What Is Site Taxonomy? The Foundation of AI-Citable Law Firm Websites

Site taxonomy is the organized, hierarchical structure that groups your law firm's content into hubs (broad topic centers) and spokes (focused chapters). A hub might be "family law"; the spokes are "divorce in California," "child custody factors," and "spousal support calculations." Every page on your site belongs to exactly one hub. This organizational model is not just for navigation—it directly affects how search engines and AI models understand your expertise and citeability.

Taxonomy differs from ontology. Taxonomy is grouping (what categories do pages fall into?). Ontology is meaning—the semantic types and relationships between pages (Is this page a Guide, a ServiceSpoke, or a LocationSpoke? Does it belong to a specific practice area or jurisdiction?). A complete information architecture uses both: taxonomy organizes the content tree; ontology wires it with structured data so machines understand the relationships.

For law firms, a well-built taxonomy accomplishes three things simultaneously: it creates topical authority (Google recognizes that you have deep coverage of a practice area), it improves AI citability (Gemini and Claude's retrieval systems favor semantically clear content that lives in topical clusters), and it simplifies internal linking (every spoke links up to its hub and sideways to sibling spokes, automatically building the topical cluster Google rewards). Without a taxonomy, a law firm website becomes a collection of isolated pages competing with each other, diluting topical authority and making it harder for AI engines to recognize the firm's expertise.

The practical impact: a firm with a clear taxonomy that has invested in a family law hub with 12 interconnected spokes (divorce, custody, spousal support, prenups, paternity, etc.) will rank higher for family law queries and be cited more by AI engines than a firm with 12 isolated pages scattered across the site with no hub.

The hub/spoke model is simple: a hub is a broad topic (e.g., "family law," "employment law," "San Francisco office"); spokes are focused chapters or subtopics under that hub. Every spoke page links back to its hub, and the hub page links down to all its spokes. This creates a topical cluster that search engines recognize as deep expertise.

A hub requires 3 or more related spokes. This rule prevents thin, duplicate-prone hub pages. If you only have two pages on a topic, they remain spokes under a broader parent hub until the topic grows. For example, if you have written two pages on non-compete agreements (one on enforceability, one on employee mobility), these stay as spokes under an "employment law" hub until you have written enough related content (employee confidentiality, trade secrets, etc.) to justify a dedicated "non-compete" hub. This guardrail ensures that every hub genuinely represents a topic your firm has invested in deeply, not a thin collection of loosely related pages.

Hubs live at clean, canonical URLs: /[hub-name] (e.g., /family-law, /employment-law, /san-francisco-office). Spokes live at /[hub-name]/[slug] (e.g., /family-law/divorce-in-california, /employment-law/non-compete-enforceability). This naming convention is predictable and portable—if you ever migrate to a new domain or CMS, the structure survives intact.

Each hub page serves two functions. First, it is a gateway: it introduces the topic, summarizes the key subtopics, and links down to every spoke. This allows a visitor to land on the hub (via direct navigation, internal links, or organic search) and quickly find the chapter they need. Second, it is an authority page: it is long, comprehensive, and rich with schema.org markup that signals to AI models that this is a canonical resource on the topic. Hubs accumulate more inbound links and are more frequently cited by AI engines than individual spokes.

Spokes are the deep-dive chapters. Each spoke focuses on one subtopic or scenario. A divorce spoke is not "everything about family law"—it is specifically "how divorce works in California." A medical malpractice spoke is not about all personal injury—it is about the specific elements, damages, and defenses in medical malpractice claims. This focus is what makes spokes citable: when an AI model searches for "how long to file a divorce in California," it finds the focused divorce spoke, not a generic hub.

The sidebar navigation is critical. Every hub page and every spoke page displays a sidebar showing all spokes in that hub (with thumbnails and brief descriptions). This accomplishes three things: (1) it helps visitors discover related chapters, (2) it increases internal link juice (every spoke page links to every sibling spoke, creating topical clustering), and (3) it signals to search engines and AI models that these pages are semantically related and part of one cluster. A visitor who lands on the custody spoke can immediately see the divorce, spousal support, and paternity spokes and navigate between them.

A well-designed taxonomy improves SEO and GEO in parallel. For traditional Google search, taxonomy builds topical authority—Google's systems recognize that your family law hub has a high density of internal links, deep coverage, and interconnected content, and it rewards the hub and its spokes with higher rankings on family law queries. For AI search engines (ChatGPT, Claude, Gemini, Perplexity), taxonomy improves citability by creating semantic clusters that the model's embedding system naturally recognizes as coherent topics.

