Canonical loop (A→B→A): how to break the circle

What it means

Two or more crawled pages have rel=canonical tags that form a circle: page A declares B as canonical and B declares A (or a longer chain closes back on itself, A→B→C→A). The chain never terminates at a stable self-canonical, so Google cannot resolve which URL is authoritative.

Why it matters

A canonical is a consolidation directive — "treat that URL as authoritative." When the target itself points elsewhere and the chain circles back, the directive is unresolvable: Google ignores the circular canonicals and picks its own canonical, which may not be the one you intended. In practice one or more pages in the loop get deindexed or rank inconsistently, and inbound link equity flowing through the loop cannot be attributed. Loops are common in CMSes that generate canonicals from URL parameters or trailing slashes — /widget canonicalizing to /widget/ while /widget/ canonicalizes back to /widget.

How to fix it

1. Identify all pages in the loop from the report.
2. Choose one authoritative URL and make it self-canonical (its canonical points to itself).
3. Update every other page in the loop to point its canonical at that single authoritative URL.
4. Add a 301 redirect from each non-canonical variant to the authoritative URL, and normalize the CMS canonical generator to emit one consistent form.

When it's not a problem

Detection is limited to the crawled URL set: if only one page of a two-page loop was crawled, the loop may not be detected (an incomplete finding). In rare CDN configurations a redirect that also carries a canonical can look like a loop but resolve correctly in practice — review the specific URLs flagged before concluding there is a production problem.

How common is it?

Not yet observed in our audit sample. We publish this explainer because the check runs on every crawl; once an audited site trips it, this page will break the issue down by platform, gatekeeper, vertical and country.