Duplicate Content & Canonicals

[sbaylor]

I am a bit confused about canonicals and whether they are "working" properly on my site.  In Webmaster Tools, I'm showing about 13,000 pages flagged for duplicate content, but nearly all of them are showing two pages, one URL as the root and a second with parameters.  Case in point, these two are showing as duplicate content:

http://www.gallerydirect.com/art/product/vincent-van-gogh/starry-night

http://www.gallerydirect.com/art/product/vincent-van-gogh/starry-night?substrate_id=3&product_style_id=8&frame_id=63&size=25x20

We have a canonical tag on each of the pages pointing to the one without the parameters. Pages with other parameters don't show as duplicates, just one root and one dupe per listing,

So, am I not using the canonical tag properly? It is clearly listed as:Is the tag perhaps not formatted properly (I saw someone somewhere state that there needs to be a /> after the URL, but that seems rather picky for Google)?Suggestions?

[TomRayner]

Hi Darin

The tag is not working because if you go into Google and enter the URL: http://www.gallerydirect.com/art/product/vincent-van-gogh/starry-night?substrate_id=3&product_style_id=8&frame_id=63&size=25x20 you will see that it is being indexed on Google.

If it's being indexed, then it runs the risk of duplicate content issues.

The tag definitely does need the /> at the end, so the correct usage of the tag would be: rel="canonical" href="http://www.gallerydirect.com/art/product/vincent-van-gogh/starry-night" />

I think if you implement that small change, there shouldn't be any problems.

Hope this helps.

[Dr-Pete]

While the properly closed tag should have "... />", that's generally only an issue in very isolated cases. I've never seen it interfere with a canonical tag. It's a harmless change to make (and it is more correct), but my gut reaction is that this will make no difference. Google should be honoring these canonicals.

One odd thing I'm seeing. If I dig into the index, I'm finding the following page:

http://www.gallerydirect.com/art/product/vincent-van-gogh/starry-night?kwid=productads-sku%255EV00551@ADL4200C160@ADL4CG@ADL4F000@ADL4ML00-adtype%255EPLA-adid%255E25137313279

This may be an ad-tracking URL (?) and it's redirecting somehow (but not with a 301 or 302) to the non-canonical URL. This may be sending a mixed signal, and ideally it would redirect to the canonical version of the URL. I'm not sure where this version is coming from, so it's a bit hard to diagnose.

[sbaylor]

Dr. Pete:

I'm looking into it to be sure, but I believe that you are correct in that this is an ad-tracking URL.

A follow up question:

The URL that is the canonical version of each page would be in the format of

http://www.gallerydirect.com/art/product/vincent-van-gogh/starry-night

However, this exact URL redirects to one with default parameters for substrate, style and frame size:

http://www.gallerydirect.com/art/product/vincent-van-gogh/starry-night?substrate_id=1&product_style_id=1&frame_id=&size=25x20

Should we change our canonical from the first URL (without the parameters) to the second URL with the parameters? Or is that a moot point with Google?

[Dr-Pete]

Argh... sorry, I didn't even check/see that. Yeah, that may be a real problem - you're basically sending two canonicalization signals that are in conflict. Is there any way to hide the defaults? If the canonicals point to (A), but then (A) redirects to (B), Google may just ignore the canonical.

Unfortunately, your options are to either: (1) hope for the best, (2) canonical to the uglier URL, or (3) kill the redirect and set the default parameters on the server-side (without resetting the URL).

I am primarily seeing the canonical URL in Google's index, so I'm not sure it's actually causing you harm. It's just not an ideal situation.

[sbaylor]

Thanks, Dr. Pete.

I'll discuss the options with our dev team and see which one will cause the least amount of developer caffeine consumption.