The problem most SEO audits miss
We worked with an Irish SaaS company last year whose product pages wouldn’t break page two for their core terms. Content was comprehensive — 2,000+ words per page. Keyword density looked fine. Backlinks were respectable. Technical SEO was clean. But Google kept ranking competitors who, on paper, had weaker pages.
The difference wasn’t in what they *said*. It was in what they *didn’t mention*.
Google’s Knowledge Graph doesn’t just count keywords. It calculates *salience* — a weighted score for every related entity (people, companies, standards, categories, sibling products) that should appear in context around your main topic. If you write about “Google Ads” but never mention Google AdSense, Google Inc, or the fact that it’s a software application, Google sees your page as incomplete. Not thin in word count. Thin in *knowledge coverage*.
This isn’t new. Google introduced semantic search with the Hummingbird algorithm update in 2012 — the moment it stopped reading “local pizza takeaway” as three isolated keywords and started understanding it as a question. That was AI in search, more than a decade ago. The tools have caught up. The strategy hasn’t changed.
How we reverse-engineered the Knowledge Graph
We wanted to understand which entities Google associates with a given topic and how heavily it weights each one. So we used Google’s own Knowledge Graph API — a free developer tool that returns entity relationships and salience scores for any search term.
We tested it with “Google Ads.”
Result:
– **Google Ads** (primary entity): salience score 32,822
– **Google AdSense** (secondary): 7,000 (roughly 21% as important as the primary)
– **Google Inc** (parent organization): lower weight, but still required
– **Category: computer program / software application**: also required
– **Key people** (Larry Page, Sergey Brin), **years** (founding, IPO), **corporate structure** (Alphabet Inc): each carrying fractional but non-zero salience
These aren’t arbitrary associations. They’re the exact weighted relationships Google uses to model the world. Every entity is classified by type (organization, person, thing, location) and assigned a score that represents its contextual importance.
You can do the same analysis for any topic. Pull the Wikipedia page. Feed it into a simple AI prompt asking for entity extraction and salience calculation (0–1 scale). You’ll get back every concept, person, standard and sibling product you need to mention — and how much weight each carries.
Paid tools like Surfer SEO, MarketMuse and Clearscope automate this at scale. But the underlying logic is free, and you can run it yourself. We did. And it works.
What we changed for the client
For the SaaS client, we mapped the entities around their three core product terms. We identified secondary entities they’d never mentioned: related standards (ISO codes, compliance frameworks), integration partners, competitive product categories, and even the job titles of typical users.
We didn’t write new pages. We restructured their existing product pages to mirror the Knowledge Graph:
– Added a “Related standards” section with two-sentence explanations of each
– Linked to a new “Integrations” hub page (one paragraph per integration, entity-linked)
– Mentioned competitor categories in a comparison table (not by brand name, but by product type — “time-tracking software,” “project management platforms”)
– Added a glossary block defining key terms Google expected (we pulled these straight from the entity extraction output)
No new backlinks. No technical overhaul. Just smarter entity coverage and internal linking.
Result: 40% increase in keyword coverage within eight weeks. Several terms jumped from page two to top five. One term hit position one and held it for three months.
The client asked what we’d changed. We told them: “We taught Google what your product is related to, not just what it does.”
The entity SEO checklist
If you want to audit your own pages for entity coverage, here’s the process we use:
**1. Identify your primary entity.**
This is your main keyword or product. If you’re writing about “Google Ads,” that’s your primary entity.
**2. Pull the Wikipedia page (or equivalent authoritative source).**
Wikipedia is the single largest source for Google’s Knowledge Graph. If your topic has a Wikipedia page, start there. If not, use the top-ranking page for your keyword.
**3. Extract entities using AI or the Knowledge Graph API.**
Feed the content into a prompt like: “Pretend you are an SEO entity extraction engine. List all primary and secondary entities mentioned in this text and assign a salience score (0–1) based on how central each is to the core topic.” You’ll get back a weighted list.
**4. Cross-check your page.**
Do you mention the top five secondary entities? If not, add them — not as keyword-stuffing, but as natural contextual references. A single sentence or internal link is often enough.
**5. Classify entity types.**
Google distinguishes between organizations, people, things (products/software), locations, and events. Make sure you’re covering multiple types, not just repeating the same category.
**6. Check internal linking.**
If a secondary entity is mentioned on another page of your site, link to it. Google uses internal link context to confirm entity relationships.
**7. Monitor keyword coverage, not just rankings.**
Use Google Search Console to track how many keywords each page ranks for (impressions). Entity-rich pages typically rank for 2–3× more long-tail variations than entity-thin pages, even if primary keyword position doesn’t move immediately.
Why this matters more now
Google’s March 2024 core update and the ongoing rollout of AI Overviews (formerly SGE) have made entity coverage even more critical. AI-generated summaries pull directly from the Knowledge Graph. If your page lacks the secondary entities Google expects, it won’t be cited — even if your primary keyword match is perfect.
We’ve seen this with multiple clients in the past six months. Pages that ranked well pre-March dropped out of AI Overview citations entirely, while competitors with stronger entity coverage appeared in the summaries. The difference wasn’t content quality. It was entity completeness.
The good news: this is fixable, and the tools are free.
What to do next
Pull your best-performing page. Run it through the entity extraction process above. Identify the top five secondary entities you’re missing. Add them — one sentence each, naturally integrated, with internal links where relevant.
Then wait eight weeks and check Google Search Console. You should see an increase in impressions (keyword coverage) before you see movement in average position. That’s normal. Entity SEO builds breadth first, then authority.
If you’d rather we do this for you, we’re happy to run a free entity audit on one page — no strings attached. Just send us the URL: [digitaltrawler.com/contact-us](https://digitaltrawler.com/contact-us)
Or book a 30-minute strategy call and we’ll walk through your entire site structure. No pressure. No jargon. Just honest advice.
Because Google’s been doing entity SEO since 2012. It’s time the rest of us caught up.