TL;DR
- Google removed breadcrumb paths from mobile search results in January 2025, keeping desktop display and the modest +5-10% CTR uplift there only.
- BreadcrumbList was untouched in Google's June 2025 structured-data deprecation and remains in the actively monitored set alongside Product and Review schema.
- Google's own AI guidance says structured data isn't required for generative AI Search inclusion — breadcrumbs don't earn you a citation by themselves.
- A 2026 audit of pages ranking for AI-search queries found BreadcrumbList on 60% of them, the second most common schema type, likely because it's a cheap way to expose site hierarchy.
- HTML breadcrumb links and JSON-LD BreadcrumbList are two separate layers: the links pass PageRank, the schema signals hierarchy. You need both — schema is not a substitute for real internal links.
BreadcrumbList is the schema.org type that marks up the trail of links showing where a page sits in a site ("Home > Blog > Schema > BreadcrumbList"). It was built for one job: letting Google render that trail as a path instead of a raw URL in search results. That job has shrunk. Its secondary job — telling any machine, search engine or language model, how a page nests inside a category structure — has gotten more relevant as AI systems try to understand sites without crawling every page.
What actually changed for Google
On January 23, 2025, Google stopped showing breadcrumb paths in mobile search results worldwide, replacing them with just the domain name. The official reason given was that breadcrumb paths were getting cut off on smaller screens and provided limited value there. Desktop results kept the breadcrumb display. That means the entire visible benefit of BreadcrumbList in classic search — the +5-10% CTR uplift estimated for pages using it over plain listings — now applies only to desktop traffic, which is a shrinking share of total search sessions for most sites.
Five months later, in June 2025, Google deprecated seven structured data types, including Book Actions, Course Info, Claim Review, and Vehicle Listing. BreadcrumbList wasn't among them. It stayed in the actively supported, monitored tier with Product and Review snippet schema. That's a signal Google still finds it structurally useful even if the visible reward on mobile has gone away.
Put those two facts together and you get an honest picture: BreadcrumbList isn't dead, but its classic SEO payoff is smaller and narrower than it was two years ago.
Why AI systems still care about hierarchy
Google's own guidance on AI-generated search results is explicit that structured data, BreadcrumbList included, is not required for a page to be pulled into generative AI Search answers. There's no special markup that buys a citation. That's worth stating plainly because vendors in this space routinely imply otherwise.
What BreadcrumbList does provide is a cheap, structured description of where a page fits relative to its siblings and parents. When a language model (or a retrieval system feeding one) is trying to figure out whether a page is a top-level guide, a sub-topic, or a deep implementation detail, the breadcrumb trail is one of the fastest signals to parse — faster than inferring hierarchy from URL patterns or nav menus alone. A 2026 audit of 60 pages that were showing up as citations for AI-search queries found BreadcrumbList present on 60% of them, making it the second most common schema type after the basics. The audit's authors described it as one of the simplest schema types to implement, which likely explains the adoption rate as much as any causal lift in citations.
None of this means breadcrumbs alone move a model toward citing you. Google itself says no special schema.org markup is needed to appear in AI Overviews, and the same logic extends to other AI answer engines. The path is one clue among several — headings, body copy, internal links — and it only helps when those signals agree with each other. A page whose breadcrumb says "Guides > Pricing" but whose body copy reads like a landing page pitch isn't going to get rescued by the schema.
The two-layer architecture, and why skipping HTML breadcrumbs is a mistake
Google spokespeople have been specific on this point: the HTML breadcrumb links on the page (the visible "Home > Category > Page" trail users click through) are treated as ordinary internal links and pass PageRank like any other link. The JSON-LD BreadcrumbList markup is a separate layer that drives the desktop SERP display and gives crawlers an explicit hierarchy signal. John Mueller has stated plainly that schema URLs aren't used the same way real internal links are — you can't drop the visible HTML breadcrumb trail and rely on JSON-LD to do the same job. Sites that implement BreadcrumbList schema but skip the clickable HTML trail are leaving half the mechanism unbuilt. Both layers need to exist, and they need to describe the same hierarchy consistently.
This matters more for topical authority clusters than for any single page. If your breadcrumb trails, your internal links, and your URL structure all describe the same category tree, you're giving both Google and AI systems a coherent map to work from. If they disagree — a page breadcrumbed under "Guides" but linked from "Resources" and URLed under "/blog/" — you're adding confusion, not clarity.
Implementation notes by platform
On Next.js, the cleanest approach is generating BreadcrumbList JSON-LD server-side from your route structure, so the schema path always matches the actual route hierarchy rather than being hand-maintained. Keep it in a shared layout component so new pages inherit it automatically instead of someone forgetting to add it.
On WordPress, most SEO plugins already emit BreadcrumbList automatically based on category and parent-page settings — check your theme isn't also rendering a conflicting breadcrumb trail, since duplicate or mismatched BreadcrumbList blocks confuse more than they clarify. If you're building an llms.txt file alongside this, keep the category structure you reference there consistent with what your breadcrumbs say.
On Shopify, breadcrumb schema for collection and product pages is often theme-dependent rather than app-dependent; verify your theme actually renders the HTML trail before assuming the schema alone covers it, especially given the platform-specific crawler constraints merchants already navigate.
Across all three, validate the output and then check it against real hierarchy — schema markup validation vs. AI usefulness covers why passing a validator doesn't mean the markup is doing anything for how a model reads your site.
If you want a baseline read on how your current structured data, including breadcrumbs, is being crawled and interpreted, CiteFlow's free scan runs a 7-dimension audit and shows a sample of AI answers with which URLs they're actually citing.
FAQ
Does BreadcrumbList schema help you get cited by ChatGPT or Perplexity?
Not by itself. Google's guidance confirms structured data isn't required for AI Search inclusion, and breadcrumbs are one hierarchy signal among several — they help when headings, links, and content agree, not as a standalone trigger for citation.
Is BreadcrumbList still worth implementing after Google's mobile change?
Yes, but size your expectations correctly: the CTR benefit now applies only to desktop results, and the bigger value going forward is giving crawlers and AI systems a clear read on site hierarchy rather than a rich-result cosmetic win.
Can JSON-LD BreadcrumbList replace visible HTML breadcrumb links?
No. Google has stated schema-only breadcrumb URLs aren't treated as real internal links the way HTML links are, so you need both the visible, clickable trail and the JSON-LD markup describing it.
Sources
Drafted with AI assistance and published by CiteFlow. Sources are linked in the text.



