Google’s John Mueller responded to a question about a website that was presumed to have been de-indexed because of a cross-domain “canonical content” issue.
Cross-Domain Canonical
The Redditor referenced canonical content. It’s unclear what they meant by that, but it sounds like they may have been referring to a cross-domain canonical.
A cross-domain canonical is a meta tag that tells search engines that the content on one site is the same as the content on another site. It’s the multi-website version of the regular canonical meta tag that handles URLs within the same website. Google treats both meta tags as a strong hint but is not obligated to obey them.
The cross-domain canonical is a way to indicate that a site was migrated from one domain name to another one, but only if a 301 redirect was absolutely not possible for whatever reason, which nowadays is not something likely to happen.
Cross-domain canonicals were then put to use as a way to indicate that the original version of a syndicated page of content was on a different URL.
Google’s guidance on cross-domain canonicals for syndicated content subsequently changed so that the best practice was to recommend the use of a meta no-index HTML element.
The guidance on syndicated content now reads:
“If you syndicate your articles to other news sites or other sites within your own network, they can add this robots meta tag to your articles:
<meta name=”Googlebot-News” content=”noindex”>
This tag stops Google News from indexing the syndicated versions of your content.
To restrict syndicated content from both Google News and Google Search, other sites can add this robots meta tag to your articles:
<meta name=”Googlebot” content=”noindex”>
This tag stops Google’s main user-agent, Googlebot, from indexing your content.”
Given that a canonical meta tag is just a hint, the reality at this point is that there is no reason to use a cross-domain canonical because there are other better alternatives to accomplishing the same thing in a more absolute manner that obligates Google to obey it, like the 3xx redirects and the meta noindex directive.
Report That A Cross-Domain Canonical Caused A De-Indexing
Someone on Reddit posted that their site was de-indexed and, upon investigation, discovered that a casino site appeared to be canonicalized as their own site.
The Redditor didn’t say how it was canonicalized as their own site because the way to do that requires a canonical meta tag on the Redditor’s site pointing to the casino site.
They dropped the URL of the casino website, which I’ve changed to example.com.
This is how they described it:
“We are seeing a slightly but growing de-indexing of our pages in Google by a very strange domain.
For context: Our pages are about companies and suppliers, and the page that google sees as the canonical one is a casino betting page (www.example.com).There is absolutely no similarity in content between our pages and this page.
Moreover when analyzing www.example.com I cannot not even find any page which goes near our content.
Does anyone may has any idea what this is about and how to fix this?”
Why The Site Seemed Canonicalized
Another Redditor (No_Wrap_9584) posted that a similar thing happened to them. What they discovered is that their site was showing a server error message and that other sites were also showing the exact same message and Google just lumped them together and after a few weeks it resolved by itself.
User No_Wrap_9584 explained:
“When searching the third-party canonical URL in Google Search, it appears indexed with the title:
“Application error: a client-side exception has occurred (see the browser console for more information).”
This is a generic JavaScript application error message. Notably, this is the same error message our own site has occasionally displayed during temporary outages when the application fails to load correctly.
This makes us suspect that at some point Googlebot may have crawled an error or fallback response rather than real page content, and then treated multiple URLs showing that same error shell as duplicates. That could explain the cross-domain canonical selection and subsequent de-indexing, even though everything looks clean now.”
Google Responds To Report Of De-Indexing
Google’s John Mueller agreed that the Redditor’s explanation could be an option and recommended using the Search Console’s live URL checker to review how Google is rendering the web page.
Mueller published a second reply with additional explanation and advice:
“Ultimately I don’t think this matters (though yes, it’s confusing). The potential outcomes are basically:
a) your page is seen as canonical, but indexed with the server message => your page is not showing in search for normal content
b) your page is seen as a soft-404 (imo this is what we should be doing) => your page is not showing in search for normal content
c) the other page is canonical => your page is not showing in search for normal content
The ideal solution is more to find ways to recognize this kind of error on your end, before you make the site live with the error. I don’t know about this site’s setup, but one thing that I’ve been doing with my smallish sites is to run a ton of automated tests before pushing the site live.
Whenever I see something go wrong, I have the code-agent add a new test for that. The tests take a few minutes to run, but you have a bit more certainty that the site – when live – will be ok. You can do similar things by either setting up site monitoring yourself or with a 3rd party tool: fetch the most critical pages “hourly” (or whatever) and check for issues, so that you can fix them before they become stable problems for search engines to pick up.”
Are Cross-Domain Canonical De-Indexings Real?
In order for a cross-domain canonical to work, a site owner’s domain would have to contain the cross-domain canonical to some other website in order for the signals to transfer to another site. If that’s what’s happening, then it could be that the site was hacked. If that’s not what’s happening, then it’s not a cross-domain canonical de-indexing.
An SEO definition of a coincidence is when two things happen independently of each other but have the appearance of being connected. For example, link disavowals take months before any effect is seen in the search results. Yet many people report having used them and seeing an effect within a few weeks. That’s a coincidence.
People sometimes get sick and say they ate something that didn’t agree with them when the underlying issue may have been that they touched a shopping cart that some toddler had infected with norovirus. It’s not really a “stomach bug,” but that’s a catchall people use. SEOs do something similar when pages stop ranking. They see that multiple pages have related content. Well, you know… if a site is about a topic, it’s natural that it’s going to have multiple pages with similar content. But some SEOs see it as evidence of the dreaded “keyword cannibalization” issue, also known as a stomach bug…
Sometimes things look like they’re related by cause and effect, but they’re not. This may be what happened here, where the Redditor thought they saw something and tagged it as the reason. If they had done a backlink search, they may have found a bunch of low-quality links and concluded that the links were the reason. But again, that’s just a hunch, a guess, not a clean case of cause and effect.
Featured Image by Shutterstock/vittaya pinpan
