Links In IFrames Pass Value In Google

Jan 9, 2012 • 8:23 am | comments (48) by twitter Google+ | Filed Under Google Search Engine Optimization
 

iframe linkMichael Martinez posted a thread at High Rankings Forums basically confirming earlier reports that Google will see and pass value of a link within an iframe.

Michael said he ran his own independent test.

Michael said:

So I designed a simple two-stage test in December to see what would happen. The test is such that if the link passes anchor text in the first stage there is no need to try the second stage.

I embedded an iframed page on one of my older sites. It is well-indexed, has never been penalized, and has sufficient inbound links that I was confident it would be recrawled and re-indexed within a short period of time. It is also a low traffic site (it used to have more traffic a few years ago) so I felt it was probably a safe place to run a link test without attracting too much attention.

The iframed page only contains one in-body HTML element, a link pointing to HighRankings.com with the anchor text of "jill whalen says search engines crawl iframes". This was, until I ran my test, an expression that did not appear in Google's search results (when you wrap quotes around it).

I let the test be through the holidays and checked it today. The expression now appears in Google's search results.

He is extremely confident about his test that Google can crawl iFramed pages and follow the links within them.

Forum discussion at High Rankings Forums.

Image credit via ShutterStock for links and frames.

Previous story: adCenter Bans Accounts By Accident
 

Comments:

Zaitbahajji

01/09/2012 02:08 pm

Mmm... I'm still skeptic about that; I really doubt GoogleBot can "read" and index any content within iframes. 

Boston SEO

01/09/2012 02:27 pm

I have to wholly disagree, chances are more likely that the iframe page was discovered in a sitemap and cached. Their is a much easier way to determine what search engines can and cant see. If you have the ability to see when your site was cached you also have the option to see only the text version. If you can see it in the text version as a link then the search engines see it.

Daphill05

01/09/2012 02:32 pm

Also skeptical.  Perhaps it's based on the location of the frame and how much, and WHAT, is in it...especially if Google is reviewing with actual people instead of bots sometimes.  My opinion is that for the sake of keeping a page strong, a page that is genuinely full of rich and useful content, but has a hundred links for internal navigation from dropdowns and footers, as well as outbound links to expert resources, you should be able to iframe these blocks of links to prevent PR leaking.

Chris Dugdale

01/09/2012 02:37 pm

Are you serious? Search engines have been comfortable with frame sets for years. The main danger with them is that the sub-frame is treated as a separate page (distinct URL after all) and the frame content is therefore not associated with the host page.

Zaitbahajji

01/09/2012 03:01 pm

well that's what I had in mind when writing it (sorry) : it's treated as a separate page, therefore the content is not associated with the page that hosts the iframe.

Osmaan Shah

01/09/2012 04:10 pm

It would be a massive SEO loophole if google followed links within embedded iframes....I find it hard to believe that google search architects would allow this to happen.

SEO Group

01/09/2012 04:42 pm

Hopefully this encourages more testing so we can be sure. But it's exciting news that will hopefully prove true soon or later.

Restaurant Vouchers

01/09/2012 04:52 pm

I doubt this is true, it would only lead to mass link spam

Michael Martinez

01/09/2012 05:01 pm

So as I mentioned in reply to your post on the forum, the iframed page was not listed in any sitemap.  There was only one link to it, the iframe code itself. I applaud skepticism but people should easily be able to replicate this test and see the results for themselves.  Whether this will encourage Webspam remains to be seen.  It was Google's decision to follow the code and allow the link to pass value.  I am not recommending that anyone use this as a linking tactic.

Michael Martinez

01/09/2012 05:02 pm

This test was not designed to determine if the content in the iframed page would be associated with the iframing page.

Michael Martinez

01/09/2012 07:08 pm

I'm not sure people fully understand what I was testing.  The test only tries to determine if Google (or Bing) will follow the link in the IFRAME ELEMENT.  Google apparently did, and Bing apparently did not.  But as a consequence of following the IFRAME ELEMENT's link, Google also followed the link on the iFramed page -- and allowed anchor text from the link on the iFramed page to pass to the destination. That MIGHT imply that some PageRank flowed down through the IFRAME ELEMENT to the iFramed page but I don't think this test is conclusive enough to show that.  The algorithm may cut new pages some slack.  This test was designed to be as simple and straight-forward as possible.

Takeshi Young

01/09/2012 07:21 pm

Another possibility is if the iframed page was generated by Wordpress or some other CMS, then it could have sent out a ping when published, which caused it to be indexed. If links in iframes were to be indexed, it seems too easy to abuse.  That would mean links in ads and widgets (including the popular Facebook widget) all pass links.  The implications of that would be quite staggering.

Chris

01/09/2012 07:35 pm

It is nice to see someone doing an actual experimental test, rather than just guessing (as so many are likely to do). 

