Content architecture and SEO

Search intent

The task a searcher appears to be trying to complete, commonly informational, navigational, commercial, or transactional. Intent is inferred from behavior and results, not directly disclosed by the person.

Query class

A category assigned to queries sharing a purpose or structure. Classes help aggregate behavior but can hide mixed or changing intent within one phrase.

Navigational query

A query intended to reach a known site, brand, or page. Brand appearance is expected here and says little about non-branded topical visibility.

Informational query

A query seeking an explanation, fact, or guidance. It may be satisfied without a click, so rankings and referral traffic measure different outcomes.

Transactional query

A query indicating an intended action such as buying, booking, or downloading. The label describes likely intent, not whether a transaction will occur.

Commercial investigation

Research comparing products, vendors, prices, or approaches before a decision. It sits between information seeking and transaction and often changes across the journey.

Entity-first content

Content organized around clearly identified people, organizations, products, places, and their relationships. It reduces ambiguity but does not create external corroboration.

Information architecture

The organization, labeling, and linking of content so people and machines can locate it. A clean hierarchy supports discovery without guaranteeing that every page is valuable.

Taxonomy

A controlled hierarchy of categories used to classify content. It improves consistency while becoming harmful when categories overlap, remain empty, or generate unlimited URLs.

Ontology

A formal model of entity types, properties, and relationships in a domain. It supplies shared semantics but cannot ensure that instance data is complete or correct.

Topic cluster

A set of related pages connected to a broader central resource. It is a publishing pattern, not a recognized engine-side unit or measurable authority score.

Pillar page

A broad page that links to deeper resources on a subject. Its value comes from useful coverage and navigation, not from the label or page length.

Content hub

A navigable collection centered on a subject, audience, or task. A hub can improve discovery while still containing redundant or weak pages.

Semantic HTML

Elements chosen for their meaning and structure rather than appearance alone. It improves machine interpretation and accessibility but does not replace clear content.

Heading hierarchy

The ordered use of heading levels to expose document structure. Visual size alone does not create a heading, and valid levels do not guarantee coherent organization.

Internal link

A link from one page to another on the same site. It supports discovery and context, but its anchor and destination must still be useful.

Measured: BROKEN_INTERNAL_LINK 2%

Anchor text

The visible or accessible label of a link. It supplies context about the destination, while repeated keyword-heavy text can be unhelpful or manipulative.

Link context

The surrounding text and structure that clarify why a link exists. It can improve interpretation, but no public engine exposes its exact weighting.

Deep link

A link pointing beyond a site’s homepage to a specific resource or application state. It improves direct discovery when the target is stable and accessible.

Broken link

A link whose target cannot deliver the intended resource. One failed check can be temporary, so status, redirects, body, and repeat observations matter.

Measured: BROKEN_INTERNAL_LINK 2%

Redirected internal link

An internal link that reaches its content only after one or more redirects. It usually works but adds latency and leaves old URL forms embedded in the graph.

Link equity

An informal name for ranking value believed to pass through links. Engines do not publish a page-level quantity that external tools can measure directly.

PageRank

A link-analysis method that models importance from the structure and weight of incoming links. Public toolbar scores are gone, and vendor authority metrics are not Google PageRank.

Dofollow

Informal shorthand for an ordinary link without a qualifying `rel` value such as `nofollow`. There is no `dofollow` HTML attribute with special ranking force.

Nofollow

A link qualification indicating that the publisher does not want to endorse the destination in the ordinary way. Search engines may treat it as a hint rather than an absolute command.

Sponsored link

A link marked `rel="sponsored"` to identify advertising, sponsorship, or compensated placement. The label communicates relationship; it does not validate disclosure law compliance.

User-generated link

A link marked `rel="ugc"` because users rather than the publisher supplied it. It identifies origin category, not quality or safety.

External link

A link to a different site or origin. It can support sourcing and navigation but does not automatically transfer trust or prove a claim.

Inbound link

A link from another site to a page. Counts depend on crawler coverage, canonicalization, and deduplication, so tools legitimately disagree.

Referring domain

A distinct domain observed linking to a target. One domain can supply many links, and domain count alone ignores relevance, placement, and authenticity.

Link neighborhood

The sites and pages connected around a document through links. It is an analytical metaphor rather than a published engine feature with fixed boundaries.

Orphaned entity

An entity declared without meaningful connections to the rest of a site’s graph or content. The node exists but offers few paths for corroboration or discovery.

Content decay

Loss of usefulness or visibility as information becomes stale, competitors improve, or demand changes. A traffic decline alone cannot identify which cause applies.

Content refresh

A substantive update to accuracy, coverage, evidence, or usability. Changing a date without changing content creates a freshness claim unsupported by the page.

Historical optimization

Updating established pages using current performance and information rather than publishing replacements. It preserves accumulated signals but can erase useful historical context if done carelessly.

answer block

A short passage, typically forty to a hundred words, placed under a page's opening that states the page's answer in a form an assistant can lift whole. It is what a speakable property should point at; writing one is the single cheapest change toward being quoted, and it only works if the rest of the page supports it.

apex domain

The bare domain without a subdomain: example.com rather than www.example.com. Which of the two serves the site is a choice; both serving it is a duplicate. The apex has one technical limit, that a DNS CNAME cannot be placed on it, which is why many hosts prefer www as the canonical host.

Measured: HOST_DUPLICATE_200 3.3%

alias domain

A domain whose homepage and machine files all redirect to one other host: a second name for the same site, not a second site. A scanner treats it as an alias rather than as a homepage that redirects off-host, because every path agrees; the alias should carry nothing the target does not.

Measured: DOMAIN_IS_ALIAS 0.9%

parked domain

A registered domain serving a registrar's or reseller's placeholder page instead of a site. Its page can run to hundreds of words and pass a thin-content check on length; a scanner recognises it by the placeholder's signatures and by every path answering the same page.

local pack

The block of map results, usually three, that a search engine shows for a query with local intent. Placement in it is driven by the business profile and its corroboration far more than by the website, which is why a site can rank on the page and be absent from the pack, or the reverse.

Google Business Profile

The business record Google maintains for a local entity: name, category, hours, service area, reviews, photos. It is the primary source for the local pack and for most assistant answers about a local business, and its facts win over the website's when the two disagree.

data aggregator

A company that compiles business listings and sells them to directories, map providers and assistants. In the United States, Data Axle is the aggregator named in Google's own provider list; a business absent from an aggregator is absent from every product that buys from it, whatever its website says.

Data Axle

The United States business-data aggregator that appears in Google's Maps provider list and feeds many directories. Its listings can be searched by phone number and claimed by the business; its site refuses datacenter addresses, so a scanner records it as refused rather than as no listing.