Semantic SEO: from intent to a connected system.
Semantic SEO here means organising content so the reader gets a complete answer and the search engine gets clear pages, links and structured data. We use “topical authority” and “cost of retrieval” as analytical models — not as named Google metrics.
What does Google actually document?
Google publicly recommends helpful, people-first content, crawlable links, descriptive headings and structured data that matches the visible content. It does not publish a formula for “topical authority” and does not require strict silos.
That distinction runs through this whole section. Where a claim comes from Google documentation, we link the source. Where it is a working model we use to make decisions, we say so — including for “topical authority” and “cost of retrieval”, both of which are analytical frames rather than named Google metrics.
Guides and implementation
Complete SEO architecture
Macro topic → cluster → page → micro-intent, with an entity–attribute matrix.
02Query templates
The query pattern as a second route to authority, the threshold for opening a new page, and cost of retrieval.
AIAI search and SEO
How an AI answer is assembled, which components make a page extractable, and what our own measurement shows.
AIGoogle AI Overviews
What AI Overview and AI Mode are, where they draw sources from, and why the share differs by page type.
BotsAI crawlers and robots.txt
Three purposes of crawling, control per agent, and what blocking actually costs.
TextWriting for AI answers
Six patterns that break a section, why repeating the name is required, and how to check your text.
CodeA cohesive JSON-LD graph
What schema does, what it does not do, and how the graph is validated.
ContentClear writing for people and machines
Direct answers, a logical hierarchy and the removal of ambiguity.
Two routes, one system
Authority is built along two axes that are easy to confuse. The first covers every entity in a class through the attributes they share. The second covers every variation of a single query pattern, and does not require the pages to be topically related at all.
| Route | What gets covered | Where authority attaches |
|---|---|---|
| Entity–attribute | All members of a class, across their shared attributes | To the topic |
| Query template | All variations of one query pattern | To the query format |
| Hybrid | Every entity, every attribute, in every variation of the pattern | To both |
Most sites do the first and stop. The second is covered on the query template page, together with the threshold that decides whether a query deserves its own URL.
Does a relevant cross-topic link “dilute” a silo?
No such general rule exists. A relevant contextual link helps both the reader and Google find and understand a related page.
That is why this hub links naturally to alt text and image context where the textual and visual layers genuinely meet. A silo that refuses relevant links is optimising for a diagram rather than for the reader.
This silo describes the method. How it is applied to a site being built from scratch, with a price and a timeline, is shown by semantic web development.
Building a site on this model?
Send your sector and the topics you plan to cover.