of letting them float as lonely PDFs and one-off pages. (Forrester)
3) “SEO” expands into “retrievability”
Forrester’s answer engine optimization angle is basically: crawlers that assemble answers need content they can parse, trust, and cite. Schema is part of that. (Forrester)
4) AI outcomes will expose your content architecture
McKinsey’s scaling-genAI work keeps emphasizing that data management is a barrier to value. If your public-facing knowledge is messy, you’ll feel it as non-human search becomes prevalent. (McKinsey)
"But we can't boil the ocean"
If you’re not trying to “build a knowledge graph program,” but you do want to make progress:
- Pick 10–20 core entities
Company, core services, core industries/use cases, key people/SMEs, and your best proof assets. - Get brutally consistent about naming + descriptions
Same service name everywhere. Same definition everywhere. Same “who it’s for” everywhere. - Implement structured data that matches reality
Reference Google’s documentation for what matters to Google, and Schema.org for the underlying vocabulary. (Google, Schema.org) - Connect content to those entities on purpose
Not just “related posts.” But rather: “this article is about this problem, for this industry, tied to this service.”
Even if you only do it for your top 20 pages, you’ll start to feel the difference in clarity.
My journey (with a dose of humility)
I’m still firmly in learning mode.
I came into the SchemaApp interview process thinking structured data was a tactic. And although SchemaApp is just one lens, their mission of: “teach marketers what semantic structure looks like in practice” seems like a good one.
There are some great talks/videos from Martha van Berkel (CEO of SchemaApp) that have been genuinely helpful without disappearing deep into computer science.
You’ll also find similar themes in entity SEO circles (iPullRank, etc.) And, with analysts (Forrester, Gartner) and consultants (McKinsey) all pointing to the same underlying constraint: AI is only as good as the structure and governance of the knowledge it’s grounded in, it got me thinking of structured data as an AI-readiness layer for your brand.
But I’m also acutely aware of my lack of practical expertise.
So instead of pretending I’ve got the definitive playbook, I’ll end with this:
Let’s compare notes.
If you’re implementing schema at scale, experimenting with entity-first content, or thinking about “answer engine” visibility, I’d love to hear what’s working (and what’s not). Drop me a note, or send an example of a site doing this well so I can learn from it.
I think after some conversations, some more learning, and some real-world field testing, I'll circle back here, and share an update on how deep the rabbit hole goes.
Stuff I read/watched:
- Forrester on answer engine optimization + schema still mattering (Forrester)
- Forrester on genAI + knowledge graphs (Forrester)
- McKinsey on data foundations for genAI (McKinsey, McKinsey)
- Gartner Hype Cycle excerpt/summaries referencing knowledge graph maturity + production challenges (Readwise, Neo4j)
- Google Search Central structured data intro/guidelines (Google)
- Schema.org overview (entities + relationships) (Schema.org)
- SchemaApp on Content Knowledge Graphs + why they matter (SchemaApp)
- Martha van Berkel videos (practical explanations) (YouTube, YouTube)
