Take the 2-minute tour ×
Webmasters Stack Exchange is a question and answer site for pro webmasters. It's 100% free, no registration required.

I work at a site which allows syndication of content (via API's and data dumps). We are finding that a number of sites which republish our content are appearing higher in Google search results, even though we're the original publisher. This is frustrating.

We are considering making rel=canonical part of our attribution requirements. Google says it's legitimate to use it across domains, and in syndication scenarios.

Have you done this, and does Google consider the canonical URL in search rankings? Will it help us in reducing such SERP "spam"?

share|improve this question
1  
What you're describing isn't spam. It's people doing what you're asking them to do--syndicating your content. Spam is unsolicited e-mail advertisements and webpages created for the sole purpose of bombarding people with ads instead of creating anything of value. If those are the types of sites syndicating your content, then you need to rethink your syndication model, or that will reflect poorly on your site (just through association). But simplying having better search ranking than you does not make a site spam. –  Lèse majesté Oct 9 '10 at 3:45
    
@Lèse really? these sites seem to be in direct contravention of the "little or no original content" rule established by Google itself google.com/support/webmasters/bin/answer.py?answer=66361 –  Jeff Atwood Dec 11 '10 at 3:12
    
@Jeff: Which sites are you specifically referring to? I'm talking about the act of using web syndication itself, which is something that many legitimate sites do. A spam site does not have to use syndicated content, and simply using syndicated content does not make a site a spam site (even if they achieve better ranking than you). Case in point, many major news publications use syndicated content from AP to supplement their own content. Is it duplicate content? Yes. But is it spam? No. And I don't think AP is promoting spamming either. –  Lèse majesté Dec 11 '10 at 10:34
    
@Lèse key words here being supplement their own content. If ALL the content is copied, what of value or interest is being created, exactly? –  Jeff Atwood Dec 11 '10 at 12:11
1  
@Jeff: It's not clear from Matt's question that those are the sites he's referring to. He simply stated that there are sites republishing his content (which is the purpose of providing a syndication API) which are ranking higher than the original content. That, to me, doesn't imply that these are (necessarily) spam sites. But maybe my interpretation of the question is incorrect. –  Lèse majesté Dec 12 '10 at 1:33

3 Answers 3

up vote 10 down vote accepted

Jeff is 100% correct in everything he said.

Another issue with requesting a syndication site to use <link rel="canonical" href="http://example.com/foo"> is that it tells Google that the syndication page should get no Page Rank and http://example.com/foo should instead get all of it.

That creates two major issues.

  1. The syndication page would not show up at all on Google searches because it has no page rank. The syndication site would not be the least bit happy about this. Making it unlikely that they would be willing to make the change if they even could.
  2. It might not affect your site the way you like because you effectively aren't getting linked to from the Syndication site. I would wonder how Google would handle this. It is true they allow for cross site rel="canonical" but I believe the purpose of that is for site migration and for having multiple sites under one host with the same content to have one defacto page versus a bunch of similar/same pages.
share|improve this answer
    
Those are some good points. I think syndication is one area where there's a legitimate reason for there to be duplicate content. In this case, it's better to leave the duplicate content alone and accept that that's what syndication is. Of course, ideally Google should give preference to the original page rather than the syndication partners' pages. Perhaps a new tag needs to be created that's something between rel="canonical" and the HTML5 <cite> tag. That way search engines can know which page is the original for legitimate duplicate content. –  Lèse majesté Oct 9 '10 at 4:01
    
confirmed, see my response from Matt Cutts below. –  Jeff Atwood Dec 11 '10 at 12:14

Adding another answer because I got a definitive response from Matt Cutts on this:

rel=canonical does work across domains, but it acts basically like a 301 [redirect], so the target site's pages would go straight to your site in Google. Any site using your content would basically be wiped out of the search engines.

Like Matt says the best way to think of rel=canonical is as a 301 permanent redirect.

Thus, requiring cross-domain rel=canonical as a set of attribution terms would be like asking them to 301 redirect to you! Ouch. :P

Knowing this, it is clear that rel=canonical is intended only for use on sites you personally have control over -- like when you move domains and you need one domain's content to replace the other.

share|improve this answer

My research indicated that requiring a link back -- and that the link NOT be nofollowed -- was by far the most important criteria.

If the "syndicating" site does not attribute the content with links back to the original that are valid for search engines to follow, search engines have a much harder time tracing where the content originated, and must apply complex "find duplicate text content across the whole of the internet" heuristics.

I'm not sure any more than that is necessary.

Related Matt Cutts video

http://www.youtube.com/watch?v=x8XdFb6LGtM

Matt said that it would be a good idea to use rel="canonical" to point back to the page where the article originated - just as he has often suggested that syndicated articles include conventional links (ie. an <a>nchor tag) pointing back to the original article.

Bear in mind that canonical isn't just slapping rel="canonical" on an <a> tag; it's more like this:

<html>
    <head>
         <link rel="canonical" href="http://example.com/foo">
    </head>
...

So it requires a different sort of work, you have to modify each page header. I'm not sure many of these "syndicators" will have that level of control versus a simple link (sans nofollow!) back to the source.

share|improve this answer
    
I'd also suggest reading Jeff's blog entry on this topic, Defending Attribution Required - blog.stackoverflow.com/2010/08/defending-attribution-required –  Scott Mitchell Oct 9 '10 at 22:24
    
@scott note that we did not originally require a followed link, but we changed that because Google's spider was missing stuff that was in our data dump some of the scrapers use ... and a missing link that is nofollowed does not help put it back in Google's index! –  Jeff Atwood Oct 11 '10 at 6:14
    
@Jeff: On a slight tangent, one thing that has bugged me is that links in a Stackoverflow answer have rel="nofollow". Shouldn't users with a certain rep get the benefit of no rel="nofollow" to the links they post? –  Scott Mitchell Oct 14 '10 at 18:30
    
@scott the website field in your profile, across any Stack Exchange website has the nofollow removed at 2k rep as a courtesy. –  Jeff Atwood Oct 14 '10 at 19:04
1  
@Jeff, I'm talking about the links in a Stackoverflow answer. For example, if I do a view/source on this very page I see the links in your answer (such as the one to YouTube) have rel="nofollow". I presume this is to dissuade spammers but at the same time it seems like you're missing an opportunity to improve the relevancy of search results for others, not to mention not "giving credit" (in Google's eyes) to the person who wrote the article/blog entry/etc. that is being linked to. –  Scott Mitchell Oct 14 '10 at 19:50

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.