While helping a company that was severely impacted by multiple algorithm updates in fall 2023, I first scanned GSC reports to get a feel for the situation. And when I looked at the inbound links report, I noticed something strange about a number of links. There were some relatively high-performing referring domains in the list, but when checking the URLs that received these links, the destination URL was different. In other words, the URL that received the link was actually from a separate URL… That can happen, but what I found was slightly different. The destination URLs came from another website. Therefore, the inbound links counted for my client’s website actually came from a third-party domain.
Dan Petrovic called this “Link inversion” in 2018. I’ve also seen this several times over the years and quickly got a feel for what was going on. This can happen when Google canonicalizes URLs from other websites into a URL on your own website. This usually happens when Google sees the same content on multiple domains and chooses a single canonical URL. And the links pointing to these third-party URLs are now transferred to the canonical URL (which may or may not be on your site).
For some websites, you may be very familiar with third-party URLs that contain the same content (e.g. other company-owned websites that publish duplicate articles or pages). However, some other websites may have no connection to your website at all. For example, my client had no idea what was happening and what effects the “signal shift” might have.
To make things even more complex, since this often happens when there is duplicate content on multiple sites, Google can always do this reverse his decision about which URL is canonical. I’ve experienced this in the past and all the signals suddenly point to a third-party URL compared to your own (and of course the third-party URL will rank in the SERPs). If you don’t know this is happening, it could well cause confusion in rankings, impressions, and clicks on these URLs.
Again, this is nothing new, but I wanted to remind you how this works since it can be difficult to track down the situation. I’ll cover more about this below.
Find the hard evidence: Visit the link report in GSC
Based on what I explained above, you might think that you can go to the URL inspection tool in GSC and find this information. Well, that depends on which end of the canonical decision you are at… The URL Inspection Tool will only show you if your URLs are canonicalized to other URLs (This could include URLs on third-party domains). However, if third-party URLs are canonicalized into your URLs, you will not see this when inspecting the URLs in GSC.
But there are also other ways to get to the bottom of this. One way to see this in action is to check your inbound links in GSC (as I mentioned earlier). If you click on a destination URL (the URL that gets a link on your website) in the report, you can see if the destination URL reported by Google is different. And if it’s different, it could be a third-party URL (on a different domain). And if that’s the case, that means your URL is receiving the incoming links and not the third-party URL. Therefore, signals, including PageRank, are applied to your URL.
Here’s an example of what this looks like (I changed the domains and URLs to hide the websites involved):
Raining on the (SEO) Parade: Manual and incomplete.
But hold your applause. There are significant limitations to the use of this approach. The inbound link reports in GSC do not provide a comprehensive list of links, especially for sites with a lot of links. There is also no API for inbound link reporting, so this process is manual and incomplete. And those are two terrible words that every SEO needs to hear. 🙂 🙂
If you export links in bulk from GSC (top level), you can get up to 100,000 links, but you won’t have destination URL information, canonical information, etc. You need to drill down to referring domains and URLs in the link reports to get this information (which is an extremely manual process).
Trouble in Paradise: Canonical role may change
Remember that the “signal shift” situation can typically occur with URLs that contain the same content (duplicates), which can also be PDFs duplicated on different websites. For example, multiple websites may host the exact same PDF and Google needs to decide which of them to rank. Depending on several factors, Google may change the canonical URL if it sees fit. And then signals including PageRank are now passed to the third-party URL. That’s why I titled this post “Signal Shifts.”
You can learn more about how Google chooses a canonical by watching this great video from the Google Webmaster Conference in 2019. I was in the audience and liked this presentation from Allan Scott from Google:
And if suddenly your URL is no longer the canonical one, your URL will no longer rank and all signals will be redirected to the canonical URL (which is now on a third-party website). Again, I know this isn’t a widespread problem on many sites, but if this occurs with URLs on your own site, you should be aware of it. For example, the client I mentioned earlier has thousands of URLs where this happens because of its setup.
Here is an example of canonical switching on my client’s website. The page literally disappears from the SERPs when the canonical page switches, and then returns when the canonical page switches back again:
The signal flow through your website:
And remember, these URLs on your site receive inbound links from third-party URLs, and your URLs then pass signals to other URLs on your site (as long as you provide internal links from those URLs). These signals are essentially truncated when your URLs are no longer selected as canonical.
And depending on how many URLs on a site are affected, other URLs being linked to could potentially experience ranking issues because the URLs lose their canonical status. And then there are the website-level algorithms used by Google, which could certainly use link signals to some extent. If this happens to many URLs, it can affect more than just the loss of the URLs’ canonical status. This could affect other URLs that are on your site.
As for turning Canonicals on and off, a few years ago I helped a health and medical site with a lot of content licensed from a third-party resource. I watched the site rise and fall over time depending on which site Google considered canonical (in terms of URL). When my client’s website was selected, it ranked at the top for highly competitive health and medical questions. If the canonical page moved to another site with the same licensed content then a boom occurred, they simply didn’t rank because those URLs were canonicalized, didn’t rank, and didn’t get the inbound links that the cluster had over time had built.
I warned my client about this many times, but due to his business model, it wasn’t a priority for him. It was just fascinating to watch the passage of time (especially with major algorithm updates, which often saw big changes on that front). The entire directory containing this content exhibited a yo-yo trend as the canonical URLs changed.
Summary: Understanding the “shift” of signals, links and rankings.
Shifting signals can be sneaky and difficult to detect, but they could occur with duplicate content on your site. And you may or may not benefit from the shift, depending on how Google’s systems view the canonical situation. If you have a lot of content that could fall into this category, I recommend you do your best to figure out what’s happening and then develop a plan to address the problem (if you can). At the very least, it might help you understand why your site appears or disappears for certain searches over time. In other words, try to understand the “shift.”
GG