# Circular Canonicals Tank Both Pages — Canonical Cluster Mapping Per-URL canonical checks catch the obvious bugs. Graph-level canonical mapping catches the subtle ones — circular canonicals (A→B, B→A), canonical pointing at a noindex page, cross-protocol drift. The mapper fetches your sitemap, builds the target graph, clusters by canonical destination, and flags every conflict type. Author: J.A. Watte Published: May 10, 2026 Source: https://jwatte.com/blog/blog-tool-canonical-cluster-conflict/ --- You can check one URL's canonical tag easily. You can't catch a circular canonical that way — you'd need to check both pages and cross-reference. Circular canonicals quietly tank both pages because Google can't pick a winner and treats the signal as conflicting. Same with canonical-to-noindex. Page A declares page B as its canonical. Page B is noindex. Google consolidates A's signals into B, which is noindex, so the signals disappear entirely. Both pages lose. [Canonical Cluster + Conflict Map](/tools/canonical-cluster-conflict/) fetches every URL in your sitemap, builds the canonical-target graph, clusters by target, and flags every conflict type. ## Five conflict types 1. **Circular canonical.** A → B, B → A, neither is the canonical of anything else. Fix: pick a primary, emit self-canonical on both pages pointing at the primary. 2. **Missing self-canonical.** Page has no `` at all. Google picks its own favorite URL as the implicit canonical — usually wrong when query parameters or trailing slashes are involved. 3. **Canonical → noindex.** Page A canonical to Page B where Page B is noindex. Kill signal. Fix: either remove noindex from B or drop canonical from A. 4. **Canonical → external.** Canonical points outside the set (could be legitimate cross-domain consolidation or a broken hardcoded URL). 5. **Cross-protocol.** HTTP page canonical to HTTPS (fine) or HTTPS canonical to HTTP (bad — protocol downgrade). Almost always a template bug. ## Why graph view beats per-URL view The conflicts that matter most are relational. You can't detect a circular canonical from one page; you need to see both canonical claims. The mapper builds the full graph across your sitemap and surfaces every edge type. A clean cluster looks like: one target URL (the canonical destination) with one or more variant URLs all canonicalizing to it. A dirty cluster has circular edges, missing edges, or edges pointing at noindex / external / wrong-protocol targets. ## The fix workflow 1. Run the mapper. 2. For each circular pair, pick the primary (usually the shorter or more-semantic URL) and update both pages' canonical. 3. For each missing self-canonical, emit a template-level `` rule. 4. For each canonical → noindex conflict, either remove the noindex (if the target should be indexed) or drop the canonical (if the source should canonicalize elsewhere). 5. Deploy. Rerun the mapper to confirm. ## Related reading - [Redirect Chain Audit](/tools/redirect-chain-audit/) — 301 chains (distinct from canonical chains) - [Index Coverage Delta](/tools/index-coverage-delta/) — what's actually indexed - [Sitemap Audit](/tools/sitemap-audit/) — XML hygiene ## Fact-check notes and sources - Google canonicalization: [developers.google.com/search/docs/crawling-indexing/canonicalization](https://developers.google.com/search/docs/crawling-indexing/canonicalization) - Consolidate duplicate URLs: [developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls](https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls) --- *The $100 Network covers keeping canonical graphs clean across site networks, where template drift multiplies fast. The mapper is the quarterly audit.* --- Canonical HTML: https://jwatte.com/blog/blog-tool-canonical-cluster-conflict/ RSS: https://jwatte.com/feed.xml JSON Feed: https://jwatte.com/feed.json Hero image: https://jwatte.com/images/blog-tool-canonical-cluster-conflict.webp