Entities and Topical Architecture: Give Every Search Task One Canonical Owner

Direct answer

Build topical architecture by mapping real user tasks to the entities, attributes, relationships, and evidence needed to complete them. Then assign each distinct task one canonical page owner. Combine close query variants when they need the same answer. Create a separate page only when the user, outcome, evidence, or format is materially different.

An entity is a specific thing or concept, such as Google Search Console, crawl budget, SEOryon, an organization, a person, or a software product. A keyword is a wording people use to refer to an entity or task. Clear entity relationships help readers and machines understand the page, but there is no public "topical authority score" that you can fill by publishing hundreds of loosely related articles.

The winning architecture is usually the smallest complete set of useful canonical pages, supported by logical navigation, breadcrumbs, contextual links, and consistent identity data. If two pages compete for the same task, merge, differentiate, or redirect them. Do not keep both merely to preserve page count.

What you will be able to do

By the end, you will be able to:

• distinguish an entity, query, topic, and user task;
• define primary and mentioned entities for a page;
• map required attributes and relationships;
• choose between a section and a separate page;
• design hubs, lessons, tools, templates, and commercial pages without duplication;
• diagnose content cannibalization with evidence;
• assign canonical ownership and merge, retain, or redirect decisions;
• connect editorial coverage to user and business outcomes.

Entity, keyword, topic, and task are not synonyms

Entity

A uniquely identifiable thing or concept with attributes and relationships.

Examples:

• Google Search Console, a software service;
• SEOryon, an organization and software brand;
• Amaury, a person and author;
• crawl budget, a technical concept;
• France, a country.

An entity can have aliases. "GSC" and "Google Search Console" can refer to the same product.

Keyword or query

The words a searcher enters. Different queries can express the same task:

• seo audit checklist
• how to audit a website for seo
• website seo audit steps

Topic

A subject area broad enough to include multiple tasks. "SEO audit" may include a definition, procedure, template, tool, prioritization method, service, and case study.

User task

The outcome the person is trying to achieve. Query wording is evidence about the task, but not the entire definition.

Architecture begins with tasks because pages exist to help people do something.

Primary and mentioned entities

Every page should have one primary entity or tightly bound entity set. Supporting entities explain it.

For this lesson:

• primary entity: topical architecture;
• closely bound concepts: entity SEO and canonical page ownership;
• mentioned entities: Google Search, Schema.org, internal links, breadcrumbs, SEOryon;
• not a primary entity: every tool named in an example.

This distinction controls depth. Define and evidence the primary entity fully. Mention a supporting entity only as far as it helps the page's task, then link to its canonical owner.

Attributes and relationships create meaning

A useful entity map does more than list nouns.

For a SoftwareApplication, relevant attributes might include name, purpose, operating model, features, pricing basis, supported markets, and current availability. Relationships might include publisher, alternatives, integrations, target audience, inputs, and outputs.

For crawl budget, attributes include applicability thresholds, capacity, demand, and observable symptoms. Relationships include Googlebot, server health, URL inventory, sitemaps, and indexing.

Ask:

1. Which entity must the reader understand?
2. Which attributes determine the decision?
3. Which relationships explain cause or context?
4. Which evidence verifies them?
5. Which page owns each deeper subtask?

The entity-task architecture canvas

Download the Entity-Task Architecture Canvas. It forces every proposed URL to justify its task and owner.

Hubs, pillars, supporting pages, and sections

These labels are useful only if they describe navigational roles.

Hub

A curated entry point that helps a person choose the right path. An Academy hub can group lessons by skill level and learning sequence.

Pillar page

A broad canonical resource for a coherent task or concept. It should not become an oversized index that gives every subtask a shallow paragraph.

Supporting page

A page that solves a narrower distinct task and links naturally to related canonical owners.

Section

The right format when a sub-question is necessary to complete the current page but does not justify a separate destination.

Tool or template

A utility-first resource. Its main intent may be using or downloading, not reading a long guide.

The architecture should reflect these roles in navigation and internal links. You do not gain authority by calling everything a cluster.

Decide whether to create a page or a section

Create a separate page when several of these are true:

• the audience is different;
• the desired outcome is different;
• the evidence set is substantially different;
• the format is different, such as tool versus lesson;
• the page can stand alone and remain useful;
• search results reveal a distinct task;
• the business next action is legitimately different;
• the subject needs enough depth that inclusion would derail the parent.

Keep it as a section when:

• it is a follow-up needed to complete the current task;
• the answer is short and dependent on parent context;
• separate results would repeat the same definition and evidence;
• the proposed page exists only for a keyword variant;
• no distinct user or completion event exists.

There is no keyword-volume threshold that makes this decision automatically.

Canonical page ownership

Canonical ownership is an editorial and architectural decision before it is an HTML tag.

For every task, record:

• canonical owner URL;
• primary intent;
• audience;
• differentiating promise;
• required evidence;
• supporting pages;
• pages that overlap;
• merge, retain, redirect, or noindex decision;
• internal-link sources;
• business outcome;
• owner and review date.

