Have you ever seen those URLs that people sometimes include on their pages, but which are not implemented as physical links? It can be frustrating if someone uses some of your content, and then gives you a citation like that, but it does happen. Did you ever wonder if Google treats those “URL Mentions” as a ranking factor?
The IMEC team decided to conduct tests to see if Google does indeed do that? In this post, we will report our results on whether or not Google places any value on these URL Mentions.
The Short Story
We tried two tests to see what would happen if we created 15 or more URL Mentions to a page to see if that would change the rankings of that page (more details on how we did that are spelled out below). Some of these were added to existing indexed pages, and some of them were included in new article posted on existing sites.
In neither case did we see any rankings change. This would imply that Google does not consider an URL Mention to be a signal. Or, at the very least, if it is used as a signal, that it’s weak in comparison to the other factors that Google considers, as there was no observable impact.
The Long Story
As mentioned above, we ran two tests. We picked queries and target sites that were fairly non-competitive in order to make it a bit easier to change the ranking position of the site:
We then asked IMEC members to volunteer their web sites for participation in the test. For each test, we created 10 original articles on the topic of our query. In the case of Giuseppe Garibaldi, the articles all were about the life of the Italian politician, and included a URL mention to the page on the www.naval-technology.com site for the aircraft carrier which is named after him.
In addition, in each test, we had additional sites that added a text blurb to existing indexed web pages on their sites. In order to help ensure that these were diverse in nature, we pre-created 10 different blurbs, so that participants would not be placing the same copy on their web sites.
As a result of all this effort, here is what we actually ended up getting for URL Mentions in the two tests:
As you can see, there were a fair number of URL mentions obtained in both tests. Here is a look at the Domain Authority from the participating sites:
Lastly, here are the results for the two tests in terms of rankings over time (note that the missing data points are times when the page was not found in the top 50 results):
Summary
As summarized above, we saw no indication that Google is using URL mentions as a ranking factor. We can’t treat this data as conclusive, however, as it’s still possible that it is factored into the algorithm, but as a relatively weak signal.
Thanks to the IMEC board: Rand Fishkin, Mark Traphagen, Cyrus Shepard, and David Minchala, and the entire group of IMEC participants! And, for completeness, here is my Twitter Handle.
Hi Eric, First of all thanks to you for sharing this case study and revealing about url mention links. I wonder if you share some sample links with us which you build up during this experiment.
Hi Kristy – unfortunately, we told the people involved to remove the pages where the URL mentions were, and the participants in that study have an expectation of anonymity. I can tell you that one of the URL mentions was on the Moz.com site.
How long did this experiment last?
We kept them in place for at least 6 weeks in each test.
Interesting results there Eric.
One thing I have noticed is that Google “appear” to be reading content “around” a link on a page and seem to be connecting the dots for matching content with the page linked to.
Anchor text, now seen as spammy in certain circumstances, and for the most not natural, can’t be relied upon like it used to dictate, but now we are seeing that “context” of both the link page and linked to page/domain are the “new anchor text”.
Anyone else seeing this?
Greg
Interesting, but I would concur that these are seen by bots. We use them to make up part of a link mix where heavily anchor texting (even with brand citations) can cause some problems, but consider that not all website will allow you to anchor text and not all people (general public beyond webmasters) understand a href and the idea of adding an achor thus I would suggest that they are and have been a part of rankings as they are… in part a link.
IDCx
Did you try and put an actuall link to see if that would change the ranking?
Only then I would think, you can truely conclude that a url as text is not treated as a link but as plain text.
I would love to see these url mentions turned into links to have a 1 to 1 comparison of links vs URL link mentions.
I could see it being a weak signal, but it I would be shocked if it had no value. A mention without an actual link still has a tie back to the site. Being a weaker signal would make sense since it isn’t a full out link.
Great article, though I must admit, I was disappointed with the outcome.
One of the issues we have with inbound.org is that because the name of the site is the URL, people are much less likely to make it a clickable hyperlink. So, we have a bunch of “mentions” of the site, but not as many links as we would like.
Such is life sometimes…
Hi Dharmesh – Agreed that the output was disappointing. But, as always we have to let the data tell the story.