Here is how it works. When Perplexity retrieves passages for a query like "can I modify spousal support in California?", it embeds the query into a vector and searches its index for passages with similar embeddings. If your spousal support spoke is well-written, semantically consistent, and embedded in a hub with related pages (divorce, alimony, family law schema), the embedding for that spoke will cluster near other spousal support content. Perplexity's retrieval system will find it and rank it highly. Conversely, if your spousal support content is isolated (no hub, no internal links, mixed terminology), the embedding will be diffuse and may not rank as high even if the content is equally good.

The topical clustering effect is measurable. A hub with 10 well-linked spokes creates more internal link velocity (pagerank) to each spoke than 10 isolated pages would. This is standard SEO: a hub-spoke structure is more efficient at distributing authority than a flat structure. For AI engines, the effect is similar: a coherent cluster of semantically related pages is more likely to be treated as a reliable source on the topic than scattered pages on the same topic.

Taxonomy also affects URL simplicity and memorability. A URL like /family-law/divorce-in-california clearly signals the topic hierarchy and is easier for both humans and machines to reason about than a flat URL like /divorce-in-california-2024. The hub prefix /family-law tells the model "this is part of the family law topic cluster" even before it reads the content. This is a small signal, but AI embeddings rely on consistent patterns, and clean taxonomy improves those patterns.

One final benefit: taxonomy makes canonical URL management easier. With clear hub/spoke relationships, you eliminate accidental duplicates. You do not have both /divorce and /family-law/divorce competing for the same traffic. Instead, you have one canonical hub page (/family-law) and one divorce spoke page (/family-law/divorce). This is especially important for law firms with many locations or practice areas: a multi-office firm with a hub/spoke structure can have one family law hub at /family-law and location-specific spokes at /san-francisco/family-law, /los-angeles/family-law, etc., all clearly related but with no ambiguity about which is canonical.

Building a taxonomy starts with research. Identify your core practice areas and jurisdictions, then cluster the content you have (and the content you plan to create) under each hub. The key guardrail: a hub requires 3+ spokes, and every spoke must be genuinely unique content, not a token-swapped or thin variant.

Step 1: List your practice areas and locations. For a mid-size family law firm, this might be: Family Law (core hub), Immigration (separate hub), and offices in San Francisco, Los Angeles, and Denver. Do not create a hub for every possible topic you might serve someday; start with what you actually practice and have (or plan to create) content for.

Step 2: Identify content clusters. Go through your existing content and group pages by topic. If you have three pages on divorce, two on custody, and one on spousal support, you have a 6-page family law cluster. If you have two pages on employment law and one on sexual harassment, you have a 3-page employment law cluster. Clusters with 3+ pages are candidates for hubs; clusters with 2 or fewer stay as spokes under a parent hub.

Step 3: Assign hub/spoke relationships. For each cluster, designate one hub page (the broad, authoritative overview) and the rest as spokes (the focused chapters). Spokes that belong to multiple logical hubs (e.g., a spousal support page that is relevant to both divorce and family law generally) should link to both hubs, but they live under only one hub canonically.

Step 4: Define URLs. Use the format /[hub-name]/[spoke-slug] with no trailing slash. Keep hub and spoke names lowercase, descriptive, and consistent. Do not change these lightly—once a URL is indexed and cited, changing it requires a 301 redirect, which costs some ranking power. Use a naming convention you can commit to. For example, /family-law, /family-law/divorce-in-california (not /family-law/divorce-california-2024 or /fw/divorce).

Step 5: Audit for anti-doorway compliance. This is critical. Before launching a spoke, verify that it contains genuinely unique, locally relevant content. A San Francisco divorce spoke must describe California divorce law specifically, reference San Francisco courts, and include local flavor. A generic divorce-in-any-state template swapped into every city is doorway content and will be demoted by Google's helpful-content system and ignored by AI engines. If you cannot source unique content for a spoke, do not create it—wait until you have enough material to make it substantive.

Step 6: Implement and test. Build the hub pages (comprehensive overviews with links to all spokes), create the spoke pages (focused chapters linking back to the hub and to siblings), and add the sidebar navigation showing all spokes. Test the internal links: every spoke must link to its hub and to at least 2–3 sibling spokes. Every hub must link to all spokes.

Step 7: Update navigation and schema. Update your site's top-level navigation to expose every hub (or at least the major ones). Add schema.org markup that expresses hub/spoke relationships: each spoke gets a `CreativeWorkSeries` schema pointing to the hub, and each hub gets a `hasPart` array pointing to all spokes. Update the sitemap to reflect the new hierarchy.

The most common mistake is creating thin hubs. A firm creates a hub for every practice area they might offer, even if they have only one page on the topic. Result: the hub page is thin (one sentence intro, one link to the spoke), and Google sees it as low-value. Guideline: if you have fewer than 3 pages on a topic, do not create a hub yet. Use the parent hub instead.

