Organize your site so every page answers one question and belongs to one topic — the taxonomy foundation that makes your site retrievable and citable by AI engines and humans alike.
Taxonomy is the grouping layer of your site — the principle that every page belongs to exactly one hub (topic cluster), with a clean /[hub]/[slug] canonical path. It is distinct from ontology (meaning/entity types), schema.org markup, or topical authority; it is the foundational *structure* that lets AI engines (and your visitors) navigate your site as coherent topic clusters rather than a flat pile of pages.
/[hub]/[slug] — root-relative, no trailing slash. This clean structure is how AI engines infer topic proximity and how humans can predict where to find content.Use the interactive map below to explore each one — click any node to read what it covers and jump to its page.
How a law firm site groups its content into retrievable, citable topic clusters.
The grouping principles that make every page discoverable and every topic coherent.
Site taxonomy is a two-layer information architecture: taxonomy (grouping pages into hubs) and ontology (classifying pages by type and wiring semantic relationships via schema.org). It is the first-class foundation of every law firm site we build.
Hub-and-spoke is a content structure where a authoritative core page (the hub) links down to narrower, specialized pages (spokes), each covering a specific scenario, location, or practice variant. Spokes link back to the hub and sideways to siblings.
Practice-area information architecture is the systematic organization of a law firm's practice areas into a hub (the main practice page) with multiple spokes (specific legal scenarios, case types, or subtopics). One page → one hub; every spoke links back to its hub and sideways to siblings.
A canonical URL is the single authoritative address for a page's content. Search engines use it to consolidate ranking signals and avoid indexing duplicate versions of the same page. Multiple versions of the same page (with or without www, with or without trailing slash, different protocols) split your authority and ranking power.
Cannibalization occurs when two or more pages compete for the same keyword or topic, confusing search engines about which page to rank. For law firms, this often happens when a practice-area page and a city-specific page target the same service-plus-location intent without a clear hub-spoke structure.
Taxonomy is the grouping layer of your site — the principle that every page belongs to exactly one hub (topic cluster) with a clean /[hub]/[slug] canonical path. It is distinct from ontology (entity types and relationships), schema.org markup (how to encode meaning), topical authority (depth and external signals), or entity identity (who you are). Taxonomy is purely about structure — how pages are organized into coherent topic clusters.
The rule is absolute: one page, one hub. A page about premises liability in Los Angeles lives in the /personal-injury hub at /personal-injury/premises-liability-los-angeles. It never belongs to both personal injury and municipal liability; it lives in one hub and links sideways to related topics.
A hub is born when you have three or more genuinely distinct, interlinked pages on one topic. Until then, those pages wait as spokes (ready to join when the 3rd page arrives) or as flat guides (if they're not specific to any hub). This prevents creating thin, orphan hubs with one or two pages.
A hub is the authoritative, overview page for one topic — the answer to the broadest question in that area. Spokes are focused pages that answer specific sub-questions, each distinct and genuinely different from its siblings. A hub on 'personal injury law' might have spokes for 'premises liability,' 'medical malpractice,' 'wrongful death,' and 'slip-and-fall settlements.' Each spoke is self-contained, but all are woven together by links.
The linking pattern: the hub links down to every spoke (often in a navigational sidebar or spoke directory); each spoke links up to its hub (breadcrumb, schema); and spokes link sideways to siblings (inline, at the end, or via a spoke sidebar). This creates a coherent cluster that engines recognize as belonging together.
When a page answers a very specific question (e.g., 'how long do I have to file a premises liability claim in California?'), it is a spoke. When it answers the umbrella question (e.g., 'what is premises liability?'), it is a hub. Hubs are few; spokes are many. A site of fifty pages might have five hubs and forty-five spokes.
| Dimension | Hub | Spoke |
|---|---|---|
| Scope | Broad umbrella topic (e.g., 'Personal Injury Law') | Specific sub-question (e.g., 'Slip-and-Fall Settlement Guide') |
| Audience | Anyone new to the topic | Someone with a specific scenario |
| Length | 1,500–3,000 words; many sections | 800–1,500 words; focused on one scenario |
| Links to others | Links down to every spoke + sideways to other hubs | Links up to its hub + sideways to siblings |
| Count | Few (one per major topic) | Many (5–15 per hub) |
| Existence rule | Requires ≥3 member spokes | Can exist alone (waits to join a hub) |
Hub vs. Spoke characteristics — how to decide which type a page should be.
