Canonical Tag

One article existing at several addresses — duplicate content — is closer to a default than a mistake. Adding a tracking parameter to the end of a URL so ad traffic can be counted creates a second address; a version with a trailing slash living alongside one without creates a third. Leave old URLs alive through a site rebuild and there is a fourth.

Pagedeck's blog had one article at four addresses. The one that showed up in search results carried the ad tracking parameter, so everyone arriving through it got recorded as ad performance.

The larger loss was the split in credit. Links other sites had given the article were spread across the four addresses, so the credit one article earned arrived in four pieces. A search engine picks a canonical among them on its own, and its pick may not be the address you want to promote. This tag is where you write that pick down yourself.

Three things.

1. Consolidating split credit. The value of links spread across four addresses accumulates on the one address named as canonical.
2. Choosing the address that appears in results. A parameter-laden URL stops surfacing in search, and the address people copy and share converges on one.
3. Saving crawl capacity. The bot stops reading the same content four times, so on a site with many URLs other pages get their turn sooner.

Write the canonical address in the page head as a full URL. A canonical page should point at itself as well; without that self-reference there is no baseline value when a parameter version appears later.

One thing to know upfront: this is a hint, not an instruction. A search engine reads it and decides, and it ignores the tag when the content actually differs. So it does not work as a way to bundle genuinely different pages together to pool their credit.

The most common accident is a whole site pointing at the homepage as canonical. One wrong line in a template does it, and then every article declares itself a copy of the homepage and drops out of search entirely.

"This post's canonical points at the homepage. That's why it isn't indexed."

A wrong tag has taken the page out of the search index.

"Put a canonical on the ad URLs too. The credit is splitting."

Make the parameter versions point at the canonical address.

"The content is different, so a canonical won't bundle them. It'll just be ignored."

The tag does not apply across genuinely different pages.

At Pagedeck, Ravi has this on the pre-publish checklist. He checks that a new post carries a tag pointing at its own address, and when an ad URL gets created he makes it point back at the original post. Theo handles the template side: the self-referencing address goes in by default and only exceptions get specified, because anything written by hand always gets skipped somewhere.