Breadcrumbs Not Showing in Google? Check Mobile Changes, Markup and Indexing

Breadcrumbs missing from Google do not automatically indicate broken markup. Check the device first: Google removed breadcrumb paths from mobile search results in January 2025, while desktop search still supports them. For desktop problems, separate page access and indexing, BreadcrumbList fields, and the accuracy of the hierarchy. Passing a test does not guarantee display.

This guide concerns the path shown next to a domain in search results. Your site's own breadcrumb navigation serves a different purpose: helping visitors reach a parent category. A change to Google's interface does not remove that navigation or make it useless.

Mobile-only absence may be expected

Google's January 23, 2025 announcement simplified the visible URL on mobile to the domain. It explicitly said that breadcrumb markup remains supported for desktop and that site owners do not need to remove it. The current Breadcrumb documentation lists desktop as the supported device category.

Save the query, observation date, device, language, region and result URL. A missing mobile path is not evidence that Google failed to understand your page. A path appearing for one desktop query is not a promise that it will appear for every query. Record functional site navigation, recognized markup and actual search appearance separately.

Observation	First check	Do not assume
Mobile shows only the domain	The announced mobile interface change	BreadcrumbList is broken
No desktop path; a test reports a critical error	The specific issue and delivered page output	Every reported issue is a JSON syntax error
Valid test; still no desktop path	Indexed version, crawl date, canonical and content guidelines	More levels or repeated submissions will force display

Make sure the versions being compared are the same

In Search Console URL Inspection, record indexed status, last crawl, page fetch, indexing permission and the user-declared and Google-selected canonicals when provided. Then run the live test. Indexed data describes a previously processed version; the live test checks current access and cannot predict Google's canonical selection or certify that a page is indexed.

If the page is not indexed, use the specific status to choose the next investigation. The discovered versus crawled indexing guide explains those branches. If Google selected another URL, compare the actual pages and inspect canonical declarations, redirects, internal links and the sitemap. See canonical meaning and verification for the implementation differences.

Google's canonicalization guidance describes preferences rather than commands. A shorter URL, a parameter or a path that differs from the breadcrumb labels does not alone establish a mistake. You can repair a known markup defect before indexing succeeds; simply do not claim that the repair has resolved the separate indexing cause.

Check the anonymous response, redirects, substantive HTML, robots.txt and indexing instructions in HTML and HTTP headers. A 200 response may contain a login screen or an empty shell. Robots.txt controls crawling rather than reliably preventing indexing, and can stop Google from reading a noindex instruction. Mark unavailable report fields as unavailable, not as evidence that everything is correct.

Check the fields without rejecting permitted omissions

Schema.org defines BreadcrumbList as an ordered chain of linked pages. Google's search feature imposes more specific requirements: a BreadcrumbList containing at least two ListItems in itemListElement, with the required properties. This does not mean every site's navigation needs an arbitrary number of levels.

• position: use integers; Google's documentation identifies 1 as the beginning. Consecutive numbering keeps the example below easy to maintain. Do not turn Schema.org's general ordering description into a claim that every gap is automatically rejected.
• name: identify the breadcrumb shown to the user. A concise label can refer to the same page as a longer H1. If item is a Thing that includes a name, Google allows the name on the ListItem itself to be omitted.
• item: specify the page URL directly, or use a Thing with @id in JSON-LD. Keep the identity and property nesting clear.
• The final item: its item property may be omitted. Google then uses the containing page's URL. This exception does not apply to every earlier item.

Google also says that neither the site's top-level path nor the current page itself must be included as a ListItem. A missing Home item or a trail ending at a parent is therefore not automatically an error. If you omit the current-page entry and the last entry represents a parent category, explicitly supply that parent's item URL. Otherwise, the containing-page fallback can pair a parent label with the current article's URL.

A working relationship does not require matching URL directories

The fictional teaching trail below is “Support → Valve guides → Manual valve check.” Its current page is https://example.test/manual-valve-check/ and its category is https://example.test/guides/valves/. The current URL does not contain every navigation level; that is compatible with a real user path.

{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    {
      "@type": "ListItem",
      "position": 1,
      "name": "Support",
      "item": "https://example.test/support/"
    },
    {
      "@type": "ListItem",
      "position": 2,
      "name": "Valve guides",
      "item": "https://example.test/guides/valves/"
    },
    {
      "@type": "ListItem",
      "position": 3,
      "name": "Manual valve check"
    }
  ]
}

To deploy an equivalent, an authorized template, CMS field or plugin must actually emit the JSON inside an application/ld+json script. This article displays teaching code; it does not activate the fictional trail on this tutorial. Replace all names and URLs before deployment. The example.test domain is reserved for this local example, not a public documentation site.

