SAP Commerce Cloud SEO: An Operator’s Guide
Quick answer. SEO on SAP Commerce Cloud is mostly template and configuration work. You control URLs through the CMS site and storefront route config, add canonical tags where the Accelerator does not emit them, keep Solr facet URLs out of the index, and manage titles and descriptions in Backoffice or SmartEdit rather than page by page. The platform ships no sitemap generator or redirect manager, so those are deliberate builds.
What matters most
- Product and category URLs are set by CMS site configuration, not by a plugin, so one change to the URL pattern rewrites every page on that site.
- Accelerator templates do not reliably emit canonical tags, so duplicate facet and pagination URLs persist until you add them in the head fragments.
- Content catalogues are versioned: an edit stays invisible to crawlers until the catalogue is synced and Solr is reindexed.
- Facet URLs are the main crawl leak on this platform; each facet needs a decision to index, canonicalise or block.
- Composable storefront SEO depends on a working SSR deployment, because crawlers that do not run JavaScript only see what the server sends.
What does SAP Commerce Cloud SEO actually involve?
SAP Commerce Cloud, formerly hybris, is a Java commerce platform with two storefronts in common use. The Accelerator is built on JSP templates and renders on the application server. The composable storefront, often called Spartacus, is an Angular application that reads product data through the OCC REST API. The SEO work is the same in principle, but the file you edit is different.
Four surfaces decide what a crawler sees. The CMS site configuration sets domains, URL patterns and which catalogue a site reads. Storefront templates write the head tags, breadcrumbs and structured data. Solr governs search results and facet links. The product and content catalogues hold the names, summaries and descriptions those templates read. None of these is an SEO plugin. There is no module to install that fixes the basics.
That shapes how you plan the work. Most SAP Commerce Cloud SEO fixes are small code or configuration changes owned by the technical team or the implementation partner. A merchandiser can improve copy all day, but the canonical tag, the sitemap and the facet rules sit elsewhere.
- CMS site and base store configuration: domains, URL patterns, catalogue assignment.
- Storefront templates: title, meta description, canonical, hreflang and JSON-LD output.
- Solr configuration: which attributes become facets, and how facet state appears in a URL.
- Product and content catalogues: names, summaries, descriptions and CMS page SEO properties.
Where to go next
Accelerator or composable storefront: what changes for SEO?
The Accelerator sends complete HTML from the server. The composable storefront sends an Angular application and needs the SSR engine running to produce complete HTML. Google renders JavaScript, but crawlers that do not run it, including most social previewers and many AI agents, only see what the server returns. If SSR is off or failing on a route, those clients get an empty shell.
URL handling differs as well. In the Accelerator the URL resolver maps an incoming path to a category or product page on the server. In the composable storefront routes are defined in front-end configuration and often sit behind a proxy or CDN. Both can produce clean URLs. Only one of them does so without front-end work.
| Concern | Accelerator storefront (JSP) | Composable storefront (Spartacus) |
|---|---|---|
| HTML delivery | Rendered on the Commerce application server | Angular app; complete HTML needs the SSR engine |
| URL patterns | Resolved server-side from CMS site configuration | Defined in front-end route config, often behind a proxy |
| Metadata source | CMS page SEO properties and catalogue attributes read by JSP | Catalogue data over the OCC API; titles set in the front end |
| Facet URLs | Query-string facets generated by Solr | Search state in route parameters; same duplicate risk |
| Sitemap and redirects | Custom controller or cron job; redirects at server or CDN | Front-end or CDN layer; nothing provided by default |
Where do SAP Commerce Cloud URLs and canonicals come from?
URLs are resolved at request time. Product pages usually sit under a category path with a product segment such as /p/, and category pages under /c/. The exact shape is configured per CMS site, so a multi-brand implementation can produce different patterns from one code base. Check each site’s configuration rather than assuming the pattern you saw on staging applies to production.
Duplicates come from three places. The same product is often reachable through several category paths. Facet state is carried in the query string, typically in the q=:relevance: format the Accelerator uses. Pagination adds another parameter. Together these produce thousands of near-duplicate URLs that a crawler will find and spend time on.
Canonical tags are where assumptions cause damage. Treat rel=canonical as work to verify rather than a feature to rely on: Accelerator templates do not emit it on every page type in every implementation. Where it is missing, add it to the head fragment for product, category and search templates, and point facet and paginated views at the clean category URL. Google’s guidance on consolidating duplicate URLs is the reference to follow.
Two more items have no default. Hreflang for country sites is hand-authored in the template or a CMS component; the platform does not build a hreflang cluster for you. Bulk redirect management is also absent from the Accelerator, so most teams keep the redirect map in the web server, the CDN or a custom resolver. If a partner tells you either is a setting, ask to see it.
How should you treat Solr facet and search URLs?
Solr returns facet links, and every combination of facet values is a distinct URL. A category with eight facets and a dozen values each generates more URLs than the catalogue has products. The decision has to be made per facet, not once for the site.
Three outcomes cover most cases. Index the facets that match real search demand and unique inventory, such as brand or size. Canonicalise the ones that help shoppers but carry no demand. Block the ones that only create combinations, such as sort orders, page numbers and price sliders.
Robots and noindex do different jobs. A Disallow stops crawling, but a URL already in the index can stay there. A noindex tag only works after a crawler fetches the page. So block the combinations you never want indexed, and use noindex, follow for facets you want removed after a crawl. Never apply both to the same URL, because a blocked page hides its own noindex tag.
