Langues

Guide Sitecore SEO

How to faire a Sitecore site rank — the two platforms que modifier everything (XP vs XM Cloud), ce que SXA donne vous out of the box and ce que it doesn't, and the platform-specific traps que catch enterprise teams: the par défaut robots.txt que blocks every robot d’exploration, aliases que créer contenu dupliqué at HTTP 200, and metadata que silently fails to inherit parce que blank isn't NULL.

Première publication : 27 juin 2026 · Dernière mise à jour : 3 août 2026 · Advanced
Langues
AIRédigé parAI v2

Sitecore is an enterprise DXP que ships with almost aucun SEO functionality out of the box — and the two product lines behave very differently. Sitecore XP is traditional server-rendered .NET; XM Cloud (renamed SitecoreAI in Sitecore's current docs) is headless SaaS où a Suivant.js front fin (via JSS, now publié as Content SDK) contrôle the HTML and vous manage SEO with Suivant.js metadata APIs. SXA adds the basics (sitemap, robots.txt management, metadata fields) but everything encore nécessite configuration. Three Sitecore-specific traps catch teams: the par défaut robots.txt blocks every robot d’exploration si the robots field is left blank (vide ≠ permissive); item aliases serve HTTP 200 on les deux URLs, creating vrai contenu dupliqué que canonical doesn't fix automatically; and blank metadata fields don't fall back to Standard Valeurs unless 'Reset Blank' is enabled, parce que blank isn't NULL. On the enterprise scale Sitecore lives at, governance — a base SEO template, validation rules, environment contrôler — matters as beaucoup as quelconque unique setting.

TL;DR — Sitecore ships with almost aucun SEO functionality; SXA adds the basics but nécessite configuration. The two platforms diverge complètement: XP is server-rendered .NET (robots d’exploration voir finished HTML, personalization renders server-side); XM Cloud is headless, où a Suivant.js front fin contrôle the HTML and vous manage SEO with Suivant.js metadata APIs (generateMetadata, MetadataRoute) — with SSG/ISR the recommended rendering modes. The platform-specific traps: the par défaut robots.txt blocks tout robots d’exploration quand the field is blank; item aliases retourner HTTP 200 on les deux URLs (vrai duplicate content, aucun canonical fix by par défaut); and blank metadata fields don’t inherit Standard Valeurs unless “Reset Blank” is enabled (blank ≠ NULL). At Sitecore’s enterprise scale, governance — a base SEO template, validation rules, environment contrôler — is the réel fonctionner. (Naming remarque: Sitecore’s docs now appel the XM Cloud product SitecoreAI; the architecture ci-dessous is unchanged.)

Evidence for this claim The article's described sitecore-seo capabilities must be evaluated against the platform's current documentation rather than assumed to be search-engine behavior. Scope: Platform-specific capability documentation. Confidence: high · Verified: Sitecore: SEO Evidence for this claim Regardless of platform, Google needs crawlable URLs, accessible rendered content, descriptive metadata, and valid search directives. Scope: Google requirements independent of platform. Confidence: high · Verified: Google Search Central: SEO Starter Guide

Two platforms, two complètement différent SEO models

The unique la plupart important chose to establish avant touching anything: qui Sitecore are vous on? The product nom is the même; the SEO mechanics ne sont pas.

  • Sitecore XP (Experience Platform). Traditional .NET, server-rendered via Razor/MVC, deployed on-premise or in managed cloud. Googlebot généralement receives finished HTML. SEO is managed à l’intérieur Sitecore — in templates, Standard Valeurs, and requête pipelines.
  • XM Cloud. Headless SaaS. Sitecore stores le contenu and exposes it over a GraphQL API (Experience Edge); a decoupled front fin — almost toujours Suivant.js construit with Sitecore JSS (JavaScript Services) — renders the HTML. SEO lives in the Suivant.js application couche. Sitecore holds le contenu; Suivant.js contrôle the output.

