Is your live site supposed to have rel canonical tags?

[Chase_Cleckner]

I recently started working for a company and got them to use Moz and I have found that our secure site and our live sites are creating "duplicate content" according to the Crawl Diagnostics feature. On our secure site we have rel canonical tags pointing to our live site. I'm not super familiar with rel canonical tags, but our developer says we're doing the right thing. Would love any insight you guys may have if this is actually duplicate content or not. Thanks so much!

[CleverPhD]

Agree with Dave's comments.  1) Get the syntax updated on your canonical links at a minimum.  2) Yes your canonical solution will "work", but it is not best practice.  This "solution" is really a last resort.  I would try and push to move away from using canonicals this way. You optimally want 1 URL.

Just to add some color, a great / classic video on this was made by Matt Cutts.  He gives all kinds of examples where you could have duplicate URLs, i.e. www vs non www subdomain, sorting parameters added onto the URL, different file extensions,  capitalization changes, etc.  He then gives 3 options to fix them.

1. Best practice: Fix your site where you only have one URL per content item and link to it consistently (Best solution)

2. Use 301 redirects to consolidate to one URL (Next best solution)

3. Use a canonical link, if you cannot do 1 or 2. (Last resort)

Note that Matt says that they treat a canonical as a strong suggestion (it is treated similar to a 301), but they do not always have to follow it.  He repeatedly says, use the first two options, and would NOT recommend a canonical as your best or first option.

My favorite quote is at 2:24 in the video, "Developers keep SEOs in business"

What your developer may notice is that Matt does say that using a canonical link for consolidating http and https will work. No one here would say that it would not, it is just not optimal.  Sure, you can use a pair of scissors to cut your lawn, "it will work".  It doesn't mean it's the best idea.  I would think any developer worth his/her salt would want to have "clean code" and having duplicate URLs is not "clean" by SEO standards

Ok, so now you need to go back to the developer or your manager with an argument that is stronger than just, "Well, some random dude on the Moz forum said that Matt Cutt's from Google said it was preferred not to use a canonical link even though it would work".  I would never want to leave you in such a position. Here is what will/can happen over time if you stay with your current setup.

1. Report consolidation issues.  When you look at GA for traffic or OSE for links, any spidering tool for technical issues, social sharing counts, you now have split data for any given page potentially.  Sure there are ways around this, but now you have to spend all your time "fixing" reports that should not be broken to start with.   Trust me, this will come back to bite you on the bum and will cripple your efforts to show the efficacy of your SEO work.  Now who really wants that?

2. Link juice consolidation issues.  With any redirect - you lose a bit of link juice.  If you have links to both sets of URLs, any single page is not getting as much credit as it should.

3. Down the line 301 redirect bloat.  If you ever change anything and need to setup a 301 redirect, now you have to setup 2 of them and having too many 301s can negatively impact server performance.

One last thing.  If you can get the URLs consolidated into one using 301s etc.  Go with the https   That is the way that we are headed with the web and so you might as well get going in that direction.

Good luck!

[Chase_Cleckner]

I really appreciate the response and the added information. I guess we will see if anyone else responds!

[BeanstalkIM]

I'd be interested in hearing what someone else has to say about the way the canonicals are coded.  You're doing yours similar to the way I do DNS Prefetching with the double slash to start the URL:

It works fine with prefetching as all the browser needs to do is find the IP of the domain but I'm not sure here how it'll handle sub-directories including www and I hate variables even when they're "it should work".  The more common way to canonicalize your secured page would be:

/>

I'd be interested to hear if anyone has any direct experience with this but at the core of technical SEO issues I always lean to "most common usage" and "how Google shows it in their examples" just to make sure there is minimal chance of hiccups or issues.

That aside though, the developer is right though I'd always still prefer to just see the pages at a single URL.  Since that can't be done however ... canonicals are the way to go.

[Chase_Cleckner]

That is correct! Here is an example of two URL's of what i'm talking about:

http://www.agroup.com/blog/5-signs-of-a-good-clientagency-relatoinship
https://agrouptt4.secure2.agroup.com/blog/5-signs-of-a-good-clientagency-relatoinship

Does this help clarify my question? I hope so!

[BeanstalkIM]

I'm not sure I entirely understand the scenario so let me note how I'm hearing it to make sure my understanding is correct to put the answer into context.  Please do let me know if my understanding of the scenario is wrong as that may well change my thoughts on it.

You note that your secure site and live site are creating duplicate content.  Of course a secure site can be live but I'm taking this to mean you have an area behind a login.  That it's creating duplicate content is making me think that a lot of the core information is the same and I'm guessing many of the same pages.

If this is all correct and you can't put the duplicated pages onto one URL only then the canonicals are the way to go and your developer is correct.