How to Fix Duplicate URLs With a Canonical Tag
A guide site I audited last year had 612 published pages and 4,180 URLs known to Google. Nobody had written 3,500 extra pages.

A guide site I audited last year had 612 published pages and 4,180 URLs known to Google. Nobody had written 3,500 extra pages. The site had a filter on its item database, a comment reply link on every post, a print view, and six months of campaign tags from a newsletter, and every combination of those produced a URL that served the same guide. Search Console had quietly sorted most of them into one row: duplicate, Google chose a different canonical than the user.
That row is the one worth understanding, because it means Google read your instruction and decided against it. A canonical tag is not a command. It is a vote, and it loses when the rest of the site votes differently. Fixing duplicate URLs properly means getting the whole site to agree, and the tag is only one part of it.
What the canonical tag is actually for
When several URLs serve near identical content, the canonical tag tells search engines which one you want in the index and which one should collect the links and ranking signals from the others. It lives in the head of the page and it should hold one absolute URL, pointing at the version you want indexed.
A self referencing canonical, where the page points at itself, is the normal state for ordinary pages. Every WordPress SEO plugin I have used outputs one by default, and that is correct. It costs nothing and it heads off the parameter mess described below, since a URL with a tracking string appended will still declare the clean version.
What the tag does not do is stop crawling. Google still has to fetch a URL to read the canonical inside it, which is why consolidating duplicates with tags alone leaves crawl budget sitting in the bin on large sites. On a 600 page fan wiki that does not matter much. On a store with 90,000 filter combinations, it matters a lot.
Where the duplicates come from
Almost every case I see traces back to one of a short list of causes, and knowing which one you have decides the fix.
| Source of duplicate | Typical pattern | Right fix |
|---|---|---|
| Campaign tags from email or social | A guide URL with a utm_source string appended | Self referencing canonical, nothing else |
| WordPress comment reply links | A replytocom parameter on every post | Canonical plus a parameter rule, or disable threaded replies |
| Trailing slash and no slash both resolving | Two live versions of one path | One 301 redirect, pick a form and keep it |
| Both www and bare domain reachable | Two full copies of the site | 301 at the server, never a canonical |
| Uppercase letters in URLs | A mixed case path serving the same page | 301 to lowercase |
| Filters and sort order on a database page | Sort and filter parameters in any order | Canonical to the unfiltered view, block the useless combinations |
| Printer friendly pages | A separate print path per article | A print stylesheet instead of a second URL |
| Pagination on a long list | Page two, three and four of an archive | Self referencing canonical on each page |
The last row is the one that trips up experienced people, so it gets its own warning. Pointing every paginated archive page back at page one used to be common advice and it is wrong. Google asks you to treat each paginated page as its own URL with its own self canonical, since page four holds content that page one does not. Collapsing them hides the links on those deeper pages, which on a wiki is often how older guides get discovered at all.
The part where canonical tags get used as a plaster
Here is my disagreement with how this tag usually gets taught. Most articles present it as the answer to duplicate content. In practice it is the weakest of the available tools and the most commonly misapplied.
If two URLs serve the same page and only one should exist, redirect. A 301 removes the ambiguity, passes signals without argument, and sends actual visitors to the right place. Canonical tags leave both URLs live, which means people keep linking to the wrong one and you keep relying on a hint being honoured. Our guide to WordPress redirects after a URL change covers the mechanics of doing that without a redirect chain.
Three more mistakes come up often enough to name. Blocking duplicates in robots.txt does not consolidate anything, because a blocked URL cannot be crawled, so the canonical inside it is never read and the URL can still show up in results with no description. Mixing a canonical with a noindex on the same page sends two opposed instructions and Google resolves it unpredictably, usually by dropping the lot. And canonicalising a whole set of thin tag archives to the homepage does not make them count as the homepage, it just signals that you know they should not exist, in which case delete them.
There is also a quiet failure mode on migrated sites: a canonical that points at a URL which itself redirects, or which returns a 404. Google ignores those, picks its own winner, and you never find out unless you look.
Getting the tag itself right
The syntax is small and the rules around it are stricter than people expect. One canonical element per page, in the head, with an absolute URL including the protocol. If a page outputs two canonical elements, which happens when a theme and a plugin both add one, Google discards both and chooses for itself.
In HTML it looks like this, written here with single quotes so it reads clearly:
<link rel='canonical' href='/guides/world-boss-rotation/' />
Use the full URL with the domain in production rather than the path shown above. Then make the rest of the site agree with it. The canonical target should be the URL in your sitemap, the URL your internal links use, the URL that returns a 200 status, and the URL in your breadcrumb. When those four line up, Google almost never overrides the declaration. When the sitemap lists one form and the menu links another, expect the overriding row in Search Console.
Checking what Google actually chose
Open the URL Inspection tool in Search Console and paste a duplicate, not the version you want indexed. The report shows both a user declared canonical and a Google selected canonical, and the whole point of the exercise is to see whether they match. If they differ, the page indexing report will usually explain which competing signal won.
Give it time before you panic. In my experience consolidation shows up within one to three weeks on a site Google crawls daily, and considerably longer on a small site crawled every few days. Changing the tag again after four days, which is the common reflex, only restarts the clock.
A rough order of work that fits an afternoon: list every URL Google knows about from the page indexing report, group them by the causes in the table above, fix the server level duplicates with redirects first, then the parameters, then confirm the sitemap holds only canonical URLs. Checking the sitemap last catches most of the leftovers, and our notes on an XML sitemap for a fan site explain what belongs in it.
FAQ
Does every page need a canonical tag?
Every indexable page benefits from a self referencing one, and plugins add it automatically. Pages you have deliberately noindexed do not need one.
Can I point a canonical tag at a different domain?
Yes, cross domain canonicals are supported and are the normal approach for syndicated copies of an article. The target has to be a live page that genuinely holds the same content.
Will a canonical tag fix a Google title rewrite?
No, those are separate problems with separate causes. Duplicate URLs confuse indexing, while rewrites come from the title text itself, which we cover in the piece on Google title rewrites.
Is duplicate content a penalty?
Not a penalty. The cost is dilution, since links and relevance signals spread across several URLs instead of landing on one, and the version Google picks may not be the one you wrote the title for.
More in Web Tips
Web Tips
How to Stop Layout Shift on Ad Heavy Game Guide Pages
The complaint that finally made me fix this arrived in our wiki Discord as a seven second clip.
Web Tips
Last Updated Dates on Guide Pages, Done Honestly
Our wiki had a raid guide that was genuinely current.
Web Tips
How to Fix Discord Link Previews for a Game Site
Someone in our clan Discord posted a link to a raid guide I had spent a weekend writing, and the embed came out as the site...