Google: Unlinked URLs Use For Discovery But Not Ranking

Sep 12, 2013 • 9:08 am | comments (25) by twitter Google+ | Filed Under Google Search Engine Optimization

robot turtleI apologize for missing this a week ago, when I was sitting in the Google hangout where Google's John Mueller went on record that Google uses unlinked URLs or domains in body content as a way of discovering new content and pages.

John Mueller, a Google Webmaster Trends Analysts. said in a video 47 minutes and 32 seconds in (I was distracted with something else), that Google uses these URLS, even though they are unlinked, to discover new pages. It does not influence rankings but it does influence discovery.

Here is the video, skip to 47 minutes and 30 seconds in:

Credit goes to for spotting this first.

Also, Dan Petrovic tested this a year ago and came to the same conclusion.

Forum discussion at Google+.

Image credit to ShutterStock for robot turtle.

Previous story: Google AdSense New Formats: Blue Circle Arrows & Cleaner Spaced Ads


Michael Martinez

09/12/2013 01:40 pm

This information came out years ago.

Barry Schwartz

09/12/2013 01:40 pm

From Google's mouth?

Michael Martinez

09/12/2013 01:43 pm

Yes. From Google.

Barry Schwartz

09/12/2013 01:43 pm



09/12/2013 01:44 pm

SERIOUSLY? "For discovery but not for ranking"?. Then ... why bother discovering it?

Michael Martinez

09/12/2013 02:13 pm

The best I can do on short notice is this article from July 2012 on DejanSEO, where he says he confirmed with a Googler: However, there are dozens of other articles or conversations going back to 2008 (at least) where people mention the fact (confirmed by Google). I just don't know what specific keywords to use in a query to find the oldest source. For what it's worth, Google has supported text-only sitemaps for years, which is just one example of how they might use unlinked URLs for discovery. Cf.

Barry Schwartz

09/12/2013 02:23 pm

I linked to that in this post above. :)

Michael Martinez

09/12/2013 02:25 pm

There are also plenty of tests going back years. But in his post he specifically says he confirmed this with Google. So, again, it's old news.

Barry Schwartz

09/12/2013 02:26 pm

that was not confirmed by google. but whatever, it is fine.

Michael Martinez

09/12/2013 02:30 pm

YES, IT WAS. He says in the article he confirmed with a Googler. Erik Dafforn posted this link on your Search Engine Land article:

Barry Schwartz

09/12/2013 02:32 pm

Okay, sorry. It was new to me. I am not perfect. :)

Michael Martinez

09/12/2013 03:24 pm

What, like *I* can claim to be perfect? :O No problemo.

troy redington

09/12/2013 04:46 pm

for meta data and ranking factors for other pages


09/12/2013 04:58 pm

OK, so wait: "for meta data" isn't really especially meaningful EXCEPT as regards what goog thinks of you, so that IS SEO. As for the "ranking for other pages" angle ... interesting, but are you saying that SEO is a one-way street? Because ... you know ... it isn't.

troy redington

09/12/2013 05:19 pm

I'm not sure where the one way / two way street thing is coming from. But to clear things up, I interpret "not for ranking" as "not for indexing" - which means they'll look at the page for any potential signal value, but they won't include it in the index and 'rank it'. This has been my experience over the last decade of google crawling totally unlinked urls, and even malformed and obfuscated urls within javascript.


09/12/2013 05:38 pm

interesting distinction: specifically, the presence of the non-hyperlinked text string will make google "know you're there" but won't get the page (being named/pointed at/referenced) indexed. OTOH: if google discovers a page via this method, then, having discovered it, won't they eventually crawl it and thereby present themselves with the opportunity TO index it? ultimately (speed of crawl/index action might be impacted, but still), I don't see a functional difference. They discover the page, meaning they eventually crawl the page.

troy redington

09/12/2013 05:48 pm

a discovered page will get crawled. I've had mixed results as to whether or not that page ever gets indexed. It's hard to prove that there were no other signals (other pages linking to it). I should do a blog post about this before I forget all the data.


09/12/2013 05:50 pm

>> I should do a blog post about this before I forget all the data.

troy redington

09/12/2013 06:19 pm

oh nice. let me know when its live!


09/12/2013 07:33 pm

Troy, the only part that will be germane HERE is this line: The granularity of what form of usage of a web site’s address will pass it the most Google juice is no longer an important enough place to spend significant resources. But you asked, so:


09/12/2013 08:56 pm

You guys are funny. It's old info, confirmed years ago. URL or it never happened! Still good for the nubs to know, in case they missed something so basic.


09/13/2013 11:18 am

OMG the drivel that is coming out of Google these days. I think it was better when they weren't transparent!


09/13/2013 07:28 pm

Its a huge problem for rss feed? adwords? foruns?

Gracious Store

09/16/2013 04:11 am

What exactly does Google mean that it uses unlinked URLs to discover new pages?

Jitendra Vaswani

09/21/2013 08:37 pm

What does google mean by Discovery, will they affect rankings Barry ??

blog comments powered by Disqus