2

I'm helping on the SEO of an international group. Among others, there's a company.com website and a company.com.sa website (Saudi Arabia). They have got a new FAQ section on the .com website, in English.

They would like to use this content in the .com.sa website, in English as well. Same content on different domains... That would be bad duplicate content wouldn't it? So I have suggested that they use a cross-domain canonical from .com.sa to .com for those pages.

My questions:

  • With this cross domain canonical tag, will the content on both appear in Google's search results?

  • Will Google be able recognize the sites are both related to one another? and therefor understand the content is not duplicate?

  • Could I just use hreflang to show that the content on .com.sa is addressed to specific users (people in Saudi Arabia),e.g :<link rel="alternate" hreflang="EN-sa" href="company.com.sa" />

4 Answers 4

2

Duplicate content is no longer determined in a linear fashion. Today, duplicate content is determined using a semantic scoring method so that near duplicate content will still be seen as duplicate. This is because spammers would simply rearrange the content to avoid content as being flagged as duplicate.

As well, n-gram phrase recognition is used to determine language separate from any other language indicator. There is no distinction between English US and English Australia for example. This method is used in a variety of ways including knowing how to index the content. Given this, English, regardless of which English, will be indexed the same.

I realize this is not what you are doing but I tell you this for a reason.

Duplicate content is duplicate content regardless of what you do or the scenario. You cannot escape having duplicate content. The exception is language, however, it is for wholly and distinct languages such as English, Chinese, and Russian.

Your notion of using a canonical tag is correct. You would have to chose which set of content should be the original. This should be the first set of content posted and found by a search engine. If both sets of content are posted at the same time, then you can chose which set of content should be the original. Since the original content has value over the duplicate, you should indicate the original content where it would have the most value. Generally, this is the primary site. It is only the original content that appears in the SERPs.

3
  • So maybe in the end, I should use both solutions : only the ".com" pages appear in SERPs (because of the canonical tag) but when he clicks on the result link, the saudi user gets the ".com.sa" page thanks to hreflang Jul 31, 2015 at 8:48
  • @whatwouldyoudo Yes. You are exactly right! In fact, the hreflang along with other indicators does determine what data centers the site should be seen in which may also specify what result links should exist. This is an area I have not studied in a while and appears to be smarter these days.
    – closetnoc
    Jul 31, 2015 at 14:08
  • @closetnoc note that if you use canoncial tag on .com.sa pointing to .com than you should not use hreflang on that page (searchviu.com/en/hreflang-canonical).
    – Adam
    Oct 23, 2018 at 7:05
0
  1. Which ever page is referenced as the canonical is the one that will show in SERP's. The other will likely not.

  2. If the content of the page is very similar or identical, yes it is duplicate content. The domain hosting the content is irrelevant.

  3. No, hreflang is used to specify pages where the content is the same, but in alternate languages. In your case both pages are English, so that would be inappropriate.

My recommendation would actually be to do nothing. Since the content is hosted on a .com.sa domain, Google/Bing/etc. should be smart enough to know to show the Saudi page to users in Saudi Arabia, and the .com page to users elsewhere. The likelihood of this happening will be increased if you have International Targeting setup in Google Webmaster Tools.

So in the end, it will still be duplicate content, so it's unlikely that a user would get both pages surfaced in a SERP. But the correct one should be surfaced for the appropriate countries.

3
  • 1
    actually about 3) - it's not entirely right. Hreflang may be used for same languages too. It's on language-country base. Actually this is one of the great things about hreflang - the option to use it on multi-countries website(s) which also share the same language.
    – ePetkov
    Jul 30, 2015 at 20:24
  • I agree with everything you say, but it don't want the whole site to lose some ranking because of a penalty. That's the risk with this duplicate content don't you think ? Jul 31, 2015 at 8:42
  • @Barnettt is correct about #3. I will modify my answer accordingly. See support.google.com/webmasters/answer/189077?hl=en for more information. But no, duplicate content penalties do not affect the whole site, it's a page-by-page thing. Basically duplicate content is "penalized", because search engine's do not want the same content to appear multiple times in SERPs, because that would be pretty useless for users. For example, if you duplicate a page on your website, you'll find that the one that was discovered first by search engines will not be penalized, but the copy will be. Jul 31, 2015 at 13:41
0

Actually setting the hreflang in the .com as you suggest - <link rel="alternate" hreflang="en-SA" href="company.com.sa/FAQ-page" /> is a way to go.

Hreflang is setted on page by page basis, so if the FAQ is the only page in the .sa domain in English - you must set it as en-SA just for that page. The cleanest way is to set hreflang for the whole website, as it's the same company anyway. Just ensure you use the language right (set all English pages as English, all Arabic pages as Arabic).

Setting these tags sitewide will actually help Google to know, that both websites are connected, and the .sa domain is for the users in SA, while the .com domain is the worldwide version.

Correct implementation will show the .sa version in the SERP for the users in SA and the .com version will be in the SERP for every other country.

Canonical shouldn't be added to these pages, at it may result in conflict.

3
  • Google admits that much of the time, the language specified by the site is wrong (in the neighborhood of about 40% at one point) and that using n-gram analysis to determine language is the most accurate way of determining a pages true language and how it should be indexed. Another consideration is that there is overlap in many languages. Google does not rely on site language indicators for indexing content because that would presumably create a mess. The pages, being both in English, would be duplicate content as they would reside in the same index and have extremely similar semantic scores.
    – closetnoc
    Jul 30, 2015 at 21:39
  • I made a mistake in my third question. Hreflang should actually be added in the ".com" website, right ? Still, as I suggested, maybe you're right and this helps Google make a connection between both websites, but I'm not sure it'd be enough to prevent penalty on saudi pages, due to duplicate content. To push the logic of "connection" to the end, why not make a clear redirection with a canonical ? What "conflict" were you thinkning of ? Thanks Jul 31, 2015 at 8:37
  • If we must believe Google - the hreflang shows Google, that the given content is not duplicate, but rather a version for the given country. Of course clear redirection is also a way, but this way you'll serve the same page in SA SERP and the rest. And yes, the hreflang's should be added in the .com, pointing to the relevant pages in .sa. About the canonical conflict - implementing it wrong along with a href may create "loop".
    – ePetkov
    Jul 31, 2015 at 14:38
0

When you're using a multi-regional website it's really hard to avoid duplicated content therefore, the best way is to use canonical tags and as mentioned in the other comments, Google sees manipulated content as a duplication.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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