---
title: "Missing canonical tag - Serpwise"
description: "Without an explicit canonical, Google picks one - often the URL with tracking params, a paginated variant, or a duplicate. Add a self-referential canonical."
url: "https://serpwise.ai/learn/missing-canonical-tag/"
source: "https://serpwise.ai/learn/missing-canonical-tag/"
---
[← Back to Learn](/learn/)
Canonicalization Warning
# Missing canonical tag
Without an explicit canonical, Google picks one - often the URL with tracking params, a paginated variant, or a duplicate. Add a self-referential canonical.
`` tells search engines which URL is the authoritative version of a piece of content. Without it, Google picks for you based on signals it can see - internal links, sitemaps, redirects - and the result is often a URL with tracking parameters, a paginated variant, or a near-duplicate.
Canonical is a hint, not a directive. Google may still pick a different URL if other signals are stronger. But a missing canonical is the loudest possible signal of indifference, and the cheapest one to fix.
## When to add a canonical
Every indexable HTML page. Including the homepage. Including pages that you think have no duplicates - query parameters, mixed-case URLs, trailing slashes, and HTTPS/HTTP variants all create duplicates you may not have noticed.
The default canonical is **self-referential**: the page’s canonical points to itself, on the preferred protocol, host, and path. No query string unless the query string actually distinguishes the content.
## How to detect it
```
curl -s https://example.com/some/page \
| grep -oE ']*rel="canonical"[^>]*>'
```
If the output is empty, there’s no canonical. If there are multiple matches, you have conflicting canonicals (also a problem).
In Google Search Console: **URL Inspection** shows “User-declared canonical” (what you said) and “Google-selected canonical” (what Google picked). Divergence means Google overrode you. Missing user-declared means you didn’t declare one.
## The fix
### Universal HTML
```
```
Rules:
- Absolute URL, including protocol and host.
- One canonical per page. Multiple `` tags confuse the picker.
- Inside ``, before any `` that could be parsed instead.
- The canonical URL must return `200` - don’t canonicalize to a redirected or 404’d target.
### HTTP header alternative (for non-HTML resources)
For PDFs and other non-HTML, set the canonical in the response header:
```
Link: ; rel="canonical"
```
### WordPress (Yoast / Rank Math)
Both plugins emit a self-referential canonical automatically. Verify it’s working - disable any caching plugin temporarily and view source on a few key pages.
To override on a specific post: Yoast → Advanced tab → Canonical URL. Rank Math → Advanced → Canonical URL.
If your theme also emits a canonical (some do, in `header.php`), you’ll get duplicates. Remove the theme’s `` and let the plugin handle it.
### Shopify
Shopify emits a canonical via the theme. Check `theme.liquid`:
```
```
If it’s missing, add the line. Shopify’s `canonical_url` already strips known tracking params and handles `?variant=` correctly for product pages.
### Static HTML
Add it to your build template. For multi-page sites, generate the canonical from the page’s URL slug at build time.
### Next.js (App Router)
Set canonical via metadata API:
```
// app/blog/[slug]/page.tsx
export async function generateMetadata({
params,
}: {
params: { slug: string };
}): Promise {
return {
alternates: {
canonical: `https://example.com/blog/${params.slug}`,
},
};
}
```
For sites with `metadataBase` set in the root layout, you can use a relative path. Verify the rendered HTML - `metadataBase` errors silently fall back to localhost during dev.
### Server-side header (Nginx / Apache)
For non-HTML assets:
```
location ~ \.pdf$ {
add_header Link '; rel="canonical"' always;
}
```
## Pitfalls
**Don’t canonicalize across substantially different content.** A blog post and its category page aren’t duplicates - canonical-ing them together hides the category page from search.
**Don’t canonicalize the wrong direction for paginated archives.** Page 2 of `/blog?page=2` should canonical to itself, not to page 1. Google deprecated `rel=next/prev` in 2019; canonicalizing all pages to page 1 hides paginated content entirely.
**Don’t canonicalize HTTPS pages to HTTP, or with-www to without-www inconsistently.** Pick one canonical form for the whole site, enforce it via `301`, and emit canonical tags matching it.
**Don’t include unused query strings in the canonical.** A page rendered at `/products/widget?utm_source=email` should canonical to `/products/widget` - strip the tracker.
**Don’t canonical to a 404 or to a redirect target.** Google ignores both and falls back to picking on its own.
## Fix at the edge with Serpwise
When the canonical strategy crosses CMS boundaries - marketing on one stack, store on Shopify, help center on a vendor - the safest place to enforce a single canonical strategy is the edge.
Serpwise can inject `` based on URL patterns: strip `utm_*` params, normalize trailing slashes, force HTTPS-only canonicals, override missing canonicals on entire sections of the site - all without touching origin templates.
[See pricing](/pricing) or run a [free AI visibility audit](/audit).
From diagnosis to deployment
## Find the issue. Ship the fix.
Use Learn to understand the problem, then run Serpwise against your own site to see what can be approved and deployed.
[Run free audit](/audit/) [Book a demo](/demo/)