A naming remarque si you’re checking Sitecore’s propre docs pendant que reading ce: as of mid-2026, Sitecore’s documentation site has renamed the composable SaaS product from “XM Cloud” to SitecoreAI — the old doc.sitecore.com/xmc/... and developers.sitecore.com/learn/accelerate/xm-cloud/... URL paths now redirection to .../sai/.../sitecoreai/... paths, and the platform’s propre landing page describes it as “a cloud-native, SaaS, hybrid headless digital experience platform.” The architecture ce article describes is unchanged — Experience Edge, GraphQL, a decoupled Suivant.js front fin — but the front-end SDK is now publié as Content SDK (ouvrir source, Suivant.js-first) plutôt que sous the JSS nom, and you’ll voir “SitecoreAI” plutôt que “XM Cloud” in current screenshots and menu étiquettes. I’ve kept “XM Cloud” in ce article parce que that’s encore the term la plupart practitioners search pour and la plupart existing implementations run on, but don’t be thrown si votre sitecore rep or the Cloud Portal now dit SitecoreAI — même platform, même SEO mechanics ci-dessous.

In les deux cas the practical baseline is the même. As Fishtank’s Marcel put it back in 2018, “Sitecore ships with virtually aucun SEO functionality (with the exception of SXA qui inclut some basics).” That’s encore the correct mental model. SXA (Sitecore Experience Accelerator) is the couche que donne vous a sitemap module, robots.txt management in le contenu tree, and standardized metadata fields. Sans it, tout of que is custom development.

Ce is squarely enterprise SEO territory. Sitecore implementations involve dedicated developers and solution partners, and — as I’ve written à propos de enterprise sites généralement — “the plus probable vous are to run into multiple tech stacks,” legacy systems, and split ownership à travers sections. Garder que in mind: on Sitecore, the fix pour an SEO problem is souvent a template modifier or a pipeline override owned by a developer, pas a setting vous flip yourself.

Metadata — and the blank-vs-NULL trap

SEO fields (page title, meta description, Ouvrir Graph tags) live in votre page données templates. The clean pattern is a base SEO template que every page template inherits from, so the fields exist everywhere. Ken Gray of Konabos notes the par défaut gap directement: “out-of-the-box, Sitecore’s données templates pourrait pas inclure some of the Meta Données fields.”

Standard Valeurs let vous définir sensible defaults pour ceux fields (Par exemple a $name token as a title fallback). But here’s le sitecore-specific trap que produces manquant metadata at scale. Sitecore’s propre docs are explicit: “Si the valeur of a field is NULL, the item contient the standard valeur pour que field as défini in the données template pour que item.” The catch is que a blank field n’est pas NULL. Quand a content author clears a meta description, it becomes blank — and a blank field fait pas fall back to the Standard Valeur unless “Reset Blank” is enabled on que field. Le résultat is pages rendering <meta name="description" content=""> au lieu de inheriting a par défaut. On a grand site, that’s thousands of vide descriptions nobody intended.

The fix is twofold: enable “Reset Blank” on metadata fields où a fallback is desired, and ajouter validation rules que enforce non-empty titles and descriptions (and character limites) so authors can’t silently ship blanks.

URL management and the alias trap

Sitecore generates URLs from le contenu-tree chemin by par défaut; clean URLs come from SXA settings or custom item resolvers.

The trap is item aliases — alternate URLs vous pouvez attach to quelconque item. Ils regarder harmless and ils ne sont pas SEO-neutral. As Dheer Rajpoot documented, “aucun redirection (aucun 301 or 302 Code d’état HTTP) se produit quand vous are en utilisant aliases in Sitecore,” which means “multiple URLs va be créé pour a unique page URL.” Les deux l’URL canonique and the alias retourner HTTP 200 with identical content — genuine contenu dupliqué, and Sitecore fait pas auto-emit a balise canonical to resolve it.

Vous have two governance-grade fixes:

  1. Override the AliasResolver in the HttpRequest pipeline to inject a balise canonical pointing at the réel URL, or
  2. Modify the alias pipeline to problème a 301 redirection au lieu de serving the alias directement.