Every page's canonical URL is /[hub]/[slug] — root-relative, no trailing slash, no query parameters. Examples: /personal-injury/premises-liability-guide, /employment-law/los-angeles-attorney, /tools/client-intake-form. This structure is human-readable, predictable and crawlable by search engines and AI bots.
Root-relative URLs (starting with /) work on any domain. Never hardcode the full domain (https://your-firm.com/personal-injury/...) in content or nav — that breaks on domain cutover and leaks staging environments. Absolute URLs are reserved for schema.org @id and sameAs (which require the final production domain) and for external references.
Hub names are singular, durable and high-level: /personal-injury, not /personal-injuries-and-tort-law. Spoke slugs are descriptive and specific: /premises-liability-guide, not just /guide-1. Use hyphens between words; never underscores or spaces. Keep slugs under 60 characters (they appear in search snippets).
Internal links flow in three directions: up (spoke to hub), down (hub to spokes), and sideways (spoke to sibling). Never link directly from one unrelated hub to another; use a topic bridge instead. This keeps the topology coherent.
Every internal link must be root-relative and have descriptive anchor text. Good: <a href="/personal-injury/premises-liability-guide">premises liability guide</a>. Bad: <a href="#">click here</a> or <a href="https://staging.firm.com/personal-injury/...">. Anchor text tells both humans and engines what the linked page is about; it should be readable if the link is stripped.
Link density: a short intro paragraph (1–2 sentences) might have 0–1 internal links; a medium paragraph (3–5 sentences) 1–2; a longer one (6+ sentences) 2–3. One link per concept per paragraph is enough. The spoke directory or sidebar on the hub page supplies the authoritative list of all spokes, so you don't need to squeeze in every spoke link in the article text.
| Direction | From → To | Purpose | Example anchor |
|---|---|---|---|
| Up | Spoke → Hub | Breadcrumb, context, read more | personal injury law |
| Down | Hub → Spoke | Spoke directory, sidebar, hub text | premises liability claims in California |
| Sideways | Spoke ↔ Spoke (siblings) | Related scenarios, comparison | also see our guide on medical malpractice |
| Bridge | Hub ↔ Hub (different topics) | Rare, with context | for tort law specifics, see employment law |
| Meta | Any → Audit page | Internal tools, checklists | see our LLM SEO checklist |
Internal linking patterns within a taxonomy — direction, anchor text, and examples.
Taxonomy doesn't directly affect rankings, but clean structure supports both retrieval and ranking. When AI engines retrieve passages, they cluster by topic proximity in embedding space. Pages from a coherent, tightly-linked hub cluster together; scattered, unconnected pages are treated as independent. A hub where every spoke links up and sideways signals topical authority to both engines and humans.
From a ranking perspective, coherent internal linking concentrates authority within the cluster and prevents pages from competing with each other for the same keyword. If you have three premises-liability pages scattered across three hubs without internal links, they compete and dilute each other's authority. If they're all in one /personal-injury hub and tightly linked, they reinforce each other's authority on that topic.
The practical takeaway: a well-organized hub with five tightly-linked spokes typically outranks and gets cited more often than fifty flat, unlinked pages on similar topics. Topology matters more than volume.
The intents AI engines fan a search into — and where we make your firm the answer.
Taxonomy is the *grouping* layer — the organizational structure of hubs and spokes. Ontology is the *meaning* layer — the entity types and relationships (ServiceSpoke, LocationSpoke, Hub, Guide). Schema.org is the *markup* layer — how to encode those meanings in JSON-LD so machines can read them. Topology is about structure; ontology is about semantics; schema is about syntax. Start with taxonomy, then layer ontology and schema on top.