Christopher Skyi

01/09/2012 08:31 pm

Great headline but completely confusing description of the setup. Sounds like google went to the 'old site,' found the page with an inframe, went to that iframed page and followed the links on that page.  But it's not clear to me what 'value' was passed. Did the 'old site' pass value to the page it iframed? If so, that makes sense -- if the 'old site' linked to the page, that page would get the benefit of the link so why should it NOT get the same benefit just because it's been iframed? I'm not sure what the big deal is if that's what happened? Why wouldn't that happen?  Perhaps one could assume that Google just didn't have the technology to do this, but it seems easy enough giving the simple structure of an iframe tag/code (i.e., extract the link and follow it).  And it makes sense to do this, i.e., instead of directly following a link, the crawler goes the page that was iframed and then follows the link there. What would be news is if a high authority site linked TO the page that was iframed and the inframing site (the 'old site' in this case) got the benefit of the link, but Google would totally discount that. Have I got it all straight?

Michael Martinez

01/09/2012 10:28 pm

The iframed page (and the entire site) is a hand-coded, flat HTML text file.  No pings were sent in the production of the test.

Joe Youngblood

01/09/2012 10:51 pm

That's the big question. Does Google pass link juice from the page where the iframe is located or from the iframed page itself. It would be reasonable that they would send the link juice somehow to the iframed page since it would cut down on malicious websites iframing content of others and ranking because the iframe counts as a link and can't be nofollowed. Also note that Google owned YouTube switched from an embed code to an iframe code a while back. not sure that means much, but i found that move interesting.

Jeff Ferguson

01/09/2012 11:44 pm

Right now this feels more like Google would index via an iFrame, but we don't have a lt of proof that any "link juice" would be passed because there is no competition for that term and therefore no determination of ranking.  Now, if you rank this test against other pages that also used the iFrame technique, all of them with varying levels of PageRank, then we would have a better indication of if rank passed or if Google was just alerted of the page's existence via the iFrame, in much the same way Google will index links found in social media, but no link juice is passed because of a nofollow.

Fedor

01/09/2012 11:50 pm

Google spiders pretty much any valid url it can find, doesn't matter where it is, it crawls urls in javascript even if the code is never executed in the page. I think you need to look at your logs and see if it followed the url after it accessed the parent page or spidered it at a later time, this would give you a better indication of what to expect in regards to linkjuice. At this point I don't think you have anything new.

Michael Martinez

01/10/2012 12:37 am

Jeff, what you're proposing sounds like it could be shaped into an interesting test (although I would want to see it run against multiple queries in parallel).  Unfortunately, I don't have enough Websites to do something like that.  And such an experiment would still have to take precautions to minimize outside interference from the Websites that scrape iFramed pages and republish them.  It may not be practical to ever try to determine if any sort of weight is flowing from a link structure like this.

bbc world news

01/10/2012 05:32 am

Yes you are right, I'm also agree with you.

write my essay

01/10/2012 11:39 am

I think you need to look at your logs and see if it followed the url after it accessed the parent page or spidered it at a later time!!!!

Patrick

01/10/2012 05:04 pm

Right. I agree with some responses above that it makes sense that a crawler would not have any trouble reading and following an href in the iframe tag (if Google and others want to do that), at some point crawling the iframe target page and also passing link juice to that target page. Why not? If you post a straight dofollow link on your website, juice flows back to the site you're linking to, so why shouldn't juice be passed back via a link in an iframe app? I don't see a spam angle here that's different from regular links. What am I missing? Thanks for posting your experiment.

Patrick

01/10/2012 05:23 pm

I said href but meant src

pervezalam

01/11/2012 05:34 am

if it is true, then it is good for all SEO and webmsters as we always avoid Iframe to use it in web page.

Vincent B. Donadio

01/13/2012 04:38 pm

I'm sorry, but the problem with iframes seems to be that Google will treat it as a separate page... if you want that content to ONLY be observable through the iframe on another page, why not include rel=canonical and the url to the page the iframe is to be visible from? That will stop Google from indexing it as a separate page, and pass any SEO credit to the page it is framed within, will it not?

Alireza Sefati

01/13/2012 06:34 pm

I never use iframes to begin with but if I ever do, will nofollow that baby

brad mrumlinski

01/13/2012 06:56 pm

To everyone who is "skeptical" of this experiment, its actually very simple... google found the iframed page when indexing its parent page. It then crawls the iframed page as if it were a unique page, and then passes link juice from this page as it would any other page. Nothing here that can be exploited from what I can see... 

Sigma SEO

01/13/2012 07:07 pm

Things looks pretty well tested here.

Michael Martinez

01/13/2012 07:27 pm

I don't know if "rel='canonical'" would prevent Google from indexing a page.  Have never tested for that.  As we can see from comments other people have left here, iframes are not exactly a popular HTML tool.  My only intention was to confirm Jill's observation (which I did, to my satisfaction) that Google was using the iFrame SRC link for discovery.  Some other interesting questions have emerged in the various discussions but I'm not inclined to pursue this further (at least not publicly).  I should have written the title of the post differently because the value I was referring to was CRAWL and not PageRank (or even the anchor text that the framed page passed to its destination).