The rel="canonical" element is not a reliable way to keep two competing editorial pages alive. If pages are functionally duplicative, consolidate their content and links, then redirect the retired URL when appropriate.

What content cannibalization really is

Cannibalization is not simply two pages mentioning the same keyword. Healthy sites repeat concepts.

It becomes a problem when multiple URLs compete to serve the same task and create harmful outcomes such as:

• search systems alternating between weaker and stronger pages;
• external links and internal links splitting across duplicates;
• users landing on an outdated or incomplete version;
• teams updating one page while another remains stale;
• analytics fragmenting the same outcome;
• crawl and index inventory growing without added utility.

Diagnose with evidence

1. group queries by user task, not exact string;
2. inspect which URLs receive impressions for that task over time;
3. compare intent, content, links, freshness, conversion, and canonical signals;
4. inspect the live search results and site navigation;
5. decide whether the pages are distinct or duplicative;
6. choose a canonical owner;
7. merge useful material and redirect where appropriate;
8. update links, sitemaps, breadcrumbs, and reporting;
9. monitor the task cluster, not one keyword.

Two pages ranking for adjacent queries can be a strength when each completes a different task. Do not consolidate automatically from a keyword-overlap tool.

Worked example: 20 SEO audit variants become five useful owners

A brainstorm produces 20 proposed articles:

• what is an SEO audit;
• SEO audit meaning;
• website audit definition;
• how to do an SEO audit;
• technical audit steps;
• SEO audit checklist;
• free SEO audit template;
• SEO audit spreadsheet;
• how to prioritize SEO findings;
• SEO issue priority;
• SEO audit tool;
• automated SEO audit;
• SEO audit software;
• SEO audit service;
• hire SEO audit agency;
• SEO audit cost;
• SEO audit example;
• SEO audit case study;
• enterprise SEO audit;
• SaaS SEO audit.

Publishing 20 articles would create repetition. Map them by task.

Owner 1: learn the concept

/academy/seo-audit-fundamentals/

Covers meaning, scope, outputs, limitations, and how technical, content, authority, and measurement work together. "SEO audit meaning" and "website audit definition" belong here as sections or aliases.

Owner 2: execute the procedure

/guides/technical-seo-audit/

Covers the ordered workflow, required access, crawling, indexing, rendering, performance, evidence capture, and validation. Enterprise and SaaS differences can be sections unless each develops a genuinely distinct procedure and evidence set.

Owner 3: use the template

/tools/seo-audit-checklist/

Provides the interactive or downloadable asset, field definitions, usage instructions, version, and sample. It links to the full procedure rather than repeating it.

Owner 4: prioritize findings

/guides/prioritize-seo-issues/

Solves the post-audit decision using impact, confidence, effort, security, revenue risk, dependencies, and recovery. This is a distinct task from finding issues.

Owner 5: evaluate a product or service

/product/seo-audit/ or another truthful commercial route.

Explains SEOryon's real feature scope, limits, data handling, pricing state, and fit. A service page should exist only if SEOryon genuinely offers that service. Do not create it to capture a commercial keyword.

What happens to the case study?

A real case study with customer permission, baseline, intervention, period, confounders, and outcome can deserve its own page. An invented case cannot. A worked synthetic example belongs inside a lesson and must be labeled.

The resulting links

The fundamentals lesson links to the procedure. The procedure embeds the checklist and links to prioritization. The tool links back to instructions. Product CTAs appear only where relevant. Breadcrumbs expose hierarchy, but contextual links express task relationships.

This five-owner system has more depth and less duplication than 20 keyword pages.

Entity clarity on the page

Use consistent names, definitions, URLs, and relationships.

For an author:

• display one consistent name;
• link every byline to the canonical author page;
• describe the real role and experience;
• reference one stable Person entity;
• do not invent awards or profiles.

For an organization:

• use the actual brand name;
• maintain a canonical about or organization identity;
• keep logo, URL, contact, and product relationships current;
• distinguish the company from the software application when necessary.

Schema.org can express these relationships, but the Schema.org documentation does not create factual authority. Visible pages and external evidence must support the graph.

Internal links, breadcrumbs, and navigation

Google's SEO Starter Guide explains that links help users and search engines discover and understand pages. Architecture should make important owners reachable through ordinary crawlable links.

Use:

• primary navigation for durable high-level destinations;
• hub pages for choice and sequence;
• breadcrumbs for hierarchy;
• contextual links for cause, prerequisites, comparisons, and next steps;
• related resources for a small set of genuinely useful continuations;
• sitemaps as discovery support, not a replacement for navigation.

Avoid orphan pages, JavaScript-only interactions without crawlable links, generic anchors, and footer grids containing every target phrase.

Topical authority: useful shorthand, dangerous score

People use "topical authority" to describe comprehensive expertise and a site's recognized usefulness around a subject. That can guide editorial quality. It becomes misleading when treated as a hidden public score with a formula such as "publish 100 articles to unlock rankings."