A hub is born when you have three or more genuinely distinct, interlinked pages on one topic. Until then, keep those pages as spokes (waiting for the 3rd member) or as flat guides. The rule prevents creating empty hubs with one or two pages, which looks thin to both engines and humans. Flat guides (like FAQ or Tools) live outside any hub if they're not specific to one practice area or place.
Hub names are the practice areas and topics your firm owns: <code>personal-injury</code>, <code>employment-law</code>, <code>ai-legal-marketing</code>, <code>tools</code>. Spoke slugs are specific: <code>premises-liability-guide</code>, <code>slip-and-fall-settlement</code>, <code>los-angeles-employment-attorney</code>. Never use acronyms or abbreviations in URLs; spell them out. Keep slugs descriptive and memorable.
No. The core rule is: one page, one hub. If a topic spans two hubs (e.g., a page on 'AI contract review for employment law'), it lives in the primary hub and links sideways to related pages in secondary hubs. Breaking this rule creates duplicate content issues and confuses engines about what the page is actually about.
Depth is measured in coherence and interlinks, not raw page count. A hub with eight genuinely distinct spokes on one practice area, each tightly linked to the others and to the hub, is stronger than forty pages scattered across multiple hubs. The sweet spot is 5–15 pages per hub, all tightly interlinked. Deep internal linking compounds authority; scatter dilutes it.
Version-control the taxonomy as a data file: a registry of hub definitions and a map of every page (slug → hub + entity type). Generate sitemaps, nav, hub directories, and redirects from this file, never hand-edit them. If a page's hub assignment changes, regenerate everything — that's what the file is for. This ensures consistency and prevents orphan pages.
Set up 301 redirects from the old flat URL to the new <code>/[hub]/[slug]</code> path. The redirect must be byte-exact and go to a real page that returns 200. Test that every old URL resolves. Once the redirect has been live for a few months and Search Console confirms the crawl, you can remove the old page entirely, but keep the 301 in place indefinitely.
Yes, but carefully. Audit first: group pages by topic, identify which groups have 3+ pages (those become hubs), and which have 1–2 (those wait as flat spokes or guides). Use 301 redirects for the cutover so you don't lose rankings. Expect to lose traffic briefly during the transition; it recovers as the new topology settles. Start with the highest-traffic hubs to prove the model before doing the whole site.
Internal links should flow within and across hubs: up (spoke → hub), down (hub → spoke), and sideways (spoke ↔ sibling spoke), with occasional bridges to secondary hubs. Never link directly from one hub to an unrelated hub without a bridge explanation. Every internal link must be root-relative (<code>/practice-area/page-slug</code>) and have descriptive anchor text. This creates a coherent, navigable graph.
Yes. A common taxonomy is to have service hubs (practice areas) and location hubs (states, cities, offices). Pages in a location hub might be <code>/los-angeles/employment-law-attorney</code> and link both up to the location hub and sideways to the service hubs. The key is that each page belongs to one hub; a page can reference multiple topics but is grouped under one primary hub.
Indirectly. Taxonomy doesn't directly affect rankings, but clean structure supports both ranking and citation. A coherent hub with tight internal linking makes it easier for engines to see your site as an authority on one topic. It also keeps pages from competing with each other for the same keyword — which can suppress both. A well-structured hub typically outranks flat competitors on the same topic.
Probably not immediately. If flat pages are already ranking and generating traffic, the risk of reorganization outweighs the benefit for those specific pages. Instead, apply taxonomy to *new* content and use redirects to gradually move high-traffic pages to the new structure over several months. Measure the impact on a small pilot (5–10 pages) before doing the whole site.
Start with a free 23-point AI visibility report — see exactly where ChatGPT, Gemini, Claude and Perplexity cite your firm today, and what closes the gap. Delivered in 24 hours. No credit card.