Canonical Caused Duplicate Pages In Google

[c41lum]

Can canonical cause a dup problem.

After much deliberation we decided to lower case all of our product page URLS. We have canonicled from the Upper-Lower case URLS to the new lower case ones.

After checking WMT today it says we have 18,000 duplicate tittles. Google seems to have taken no notice of the canonicle tag at all! has this happened to anyone else and is there a fix?

We are internally linking to the lower case URLS.

[rainborick]

The reports in Webmaster Tools are often several days out of date, and mostly advisory. It can also take quite some time for a rel="canonical" tag to be accepted and integrated into their index. I would wait a while and then check the site: operator to see if the old, mixed-case URLs are disappearing to determine if the canonical tags are being accepted in your case.

[c41lum]

The strange thing is we did everything exactly as google recommends.

We decided on a lower case URL structure because this would cause less 301/302 issues with our internal and external linking. We then canonical the old indexed urls to the new ones.

Now POW~! duplicate tittles .....

[tedster]

This experiences reinforces the idea that you should always handle as many canonical problems directly on your server as you possibly can. If the mixed case URLs still resolve 200 OK, then technically you still have duplicate titles. If they redirect, then the duplicate title messages should vanish eventually.

[c41lum]

We did it this way so that we wouldn't be doing tens of thousands 301 redirects, it would also help to clean up our site up and stop us internally linking to 301 redirects.

Surely by classing these pages as duplicates this goes against everything their engineers recommend i.e [googlewebmastercentral.blogspot.com...] . (sorry for the link)

Am I missing something or is G telling us one thing and totally ignoring there own advice.

[tedster]

What you've got is an FYI message in Webmaster Tools - that's all. And thought we may sometimes think of Google as this monolith, in reality the internal teams are often quite disconnected and they do launch changes without a lot of internal vetting. It's perpetual beta.

Has your traffic been affected? I'm guessing the answer is no.

Does your internal navigation still show mixed case URLs? If so, I'd definitely fix that. And then I'd at least place 301 redirects on any URLs that have external backlinks.

[c41lum]

The thing is, that was a major post in the webmaster central blog at the time and as far as I know one of the definitive pieces of info that google have produced on how to handle the canonical tag. That aside, we have only seen a small drop off in traffic, nothing major.

We have sorted all our internal linking out, all internal links are now lower case.

Shouldn't any 'juice' still flow through the canonicle pages.

[tedster]

Yes, Google says they will treat links to a canonically incorrect URL as if there were a 301 redirect, and as long as they accept the "strong hint" shown in the canonical link tag. And up to now, that is what seems to happen - at least in the great majority of cases.

[netmeg]

Have you heard, or is it likely, or even possible, that a canonical tag could cause Google to change a page title?

I have a mobile version of a page. There's a browser detect on the regular page (which has been indexed for years with the same page title) which pops the user over to the mobile version. The mobile page has canonical tag pointing back to the home page. Today, all of a sudden, the home page has the Example.com - Mobile title in Google. No loss of rank or traffic. Just the title change.

[tedster]

Interesting report, netmeg - no I hadn't heard of a canonical tag creating a title change. Just yesterday we had another unusual report of a Google title tag change [webmasterworld.com]. Maybe it's a current area of focus and testing.

[aakk9999]

@c41lum
I would not worry about duplicate titles, just give it a bit of time. We have recently done similar on one of sites (we implemented canonical on cca. 700 pages) and a few days later we got reported 600 duplicate titles. Today, ten days down the track, the number of duplicate titles has fallen to just 17 - Google sorts it out as it re-crawls the pages.

From the number of duplicate titles you have, it is obvious that your site is much larger so depending on your Google crawl rate it may take longer time for Google to go through all pages and sorts them out, but it will happen.

[added] As you are internally linking to lower case URLs, Google suddenly found lots of new URLs and these URLs have canonical set to its own URL. In order for Google to recognise canonical, it has to re-crawl what it knows as EXISTING mixed case URLs to see canonical tag on these pages now pointing to newly discovered lower case URLs. This is why it takes longer - seems Google always gets busy with new URLs first and as it re-requests previosly existing URLs the number of duplicate titles will start to fall.

In fact, monitoring the number of duplicate titles over the next few weeks (months?) can give you another indication of the rate with which Google re-crawls your previously existing URLs on the site. You should also be prepared for some duplicate titles to take long time to dissapear if the page was seen as "low value" page and as such not crawled very often by Google.
[/added]

[jdMorgan]

Great description, aakk9999.

Folks seem to expect "instant results" these days, and you summarized the actual process and the solution (wait awhile) quite well.

Jim

[c41lum]

Thanks aakk9999, some good info.

When I checked WMT today our crawl has gone through the roof,
a few key product pages have seen a drop of 3/4 places.

Ill keep you posted on any other changes I spot.

[g1smd]

I made some changes to several sites towards the end of 2009. These are changes that WebmasterTools still hasn't fully digested here in August 2010.

[netmeg]

Interesting report, netmeg - no I hadn't heard of a canonical tag creating a title change. Just yesterday we had another unusual report of a Google title tag change [webmasterworld.com]. Maybe it's a current area of focus and testing.

It lasted two days, and then went back to normal.

<shrug>

[c41lum]

Im seeing a lot of canonical Duplicates in G today.

Blue-Widget.html (now with canonicle to below)
blue-widget.html

We have also had some big drops in positions.

[setzer]

The data in WT is way off the chart! Just yesterday it reported I had 47000 impressions, up from a 15,000 average. There was no spike in traffic that day. I don't even really pay attention to it anymore.