Breadcrumbs and SEO: What They Still Do in 2026
Table of contents
Breadcrumbs are the small trail that tells you where you are inside a site: Home, Category, Article. Plenty of guides still sell them as a way to win extra space in Google, and that is no longer true. They still earn their keep, just for different reasons: they link, they order, and they orient.
What Breadcrumbs Are and Which Ones Matter for SEO
A breadcrumb is secondary navigation. It shows the hierarchical path from the homepage down to the page you are reading. It does not replace the main menu: the menu moves sideways (services, blog, contact), while the breadcrumb moves upward and always offers a one-click return to the start.
There are three types: attribute breadcrumbs (file path), history breadcrumbs (where the visitor came from) and hierarchical breadcrumbs (where the page sits in the site). Only the hierarchical ones matter for SEO: they describe a stable structure a search engine can read and that does not change with each session.
The idea behind this article: they are not decoration for the results page, they are real internal links and explicit hierarchy.
What Changed in January 2025 and What Didn't
On 23 January 2025 Google announced it would stop showing breadcrumbs in mobile search results. The stated reason was space: on a small screen the trail gets cut off and is less useful. On mobile only the domain remains. Desktop keeps the domain plus the breadcrumb, exactly as before.
Most SEO writing covered the flashy part and skipped the important one: the announcement said that if you use breadcrumb markup there is nothing to fix. BreadcrumbList is still supported, the breadcrumbs report in Search Console is still live, and the rich results test still validates it. The only thing lost was the visual trail on phones.
Worth killing a myth too: structured data does not improve rankings. Google's structured data policies state that a manual action for structured data "doesn't affect how the page ranks in Google web search", only its eligibility for a rich result. And even with correct markup, there is no guarantee the rich result will show.
The practical takeaway: if you added breadcrumbs only to look pretty in Google, the incentive just dropped. If you added them for structure, nothing changed at all.
The Three Jobs They Still Do for SEO
1. Real internal links
Every level of the trail is a normal link pointing to the parent category. That feeds internal linking with descriptive anchor text repeated across hundreds of pages: the real category name, not a generic "click here". It also reduces click depth for category pages, which are usually the ones carrying a site's commercial traffic.
2. An explicit, consistent hierarchy
The breadcrumb and the URL should tell the same story. If the URL is /blog/technical-audit and the trail says "Home, Services, Audit", the message contradicts itself. Google recommends that a breadcrumb represent a typical user path to a page rather than mirroring the URL structure. That is a sensible recommendation: breadcrumbs are navigation, and navigation should make sense to a person. If your URLs and slugs are already clear, the trail almost writes itself.
3. Fewer orphans, better orientation
Every page with a breadcrumb has at least one navigable path back. That helps against orphan pages and gives context to anyone landing on a deep page from a search result: without touching the back button, they know which section they are in and what else lives there.
One caveat about AI assistants: they read HTML too, and a readable hierarchy helps them understand the document's context, but there is no evidence Google uses breadcrumbs to generate answers. Do not promise that in a proposal.
Writing the Trail in HTML Without Breaking Accessibility
Four rules before the code. Use two to four levels: more does not fit and does not get read. Keep labels short, with the real section name. Never start with the full domain. And the current page goes last and does not link to itself: a link to where you already are is noise for keyboard users.
The W3C WAI authoring practices ask for the trail to sit inside a navigation landmark with an accessible name, use an ordered list, and mark the current page with aria-current="page". If the last item is not a link, that attribute is optional, but adding it does no harm.
<nav aria-label="Breadcrumb">
<ol>
<li><a href="/en/">Home</a></li>
<li><a href="/en/blog/">Blog</a></li>
<li aria-current="page">Breadcrumbs and SEO</li>
</ol>
</nav>
Separators belong in CSS: li + li::before { content: ">"; }. That keeps the screen reader from announcing a symbol between levels, which is exactly what the W3C design system avoids. If you do use an element for the separator, mark it aria-hidden="true". And when a page has two breadcrumb bars, each needs its own distinct aria-label.
The BreadcrumbList Markup Google Understands
JSON-LD is the recommended format. You need at least two ListItem entries: with a single level there is no path to declare. Each entry requires name and position, and position starts at 1 and runs consecutively. The item URL field is required except on the last entry, where Google falls back to the URL of the containing page.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "BreadcrumbList",
"itemListElement": [
{ "@type": "ListItem", "position": 1, "name": "Home", "item": "https://yourdomain.com/" },
{ "@type": "ListItem", "position": 2, "name": "Blog", "item": "https://yourdomain.com/blog/" },
{ "@type": "ListItem", "position": 3, "name": "Breadcrumbs and SEO" }
]
}
</script>
Three details that save arguments in an audit. Google does not require the domain as the first entry, nor the page itself as the last one. If a page belongs to two categories, declare two BreadcrumbList objects, one per valid path, never one blended trail. And data-vocabulary.org is no longer eligible: if an old template still uses it, migrate it to schema.org.
The rule that settles almost every edge case: markup and visible HTML must describe the same trail. Marking up a path the user cannot see is one of the reasons Google ignores a rich result, because the policies ask you not to mark up hidden content.
Mistakes That Show Up in Every Audit
Duplicate markup
The theme and a plugin both output a BreadcrumbList, with different names for the same page. Google allows several paths, but not two contradictory versions of the same one. Turn one off instead of adjusting the second.
Breadcrumbs full of parameters
The trail links to /category?color=red&sort=price and turns navigation into a near-duplicate URL factory. Point it at the clean category URL. If the site lives on filters, the real problem is faceted navigation.
Breadcrumbs that only exist after JavaScript
If the trail appears on hydration, the initial HTML carries neither the link nor the markup. Check with "view source": the inspector shows the modified DOM and will lie to you.
Decorative trails
Icons as the last item, images instead of text, or a giant "Home". None of it helps crawling and all of it breaks on mobile.
Fighting with the H1
Repeating the page title as the last level is correct, but duplicating the full h1 inside the bar eats the title's space. The last level can be a short version. It ties into the H1, H2 and H3 heading structure used elsewhere on the site.
How to Check It in Ten Minutes
Four commands and one manual review are enough to know whether the trail really exists and whether Google can read it:
curl -s https://yourdomain.com/page | grep -o BreadcrumbListto confirm the markup is in the served HTML, without running JavaScript.- The rich results test, but on the published URL rather than a pasted code snippet.
- Search Console, Enhancements and then Breadcrumbs: valid, invalid and the exact error reason.
- URL inspection and rendered HTML, to tell "it is not there" apart from "JavaScript paints it".
- The browser accessibility tree: the bar should announce as a "Breadcrumb" navigation and the last item as current.
If the Search Console report is empty but the code looks right, check how many pages carry breadcrumbs: one article will not produce a report worth reading.
Copy-Paste Checklist
BreadcrumbListJSON-LD present in the initial HTML, verified with curl.- At least two
ListItementries,positionstarting at 1 with no gaps. - Short
namevalues using the real section name, not the full article title. - Last level without a link and with
aria-current="page"where it applies. - One
BreadcrumbListper path, with no contradictory versions. - No filter or sort parameters in breadcrumb URLs.
- Consistency between the trail, the URL and the
h1. - Separators handled in CSS or marked
aria-hidden="true".
Breadcrumbs will not bring the mobile trail back to your search result. What they still do, and that is real business value, is build a meaningful internal link network and make clear where each page lives. You measure that in crawling and click depth, not in a screenshot of Google's results page.