On-Page SEO: Complet Guide

A practical map of lune page-level signals vous contrôler, ce que chaque un peut influence, and où to trouver le site's detailed implementation guides.

Première publication : 18 juil. 2026 · Dernière mise à jour : 3 août 2026 · Advanced
Langues
1 indice probant sur cette page

On-page SEO is lune page-level fonctionner vous contrôler: the réponse and evidence in the principal content, HTML structure, headings, lien internes, titles and snippets, images, and données structurées. Ne faites pas collapse ceux elements into un liste of ranking factors. Some aider a search system match and rank une page; some faire lune page eligible to be indexé or affiché in a particulier fonctionnalité; some influence how le résultat is presented; and some primarily aider personnes utiliser and comprendre lune page. Commencer by making lune page's objectif and réponse clair, alors vérifier que the rendered HTML exposes the intended content and signals. Utiliser ce guide to choisir the correct subguide, and utiliser the separate on-page checklist quand vous are ready to execute an audit.

TL;DR — Treat une page as a contract entre intent, evidence, rendered HTML, retrieval signals, presentation contrôle, and human usability. Diagnose the failed couche avant modification copy. Separate candidate eligibility from relevance, ranking, result assembly, and accessibility. Maintain un owning URL per task, route specialized implementation to the existing subcluster guides, and validate les deux source/rendered output and observed search behavior.

Model lune page as a layered contract

A utile on-page examiner follows lune page via several layers:

  1. Task ownership: qui audience, question, and outcome ce URL owns.
  2. Réponse: the direct réponse, traiter, evidence, exemples, and limitations.
  3. Information structure: sections, headings, listes, tables, and relationships.
  4. HTML exposure: ce que apparaît in the initial réponse and rendered DOM.
  5. Search contrôle: title sources, snippet contrôle, index directives, canonicals, and pris en charge données structurées.
  6. Connections: lien internes into and out of lune page, with utile anchor text.
  7. Media: discoverable images and video, accessible alternatives, and performances.
  8. Observed outcome: indexation, requête fit, result presentation, usage, and business behavior.

A échec décrit as “on-page” peut begin in quelconque couche. A manquant réponse may be an editorial gap. A manquant réponse in rendered HTML may be a rendering defect. A correct title element que n’est pas affiché may be a result-assembly choice plutôt que a broken tag.

Separate the gates

The la plupart important avancé distinction is entre entering a candidate définir and winning dans it.

CoucheTypical questionEvidenceNe faites pas conclude
Discovery/accèsPeut the system reach l’URL and resources?liens, réponse, robots, renderque accès signifie indexation
Index/eligibilityPeut lune page or fonctionnalité be considéré?canonical/index state, pris en charge markup, policiesque eligibility guarantees afficher
RelevanceFait le contenu réponse ce requête or subtask?query-page comparison, passage coverageque relevance alone determines rank
Ranking/rerankingQui eligible candidates are preferred?observed results and controlled testsa fixed public weight pour un element
PresentationQui title, snippet, image, or fonctionnalité is affiché?live result and Search Console contextque supplied metadata is toujours utilisé
AccessibilityPeut personnes operate and comprendre lune page?manual and assistive-technology testingque an SEO robot d’exploration proves conformance

The practical benefit is meilleur prioritization. Si lune page is non indexée, rewriting the meta description is downstream of the réel problem. Si a rich result is absent, premier establish eligibility, alors remember que Google ne fait pas guarantee afficher.

Design a content contract

Pour every important page class, define:

  • the owning intent and excluded intents;
  • l’URL canonique and attendu index state;
  • the requis réponse blocks and evidence owners;
  • acceptable freshness and examiner triggers;
  • title and heading generation rules;
  • requis internal-link relationships;
  • media and alternative-text rules;
  • structured-data eligibility and visible-content dependencies;
  • validation checks and accountable team.

Ce is plus durable que a one-time score. It aussi rend template regressions testable avant publication.

Route the implementation to its owner

Header system

The header-tags hub owns heading levels, H1 questions, hierarchy, and navigation implications. Utiliser headings to expose a logical outline and utile section étiquettes; ne faites pas invent a ranking-weight ladder pour H1–H6.

HTML system

The HTML SEO hub owns source-versus-rendered HTML, semantic HTML, crawlable anchors, language attributes, and parsing échecs. Google’s lien guidance is explicit à propos de crawlable anchor markup and utile anchor text.

Image system

The image SEO hub owns image discovery, responsive delivery, page context, filenames, texte alternatif, and image performances. Google’s current image guidance distinguishes discoverable HTML image elements from CSS background images and connects texte alternatif with les deux image understanding and accessibility.

Metadata system

The meta-tags hub owns title and snippet inputs, robots metadata, snippet limites, and social metadata. Google may assemble title liens from several page signals and usually builds snippets from page content, parfois en utilisant the meta description. Supplied text is an input, pas an instruction que doit be affiché.

Evidence for this claim Google primarily creates snippets from page content and may use the meta description when it better describes the page for a query. Scope: production Confidence: high · Verified: Control your snippets in search results

Structured-data system

The structured-data hub owns schema.org vocabulary, formats, feature-specific requirements, and validation. The Schema.org documentation defines the shared vocabulary, pendant que Google’s structured-data introduction frames markup as standardized clues pour understanding and pris en charge search fonctionnalités; it ne fait pas turn markup into a general ranking guarantee.

Manage overlap and cannibalization

Two pages peut mention the même entity sans competing. The problem is ambiguous ownership of the même reader task.

Utiliser a simple ownership record:

FieldExemple
Owning taskExpliquer the on-page system and route to detailed guides
Principal audienceSomeone deciding ce que kind of on-page fonctionner is nécessaire
Requis réponsescope, mental model, library map, diagnostic routing
Explicit exclusionstep-by-step audit execution
Handoffon-page SEO checklist

Quand two URLs apparaître to propre the même task, choisir an owner, narrow the autre page, strengthen the linking relationship, and confirmer the titles and introductions reflect the distinction. Ne faites pas merge pages solely parce que a outil reports overlapping words.

Tester changements as hypotheses

An on-page edit devrait state:

  • Problem: ce que observed behavior is incorrect?
  • Couche: task, content, HTML, eligibility, presentation, or accessibility?
  • Modifier: ce que unique material variable va be altered?
  • Attendu observation: ce que devrait modifier, où, and pour whom?
  • Window: quand va exploration, processing, and trafic cycles faire evaluation fair?
  • Guardrail: ce que utilisateur or business outcome doit pas obtenir worse?

Pas every page supports a causal SEO tester. Pour a low-traffic page, the honest result may be “implementation verified; performance effect not determined.”

Add an expert note

Pin an expert quote

New person? Create their unclaimed profile at /admin/experts/ → Pin a quote first.