Vincent B. Donadio

01/13/2012 08:03 pm

Canonical is supposed to tell Google to index the canonical URL *instead* of the current URL, like if you had a bunch of URLs for sorting the same products by different parameters, the rel=canonical for each page would tell Google to index only the default sorting page. I understand what you were trying to achieve, though, and you did, in a very easy-to-replicate experiment. iframes are definitely not the best way to accomplish some things, but it's curious as to why Google changed YouTube embed codes to iframe.. and Facebook's applets use iframe, as well (multiple types, but iframe is a very popular one). Good job on the experiment, though.

Barbio

01/13/2012 10:23 pm

I agree and I think somehow it's logical. If you embed a page inside another through an Iframe you are showing content of both pages. The YouTube it's a good example. If you have a blog post with a phrase and a youtube video embed it should pass link juice.

Str82u

01/14/2012 03:30 pm

I thought everyone knew this, Google had been indexing the pages I pull in iframes, so much so that I've restricted access through robots.txt and the robots meta-tag tell them to not cache or follow.

Sytern

01/15/2012 03:45 am

Well, nice test and pretty interresting to know it. But personnally I still prefer not using Iframe simply from a coding point of view.. I think they're out since a long time. Thanks for the info

Mooersrealty

01/15/2012 12:57 pm

labeling the embed especially on video code is needed to help recognize what this link to a video, map, etc is. Even from Google's own treasure chest.

Patrick

01/17/2012 04:16 pm

HTML5 dropped frames but kept the iframe tag and made some modifications to it

Privatdetektiv

01/19/2012 10:03 pm

google found the iframed page when indexing its parent page. It then crawls the iframed page as if it were a unique page, and then passes link juice from this page as it would any other page. Nothing here that can be exploited from what I can see To everyone who is "skeptical" of this experiment, its actually very simple...

Local Internet Marketing

01/21/2012 07:56 am

Not a big secret, but a good point, nonetheless. 

Web Application development

01/25/2012 05:15 am

Search engines robots very much any legitimate url it can discover, doesn't issue where it is, it crawls web addresses in javascript even if the value is never carried out in the site. I think you need to look at your records and see if it followed the url after it used parents or guardian web page or spidered it later, this would offer you a better clue of what to anticipate in regards to linkjuice. At this factor I don't think you have anything new.

Seo

01/26/2012 02:23 pm

What impact do you think this would have on the use of twitter feeds as an embedded iframe?

Cambridge web design company

01/26/2012 02:27 pm

If the iframe destination had many links, could this penalise the site then?

audette

02/01/2012 03:51 pm

Barry - we conducted a test that validated Michael's findings: http://www.rimmkaufman.com/blog/do-search-engines-follow-links-in-iframes/31012012/

STR82U

02/01/2012 05:59 pm

There is no positive link juice passed here, when one of these types of sites comes online, like the situation below, the linked site on my side suffers. Not all the links listed like this contain iFrames, about half do.   "Links from thiswebsite.com to /something.php?news.1htt​p:/​/ww​w.thiswebsite.c​om/​some​-topic/​078​8.Junk-​Page​.ht​mlVia this intermediate link: htt​p:/​/ww​w.mywebsite​.co​m/something.ph​p?news​.0"   This has been going on at least 1.5 years, maybe more - I only took notice of this because of one particular site.

Michael From Seattle

10/01/2012 04:02 pm

I know for a fact that Google will IGNORE an iframe that points to a different domain, even a different subdomain of the same domain. Presently, at least, it will only crawl an iframe if it links to your own TLD. Google's wising-up. Presently MSN (Bing, Yahoo) is crawling the iframes regardless of their domains.

james

10/01/2012 04:47 pm

I think iframes are great for menus on your website. You can easily update your menu with an iframe, instead of having to update each page on your website with the new link. Time consuming if you do every thing by hand and don't rely on wordpress. An example: World-Of-Lucid-Dreaming.com ranks #4 for "Lucid Dreaming" and has the entire right side of their website in an iframe menu.

Carla Ackley

11/11/2012 02:20 am

Speaking of iframes. Does anyone know what an iframeping does? Can it be used in SEO? Or is this just something to keep a session open? For example: if (window.top !== window.self) { document.write('') }

NearlyNormal

11/11/2012 10:43 am

You would be much better off using a proper template. iFrames are a useless solution to that problem.

muara fatan

03/02/2013 05:04 pm

Do You mean that the link from where the iframe source or embed?

robert bob lalas

03/29/2014 08:19 am

you should have prof about this test could you look your up google webmaster if the links is in there. If it is in there then their is link juice. give screenshots.

blog comments powered by Disqus