The catalogue and the search index are separate stores. After a catalogue sync, reindex Solr, or facets and search results will lag behind the pages. A new facet attribute also needs to be present in the indexed type and requires a full index before it appears. Merchandisers who enable a facet in Backoffice and see nothing usually have an indexing problem, not a UI bug.
Where do titles, descriptions and structured data come from?
There are two layers. CMS pages carry SEO properties for page title, meta description and meta keywords, edited in SmartEdit or Backoffice. Product and category pages read from catalogue attributes: name, summary and description. Which layer wins depends on the template that renders the page.
Check the template before promising a change. A product page title is often assembled from the product name plus brand or category, and a category page title often comes from the category name alone. If the template ignores the meta description field, filling that field changes nothing on the page. This is the single most common reason a metadata project shows no result.
Headings are the other quiet problem. Shared CMS components such as breadcrumbs, promo slots and link lists can each output an H1, and they render on every page that holds them. Two H1s on a category page is common and easy to miss in review.
Structured data is not generated by the platform. Add Product and BreadcrumbList JSON-LD in the template or as a CMS component, and make sure price and availability match what the page displays. For thousands of SKUs, edit catalogue attributes with ImpEx or Backoffice in bulk. SmartEdit is a per-page tool and will not scale to a catalogue.
What breaks during an SAP Commerce Cloud migration or replatform?
Content catalogues are versioned, with staged and online copies. An edit is not live until the sync job runs, and crawlers only ever see the online version. A skipped sync on launch day looks exactly like a content problem, which is why it burns hours.
Solr is a separate index. Sync the catalogue, then run a full index. Pages will render correctly while search and facets return stale or empty results, and that gap is easy to miss when the team is checking pages by eye.
URL patterns are the largest single risk. Changing a site’s URL pattern at launch rewrites every product and category URL at once. Decide the pattern before go-live and hold a redirect map outside the platform, because the Accelerator gives you no editor for it.
Product code changes are the second largest. A PIM migration often assigns new codes, and on a path-based URL that moves the page. Keep the old code as an alias or redirect it. Non-production environments are the third: CCv2 multiplies them, so confirm every one is behind authentication or blocked, and that no staging host is indexable.
For a composable go-live, add three checks to the launch list: SSR running on every route, sitemap generated by the front end, canonical output verified in the server response rather than in the browser.
How do you measure SAP Commerce Cloud SEO honestly?
Search Console, set up per domain and per CMS site, is your primary source. The platform’s own reporting is merchandising reporting; it will not tell you how a template renders to a crawler. Pair it with server or CDN logs, which are the only way to see how much crawling goes to facet URLs. On CCv2 those logs are not in Backoffice, so ask for access early.
Some things are genuinely hard to measure, and saying so is more useful than a dashboard. Crawl waste on facets is inferred from logs, not reported. Citations from AI assistants are not attributable in any reliable way yet. SSR failures on individual routes show up as traffic loss long before they show up as an alert.
What you can do is make the work reviewable: a canonical tag confirmed in the raw HTML, a facet list with an explicit decision for each, a sitemap that reflects the online catalogue, and a redirect map under version control. If you want a second pair of eyes on those four, our SAP Commerce Cloud practice and the audit below are the places to start.
Frequently asked questions
Does SAP Commerce Cloud have built-in SEO features?
Partly. Every CMS site has configurable URL patterns, CMS pages carry SEO properties for title and meta description, and the storefront will serve a robots.txt if you put one there. What it does not ship with is an XML sitemap generator, a redirect manager, or canonical tags guaranteed on every template. Plan for those as small builds rather than settings.
Should I block facet URLs in robots.txt or noindex them?
Both, for different cases. Use Disallow for combinations you never want crawled, such as sort orders and page numbers. Use noindex, follow for facets you want removed from the index, because a crawler has to fetch the page to see the tag. Never apply both to the same URL, since a blocked page hides its own noindex tag. Then canonicalise the facets you keep.
Can a merchandiser change a meta title without a developer?
For CMS pages, usually yes. SmartEdit and Backoffice expose SEO properties if your permissions allow it. For product and category pages, the title is often assembled in the template from the product name, brand and category, so the field you edit may not be the one that renders. Ask to see the template logic before promising a change.
Do I need to reindex Solr after changing product data?
Yes, if the change should affect search or facets. The catalogue and the Solr index are separate stores. A catalogue sync makes the change live on the page, and a Solr update or full index makes it visible in results and facet counts. A new facet attribute also needs to be in the indexed type and requires a full index.
Is the composable storefront bad for SEO?
Not inherently. It produces the same HTML as any Angular application, which works when the SSR engine is running and fails when it is not. Google renders JavaScript, but crawlers that do not, including most social previewers and many AI agents, only see what the server sends. Treat SSR as a production requirement and monitor it after every deployment.
Sources
- SAP Commerce Cloud product documentation on the SAP Help Portal
- Google guidance on consolidating duplicate URLs with canonical tags
- Google guidance on JavaScript SEO and server-rendered HTML
- Google introduction to robots.txt and crawling control
Get an SAP Commerce Cloud SEO audit
We check the things this article names: canonical output in the raw HTML, facet decisions, catalogue sync and index state, and the redirect map. You get a prioritised list your technical team can act on.