The final “Manual valve check” entry omits item and represents the containing teaching page. Keeping only the first two entries still meets the quantity requirement, provided the explicit Valve guides URL is retained. Alternatively, the first entry can use a Thing containing @id and name and omit the ListItem-level name. Review the entity relationship, rather than requiring every valid implementation to look exactly like this simplified example.

The image combines actual local browser captures before and after keyboard navigation. Tab focuses Valve guides; Enter opens its category; the category entry links back to the guide. Local intercepted responses supplied these fictional pages. Both languages were exercised at 1440, 1024, 820, 768, 390 and 320 pixels. This was not a public HTTP test, a Google tool run or a customer project.

Separate fixed-case checks covered the full trail, omission of the current entry while keeping the parent URL, a named Thing, and two genuine paths in the teaching hierarchy. A missing nonfinal item and a repeated position were negative cases. That checker tests these chosen fixtures; it is not a complete Schema.org or Google eligibility validator.

Multiple trails are supported; conflicting identities need repair

Google's multiple-trail example allows more than one way to reach the same page. Check whether each path exists and whether names identify the correct destinations. Counting two BreadcrumbLists or two script blocks is not a sufficient reason to delete either.

A real conflict would be a theme linking Valve guides to its category while an old plugin assigns the same label to an unrelated promotion. A migrated category with a stale destination is another actionable defect. Trace the producer—theme, plugin, template or API—and repair the relevant data source. Do not remove an entire data graph that also contains useful Article or Product entities.

Google recommends a typical user path rather than mirroring URL structure. Do not migrate established URLs merely to add breadcrumb directories. A parameterized URL also needs a purpose-based assessment; it is not automatically invalid. Decide whether filters, promotions and tags represent appropriate, stable navigation destinations before treating them as parents.

Inspect output after templates, JavaScript and caches

Save the same URL's initial response and rendered DOM. Find BreadcrumbList and inspect the names, positions, destinations and surrounding content. Include RDFa and Microdata in the inventory. In a CMS, the visible trail and machine-readable data may come from different components, so changing a theme label may leave a plugin's output untouched.

Google's JavaScript structured-data guide says it can process markup available in the rendered DOM. Absence from initial HTML is not automatic failure; presence in your browser is not proof that Google received the same resources. Test the actual URL, examine rendered HTML and investigate failed scripts or data requests. The JavaScript rendering diagnostic develops that comparison without inventing a fixed injection deadline.

After deployment, compare origin and public responses and any relevant cache layers. General Schema validation, Google's Rich Results Test and URL Inspection answer different questions. Fix critical errors and review noncritical issues without treating them as identical. Google's structured-data policies also require truthful, relevant content: technical tools cannot certify every quality requirement or guarantee appearance.

Test the site's navigation as a visitor

Enter a deep page directly, follow its parent links and confirm that the category leads back to the content. Repeat with a keyboard and a phone. The trail should represent hierarchy rather than browser history. Make the current-page state clear and ensure earlier entries reach the pages they promise; do not fabricate links for category labels that have no page.

NN/g's usability guidance favors a clear primary visible trail and allows task-based shortening on mobile, while warning against tiny targets and excessive screen use. Those are interface recommendations, not a Google prohibition on multiple structured trails. The underlying relationships should agree; abbreviated labels and responsive presentation still need practical review.

There is no breadcrumb rule requiring five articles per category or 300 words per page. Keep a category when it helps visitors find relevant content. For issues spanning templates and content groups, Guangsuan's SEO service overview in Chinese describes the broader maintenance work.

Verify the repair without promising a fixed display date

1. Read back the delivered HTML and rendered page. Confirm that fields, navigation and destinations changed, including cached public responses.
2. Retest affected URLs. With the appropriate Search Console access, inspect the live result and validate the specific reported issue.
3. For a few important URLs, request a recrawl when warranted; for larger updates, maintain an accurate sitemap. Google's recrawl guidance specifies quotas and says repeat requests do not accelerate crawling. Timing and inclusion are not guaranteed.
4. Later, review the crawl date, indexed version and desktop search observations under recorded conditions. Track impressions, clicks and genuine enquiries separately.

Successful deployment, valid markup, a new crawl, search appearance and business improvement are different outcomes. Preserve the same sample URLs; record missing evidence as unavailable instead of zero. The Chinese version of this guide covers the same method and local example for a bilingual team.