Google's ranking systems guide documents systems and signals at a broad level, not a marketer's topical-authority metric. Its helpful content guidance favors a clear site purpose and substantial value.

Use an evidence-backed coverage model:

• are the important user tasks covered?
• does each have a strong canonical owner?
• are primary entities and relationships clear?
• is evidence original and maintained?
• can users navigate the learning or buying journey?
• are pages earning qualified visits, links, citations, product use, or learning completion?

Delete or merge pages that do not earn a distinct role.

Architecture governance

Every proposed page enters a route-and-intent register before drafting. The editor checks ownership. The product team confirms the outcome and CTA. The evidence owner confirms source availability. Technical implementation checks route, canonical, locale, structured data, sitemap, and links.

For a multi-locale site, do not create hreflang equivalents until translations are complete and genuinely equivalent. Locale paths need consistent ownership without automatic thin duplication.

Review the map quarterly and after product changes. Log redirects and preserve them. Make import jobs idempotent so rerunning a content migration does not create duplicate routes. Validate route uniqueness in the build.

Exercise: consolidate a proposed cluster

Take ten proposed URLs around one topic and complete the canvas. For each, write:

1. user task;
2. primary entity;
3. required evidence;
4. proposed canonical owner;
5. page versus section decision;
6. overlapping existing URL;
7. merge, retain, or redirect action;
8. internal links;
9. user and business outcome;
10. confidence and validation plan.

Evaluation rubric

Any architecture that preserves a known duplicate without a reason remains incomplete.

Common mistakes

Publishing one page per keyword

Queries are observations, not an automatic URL list.

Treating entity SEO as adding schema

Entity clarity comes from accurate content, identity, relationships, evidence, and architecture. Markup is one representation.

Building a giant pillar with shallow sections

Breadth without task completion is not comprehensive.

Consolidating pages merely because they share words

Different tasks can mention the same entity. Diagnose intent and outcomes.

Keeping old pages to avoid losing page count

Page count has no intrinsic value. Preserve useful content and signals through careful consolidation.

Measuring only traffic

Learning completion, tool use, qualified signup, retention, links, citations, and reduced support demand may be better outcomes.

Final checklist

• Each proposed page has a defined user task.
• The primary entity is explicit.
• Required attributes and relationships are mapped.
• Evidence needs are known before drafting.
• Query variants serving the same task are consolidated.
• A separate page has a distinct outcome, audience, evidence set, or format.
• Every task has one canonical editorial owner.
• Existing overlaps have merge, retain, or redirect decisions.
• Canonical HTML signals agree with editorial ownership.
• Hubs help users choose rather than duplicate every answer.
• Breadcrumbs reflect real hierarchy.
• Contextual links express prerequisites and next tasks.
• Important pages are reachable through crawlable links.
• Author, organization, and product entities remain distinct and consistent.
• No hidden topical-authority score is presented as fact.
• User and business outcomes are assigned.
• Route uniqueness and redirects are tested.
• The architecture has an owner and review date.

Frequently asked questions

What is entity SEO?

It is the practice of making specific things, their attributes, and their relationships clear through accurate content, architecture, links, identity, and appropriate machine-readable data. It is broader than schema markup.

What is topical authority?

It is useful shorthand for demonstrated depth and reliability around a subject. Google does not publish a simple topical-authority score or a required number of articles.

Should similar keywords have separate pages?

Only when they represent materially different tasks, audiences, evidence, formats, or outcomes. Close wording variants often belong on one canonical page.

How do I fix keyword cannibalization?

Group queries by task, identify competing URLs, compare their purpose and performance, choose one owner, merge valuable material, redirect retired duplicates, update links, and monitor the task cluster.

Does a topic cluster need one pillar page?

Not always. It needs clear navigation and canonical ownership. Use a hub when it helps users choose; use a broad pillar only when it can complete a coherent task.

Sources and methodology

Learner research showed recurring confusion about cannibalization, topical authority, entities, and where blog or guide pages should live. Those questions shaped the answer-first ownership model. No community claim is used as factual evidence.

1. Google Search Central, In-depth guide to how Google Search works, checked 31 July 2026. First-party overview of crawling, indexing, and serving search results.
2. Google Search Central, SEO Starter Guide, checked 31 July 2026. Official site organization, links, URLs, and helpful page guidance.
3. Google Search Central, A guide to Google Search ranking systems, checked 31 July 2026. Official overview, used to avoid inventing a public topical-authority score.
4. Google Search Central, Creating helpful, reliable, people-first content, checked 31 July 2026. Official guidance on site focus, first-hand expertise, substantial value, and people-first purpose.
5. Schema.org, Documentation, checked 31 July 2026. Vocabulary reference for entities and relationships, not evidence of factual authority or ranking.

Previous: Information Gain and Evidence Engineering
Next: /academy/internal-linking-navigation/ is the planned Step 16 route and should remain unpublished until its full lesson exists.