What Is Breadcrumb?

How Breadcrumbs Work

A breadcrumb does one job for a human and a second job for a machine. For the human, it is orientation: Google describes a breadcrumb trail as something that “indicates the page’s position in the site hierarchy,” letting a visitor climb “up the site hierarchy, one level at a time.” Land on a specific running shoe from a search result and the trail Home › Shoes › Running Shoes tells you instantly where you are and gives you one-click routes back up to broader pages.

For the machine, the same trail becomes a structured signal when you mark it up with BreadcrumbList structured data. Google reads that markup to categorize the page and can then render the trail inside the search result itself. The markup is small. A BreadcrumbList contains an itemListElement array, and each element is a ListItem with three parts: a name (the label a user sees), a position (its order in the trail, starting at 1), and an item (the URL of the page that crumb represents). Google requires at least two ListItem entries for a valid trail, and the last crumb — the current page — can omit its URL because you are already there.

Because each crumb is also an ordinary internal link, breadcrumbs quietly reinforce site architecture. They create consistent, crawlable paths from deep pages back up to category and section pages, which spreads PageRank toward those hubs and reduces the click depth between the homepage and buried content.

What Google Requires

Google’s guidance on breadcrumb markup is specific and worth following to the letter:

• Represent the user path, not the URL. Google recommends providing “breadcrumbs that represent a typical user path to a page, instead of mirroring the URL structure.” The way people navigate to a page is often not the same as its folder path.
• Skip the redundant crumbs. Including the top-level domain or the current page in the trail is unnecessary.
• Order matters. position must reflect the true order of the trail, from the broadest ancestor down to the page.
• A single page can carry multiple trails. The same page can belong to more than one path — for example a shoe reachable under both a brand and a category — and Google can use different trails to contextualize it for different queries.

Example of a Breadcrumb

Google’s own structured-data documentation supplies the canonical worked example. Its reference BreadcrumbList describes a page reached by the path Books › Science Fiction › Award Winners, encoded as an ordered list of ListItem objects. The first item carries name: "Books", position: 1, and an item pointing at the books landing page. The second carries name: "Science Fiction", position: 2, and the URL of the science-fiction category. The third, the award-winners page itself, carries position: 3 and, per Google’s guidance, need not repeat its own URL.

The instructive detail is what the example does not do. It does not mirror a URL such as /books/sciencefiction/awardwinners/ folder-for-folder as a matter of principle; it encodes the navigational hierarchy a reader would actually follow. Google’s documentation drives this home by showing that a single page can be reached by more than one breadcrumb trail — the same content might sit under a genre path for one query and an “award winners” path for another — and each trail can legitimately appear as its own BreadcrumbList. When that markup is valid and matches the breadcrumb visible on the page, Google is eligible to show the trail as a breadcrumb rich result in Search on desktop, replacing a bare URL in the listing with a readable path that tells searchers where the page lives before they even click.

Frequently Asked Questions

The Bottom Line

A breadcrumb is the small trail that answers "where am I in this site?" — a stack of links from the homepage down to the current page. Its SEO value is twofold: the BreadcrumbList markup can surface the page’s context directly in Google’s result, and the crumbs themselves are internal links that reinforce your category structure. Build the trail around how people actually navigate, keep the visible version and the structured data in lockstep, and it will do quiet work for both readers and crawlers.