That said, there were limits to this test, and it’s still possible that Google uses it as a signal in some scenarios. Just not the ones that we tested. For example, perhaps on a much larger scale, it might start to carry some weight.
Yes I agree, as Google has evolved, my observations are that they are reading text around a link now, so submitting the old “anchor text = about the link” approach.
I have seen several case studies of sites without any anchor text links ranking for the content surrounding the link page and target page.
I wonder if 6 weeks was truthfully long enough… As we all know Google is always changing something. They may have changed their intellectual direction to continually keep us in the dark.
I agree with Dr. Rovbert, a 6 weeks time may not be sufficient to get results with link mentions. Surely link mentions may not be as strong signals as a direct hyperlinks, but honestly speaking we found link citations pretty helpful for our website. I believe some more time would have been better for this test but yes, link mentions are those weak signals which can not rank a site independently. Though I firmly believe that getting a contextual link citation from an authoritative site is a huge plus for the mentioned website.
Best Regards
Soumya Roy
How can I find mentions pointing to my site, so I can contact the webmasters to convert them to links?
Both Moz and Majestic offer a way to find URL mentions.
I would maybe use Google Alerts, social mention.com and or socialreport.com as well as those mentioned by Eric
Dr. Robert, those tools will find mentions of your brand, but not necessarily URL mentions (URLs that are not links), unless the name of your brand or site is in the URL.
Mark – Ricardo P. simply asked about mentions to contact webmasters about adding his link.
I have heard in the past that Google does crawl/follow non-mentioned links and therefore they can help aid content discovery. Did any of your tests look into the validity of this? If not it might be an interesting follow up.
Hi David – we didn’t test that, but I bet that they do!
I agree. On its own, this test is fairly strong evidence that Google does not treat an orphan URL as a “link”, but without a comparison with a hyperlinked URLs from the same pages, same context and same keywords, it leaves a big question mark.
These results are weird, in my opinion. Less is good , thats what i learn.
One thing I’d also be curious about is in the case of a raw URL on a page that is automatically converted into an embed. In WordPress, simply pasting in a link to other WordPress sites will render as an actual embedded preview/excerpt of that linked page.
Hi Dustin – not sure what you mean by an embed. Can you send me an example of that? The bottom line issue is whether or not Google can see in the source code of the page a clean link.
WordPress just came out with an update that turns any raw link to another WordPress powered page into an embedded rich snippet. You can test it out by pasting a link to any one of your blog posts into a new blog post and previewing what it will look like. This is new with WordPress 4.4.
I think it we shouldn’t be surprised about that, although it’s a good thing to know. Taking into account that google indexes content according to its value and relevance for the user, it seems quite obvious that a link it’s far more relevant than just a bunch of text, even if it follow a link pattern. After all, a link usually stands out the page, as bold or italic fonts. Also, blind users have tools to jump from link to link (besides the tab key),.
^_^ I see what you did there Eric
I’m a little late to this article, but I can understand why Google would not want to give a ranking boost to URL mentions. It makes sense.
Interesting , Thanks for sharing this experiment with us. need to make some changes then 😉 New to all this, thanks for sharing anyway.
That’s a kind of experiment I was really thinking off. Glad it’s already been done. Similar to url mention, I was wondering if the specific targeted keyword that comes before and after a linked website does have any sort of impact as well? Too many variables to judge probably
It’s hard to test that idea, you’re right. But, to respond based on my gut intinct, it feels like the context of a linking page, as well as the text nearby the link, shoult count for something.
Umm does this really happen? Its an interesting experiment and thanks for the great information shared. But why does not google want link URL to be ranking factor.Kind of amazing matter. But I happen to get it a bit. Thanks again for the good work.
The study was designed to see if they use it as a way to discover new URLs, not to see if they use it as a ranking factor. It does not appear that they do either.
Wow!! Really interesting experiment.Thanks for sharing with us.Good work.
I would concur that these are seen by bots. We use them to make up part of a link mix where heavily anchor texting (even with brand citations) can cause some problems, but consider that not all website will allow you to anchor text and not all people (general public beyond webmasters) understand a href and the idea of adding an achor