My recommendation is the second: treat aliases as redirections, pas alternate accès URLs. Ken Gray’s advice s’applique broadly ici — “utiliser Sitecore’s canonical lien management to specify the preferred version of une URL” — but on aliases specifically, a 301 is cleaner que relying on canonical hints.

Sitemaps

On SXA, configurer le sitemap at site/Settings → Moteur de recherches Sitemap → Sitemap Mode. Two relevant modes: Stored in cache (the par défaut — regenerates dynamically; bon pour frequently mis à jour or Azure-hosted sites) and Stored in fichier (a static fichier; meilleur pour grand sites que modifier infrequently, avoiding regeneration overhead). SXA automatically appends the sitemap URL to robots.txt, and le sitemap lives at /sitemap.xml. A courant échec: si TargetHostName isn’t configuré, le sitemap renvoie a 404.

On XM Cloud + Suivant.js, generate le sitemap programmatically with MetadataRoute.Sitemap, querying Experience Edge over GraphQL — qui lets vous exclude non-indexable URLs at the application couche. Sitecore’s propre guidance notes que “Next.js offers built-in sitemap and robots.txt generation.”

Robots.txt — the par défaut que blocks everything

Ce is le sitecore trap with the highest blast radius. On SXA, robots.txt is configuré in le contenu tree (site Settings → Robots content field), and vous doit republish le site après a modifier. The gotcha, straight from Sitecore’s docs: “If no rules are added, the system writes: ‘User-agent: * Disallow: /’” — qui blocks every robot d’exploration. An vide robots field n’est pas a permissive robots.txt; it’s a site-wide block. Vous doit explicitly définir:

User-agent: *
Allow: /

On la plupart platforms a manquant robots.txt signifie “crawl everything.” On Sitecore it peut mean the opposite, so verifying production robots.txt is a non-negotiable launch step. On XM Cloud + Suivant.js, utiliser MetadataRoute.Robots pour a code-driven, type-safe robots.txt. Separately, CM (Content Management) instances and QA/staging environments devrait toujours disallow tout — seulement the CD (Content Delivery) production instance devrait be crawlable.

Multilingual, hreflang, and multisite

Sitecore stores language versions as separate versions of the même item, pas separate items. Que, combined with language fallback (item-level or the plus granular field-level, e.g. an es-MX → es-ES → en chain), peut serve the même content at multiple language URLs — a duplicate-content vector unless hreflang signals the relationship. And hreflang is pas automatic in standard Sitecore; it has to be ajouté to templates (SXA peut generate it quand configuré), with complet absolute URLs, bidirectional références, and an x-default. As John Mueller put it (cited by Jakub Koba), “TBH hreflang is un of the la plupart complex aspects of SEO (si pas the la plupart complex un).” Don’t underestimate it.

Sitecore aussi natively supports multisite — nombreux sites in un install, parfois sharing content. Shared content à travers site boundaries nécessite a deliberate canonical strategy, and chaque site nécessite its propre sitemap and robots.txt. Ce is exactly the kind of split-ownership, shared-infrastructure complexity I flagged in enterprise SEO technique: “Parfois différent personnes are responsible pour différent sections of the website or même différent pages, qui peut faire maillage interne time-consuming.”

Headless rendering strategy (XM Cloud)

Un layer-ownership point worth being explicit à propos de, since it trips personnes up: XM Cloud environments have les deux an editing host and a rendering host, and ils ne sont pas the même chose. Sitecore’s propre docs draw the line clearly — the editing host exists seulement to power the WYSIWYG editing experience à l’intérieur Page Builder/Design Studio, “is not set up or scaled for serving live traffic,” and sees internal (author-only) trafic. The rendering host is the public-facing Suivant.js app — hosted on Vercel, Netlify, or Azure, consuming content from Experience Edge, and scaled to serve réel site visitors. Robots d’exploration seulement ever hit the rendering host; the editing host’s rendering behavior (or an editing-host URL leaking into a sitemap or getting lié to) n’est pas representative of ce que Googlebot voit and isn’t something moteur de recherches devrait be able to reach at tout.