The second mistake is mixing taxonomies. A firm has both /family-law and /divorce as hub pages, with overlapping content. A visitor (or a search engine) does not know which is canonical. This creates internal link competition and dilutes topical authority. Guideline: one hub per distinct topic. If divorce is a subtopic of family law, it lives under family law as a spoke, not as a sibling hub.

The third mistake is hardcoding location-specific facts into a shared template. A /san-francisco/family-law spoke page copy-pastes the same text from /los-angeles/family-law, only changing the city name. Result: false claims. The San Francisco page falsely claims to serve Los Angeles; the LA page falsely claims to serve San Francisco. This is doorway content. Guideline: each location spoke must have unique content. Research the real court locations, judge names, local family law bar, and regional quirks for each city. If you cannot source unique content, the page waits.

The fourth mistake is creating a hub for every query variant. A firm creates separate hubs for "family law," "family law services," and "family law in California." Result: thin hub pages with minor wording differences. Google and AI engines see these as duplicates and may penalize both for perceived keyword stuffing. Guideline: one hub per topic, query variants served by the spokes and FAQs. /family-law is the hub; a spoke on "family law in California" lives at /family-law/california-family-law.

The fifth mistake is orphaning pages. A firm redesigns the site and forgets to add internal links from the hub to a spoke. That spoke page no longer receives internal link juice and becomes harder to find and rank. Guideline: implement a programmatic check: every spoke must have at least one link from its hub, and the hub must link to all spokes. Automate this in your build pipeline.

The sixth mistake is inconsistent entity terminology. A page alternates between "statute of limitations," "filing deadline," and "prescriptive period." AI models embed text, and inconsistent terminology creates noisier embeddings. Guideline: define a canonical term for each key concept and use it consistently. If you must use a synonym, do so once and link it to the canonical term via hyperlink or schema.

Taxonomy is not a one-time design decision—it is the living foundation of your content strategy. As you create new content, you place it within the taxonomy, which ensures it clusters properly with related content and inherits the hub's authority signals. This is the bridge between taxonomy and the content engine.

The content engine (research → semantics + vectorization → clustering → hub/spoke authority) relies on a taxonomy to be effective. If you do not have a taxonomy, the content engine can still identify natural clusters in your content, but those clusters will not map to your site's structure. Instead, you will create new content in a vacuum, and it will not benefit from hub/spoke linking. With a taxonomy, every new spoke you write automatically becomes part of a topic cluster, gains links from the hub and its siblings, and inherits the topical authority the hub has accumulated.

This is also why GEO (Generative Engine Optimization) depends on taxonomy. AI models understand topical authority through semantic clustering and internal linking. A firm with a clear taxonomy—deep family law cluster, distinct employment law cluster, location-specific spokes—signals expertise across multiple dimensions simultaneously. The model recognizes that the firm specializes in family law (because the cluster is deep), that it has local presence (because location spokes exist), and that it is organized and thoughtful about its expertise (because the taxonomy is coherent). This reputation translates into higher citation likelihood.

To maintain taxonomy integrity over time, keep a single source of truth: hubs.json (registry of all hubs, with metadata like name, topic, target audience) and page-map.json (every page on the site, with its hub assignment, entity type, and semantic relationships). When you add new content, you add it to page-map.json, and then regenerate the outputs: navigation, spoke sidebars, sitemaps, canonicals, and schema. This prevents drift. A page that is created but not added to page-map.json is invisible to this system and will not benefit from taxonomy-based interlinking.

As a law firm grows across locations and practice areas, taxonomy becomes critical for avoiding duplicate content and maintaining clear authority. A firm with a San Francisco office and a Los Angeles office faces a choice: one /family-law hub with separate /san-francisco/family-law and /los-angeles/family-law spokes, or separate hubs per location.

Best practice: one practice-area hub (e.g., /family-law) with location spokes underneath (/family-law/san-francisco, /family-law/los-angeles). This preserves topical authority (all family law content clusters under one hub) while allowing location specificity (each spoke describes how family law applies in that jurisdiction). The alternative—separate hubs per location (/san-francisco/family-law, /los-angeles/family-law)—creates duplicated content and splits authority.

A variant is a hub-per-location with practice-area sub-hubs (/san-francisco, /san-francisco/family-law). This works if locations are distinct enough to merit their own brand presence and if you have enough content per location-practice combination. But it requires more content to avoid thin pages. Most firms find the practice-area primary structure simpler and more scalable.

The key guardrail for multi-location sites: every location spoke must contain genuinely unique information. The San Francisco family law spoke discusses California divorce law, San Francisco courts (Superior Court, judges, courthouses), San Francisco demographics, and local attorneys. The Los Angeles spoke discusses California law (the same, since state law applies statewide) but different courts, judges, and demographics. Sharing state law is correct; sharing city specifics is incorrect and doorway-like.