What a canonical tag actually does
A canonical tag is a hint, not a directive: it tells search engines which URL you consider the one worth keeping when several serve the same content. Google treats it as a strong signal but resolves duplicates with its own judgment too — which is why a canonical that contradicts the redirects, the sitemap and the internal links gets partially ignored, and the site loses the argument.
| Canonical state | Consequence |
|---|---|
| Missing | Google picks a favourite among your duplicates — not necessarily yours |
| Self-referencing | Safe default; blocks parameter and slash variants from splitting the page |
| Points at a duplicate | Legitimate for print/variant pages; consolidates signals onto the winner |
| Cross-domain | Tells Google another site owns the content — almost always an accident |
Should the canonical point to itself?
Yes, by default. A self-referencing canonical on every indexable page is the cheapest insurance in SEO: when a tracker parameter, a trailing slash or a campaign link creates a variant, the canonical tells Google the variant is the same page. Stores with faceted navigation need this more than anyone — a collection page reached via ?sort=price is a different URL, and without a self-canonical each sorting order is a separate candidate for the index.
The cross-domain canonical trap
The most damaging version is invisible: a page whose canonical points at a different domain. It happens when content is syndicated or copied with tags intact, when a staging domain leaks into production templates, or when an app "helpfully" adds a canonical. Google reads it as you saying this page is a duplicate of someone else\u2019s — and drops it. If you did not deliberately syndicate, a cross-domain canonical is an error, and it outranks every other canonical problem in urgency.
Canonical errors produce no error message anywhere. The symptom is a page that ranks badly or not at all while everything about it looks healthy — which is why the check reads the live tag, not your template.
Parameters, campaigns and canonicals
Email campaigns and ad platforms append parameters to your URLs. Those links are fine from other sites, but if any of your internal links carry parameters (utm anything, sort, filter), the parameterised URL is crawlable — and with no self-canonical it is a duplicate candidate. Fix the internal links to use clean URLs and let the canonical handle the external ones.
Run the free canonical check — it reads the canonical on your sampled pages, flags missing, broken and cross-domain ones, and compares each against the page URL. Related: duplicate content check · URL parameters check · pages not indexed
Shipwork reads the canonical on your sampled pages and reports missing, invalid, cross-domain and non-self-referencing canonicals — the states that split or surrender ranking. Free, no account, no signup — paste your store address.
Check my canonicalsQuestions
- What is a canonical tag?
- A link element naming the preferred URL for a page. When duplicates exist, it tells search engines which version to index and credit.
- Should every page have a canonical?
- Every indexable page should — a self-referencing one by default. It is the cheapest protection against parameter and variant duplicates.
- When is a cross-domain canonical correct?
- Only when you deliberately syndicate content to another site and want them to rank. Otherwise it tells Google to drop your page in favour of the other domain.
- Does a canonical fix duplicates immediately?
- No — it is a hint Google weighs against redirects, links and content. Expect the preferred version to consolidate over the following crawls.
Keep reading