On XM Cloud the rendering mode vous choisir in Suivant.js is the SEO decision. Akshay Sura (Konabos) lays out the four options cleanly:

StrategySEO impact
SSG (static)Meilleur — “SSG pre-renders HTML at build time… Search engines can easily crawl the pre-rendered HTML.”
ISR (incremental static regen)Bon — static performances with background freshness; recommended pour content at scale
SSR (server-side render)Bon — “Fully rendered HTML is ready for search engines to index.”
CSR (client-side render)Worst — “Search engines may struggle with indexing JavaScript-rendered content.”

The recommendation: SSG or ISR pour le SEO-critical content; reserve CSR pour interactive UI seulement. Metadata is définir via generateMetadata (App Router). David Austin (Fishtank) notes a performances nicety: “Tout récupérer calls dans generateMetadata are memoized, meaning identical URLs are seulement récupéré une fois à travers the application, preventing redundant requêtes.” And as Sebastián Aliaga puts it, “The dynamic approach is the meilleur méthode pour Sitecore Headless as you’ll be able to prendre what’s partie of lune page’s layout données and incorporate it.”

Personalization — don’t cloak

Sitecore’s personalization is a réel SEO consideration. On XP, personalization renders server-side, so Googlebot voit the par défaut (un-personalized) experience — assurez-vous que par défaut is complet and SEO-optimized, pas thin. On XM Cloud, client-side personalization (JSS) peut hide content from robots d’exploration que don’t run the JavaScript; pre-render the par défaut with SSR, or personalize at the edge.

The hard rule à travers les deux: jamais serve différent content to robots d’exploration que to utilisateurs — that’s cloaking, and it’s a guidelines violation. Jamais personalize the principal H1, FAQ content, or données structurées with client-side rules. Personalization devrait ajouter to, jamais replace, the canonical content.

Données structurées

Utiliser JSON-LD in <script type="application/ld+json">. On XP, render it from template fields in votre Razor view or via pipeline; on XM Cloud, model schema fields in templates, retrieve over GraphQL, and render in the Suivant.js component. Prioritize FAQPage, HowTo, Product, Article, and BreadcrumbList. Vous pouvez deploy via a tag manager — Martha van Berkel (Schema App) notes teams “typically utiliser JavaScript to deploy Balisage de données structurées to Sitecore… les deux efficient and scalable” — but be aware client-side injection signifie AI robots d’exploration may miss it, so préférer server-rendered JSON-LD pour maximum coverage. Peter Lambrou (Codehouse) frames the payoff simply: “Ajouter balisage de données structurées to lune page HTML to faire votre résultats de recherche apparaître plus attractive.”

Enterprise governance — où the réel fonctionner is

Unique settings aside, the SEO outcome on a big Sitecore install is decided by governance. I’ve said avant que enterprise sites are où SEO technique shines precisely parce que “enterprise sites peut have complex infrastructures and a lot of legacy systems in placer” — and “I doubt there’s a major website que is technically perfect.” The recurring Sitecore-specific governance items:

  • Template sprawl. Multiple templates pour the même objectif, chaque with différent (or manquant) SEO fields. Audit les; enforce a base SEO template que every page template inherits.
  • Metadata validation. “Reset Blank” plus field-level validation so authors can’t ship vide or over-length titles and descriptions.
  • Alias governance. A policy que aliases are seulement créé with a canonical override or as 301s.
  • Environment contrôler. CM/QA/staging disallowed; production explicitly allowed and verified.
  • Budget d’exploration at scale. Faceted navigation, language versions, and parameter URLs peut blow up l’URL space; curate sitemaps and govern robots/noindex tightly. (Voir budget d’exploration.)

Add an expert note

Pin an expert quote

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