Google+ Authorship and Canonical Pages

[Panthro]

I've noticed that when verifying Google+ authorship on Google's Structured Data Testing Tool for various sites that only the www OR non-www version of a page will show verified authorship - not both. This, obviously, depends on which link is connected from the author's Google+ profile. Matt Cutt's site, for example, is only verified on the www version.

Does this mean that we should be verifying Google+ authorship to both www.mydomain.com as well as domain.com to give ourselves the greatest chance of having our nice little photo show up in the SERPs? Or, if our pages are properly tagged, will only the canonical version of the page be served in the search results (so we only have to G+ verify the canonical page)?

[goodroi]

I haven't noticed this before because I tend to pick one version (www. or non-www.) and stick with it. The other version is redirected so it should never appear in the serps.

I am not sure why you would want to verify both versions, at a quick glance that seems it could introduce possible dup issues. I haven't had a lot of sleep the last few days so maybe I am overlooking something.

[rainborick]

I'm sure that the reason you're seeing this is that only one version of the URL is actually indexed, so the Testing Tool doesn't have any reference data for the non-canonical URLs. Even when webmasters don't provide canonicalization signals via Webmaster Tools, the rel="canonical" tag or 301 redirects, Google's own canonicalization processes will filter out most duplicate pages.

[Panthro]

That's what I figured. But then I'm trying to think of any examples where I've seen both www and non-www pages from the same site returned in the SERPs - ie, examples.com/pages.html and www.example.com/other-page.html. I don't think that happens, but I haven't really paid attention to that much, tbh.

It's a case like that were I wonder about verifying authorship on both versions of a page, particularly if pages are verified on a page-per-page basis (as opposed to one Authorship connection in a universal footer) or multiple authors on a single domain.

[rainborick]

When left on its own, Google's canonicalization is generally done at the URL level, not on the domain. So it's not unusual for sites to have pages show up for variations of their domain name/URL and sometimes run into duplicate content issues. But, as I say, when a URL *has* been canonicalized by Google, it's no longer indexed, in which case, Google's tool won't find the underlying data required for author verification for that URL. You can only verify the canonical. And this all assumes that Google has ever seen the non-canonical version of the URL. Naturally, you'll also be unable to verify a URL that's never been crawled.

[ergophobe]

Just recently I have looked a few sites that were doing poor canonicalization at the server level (accessible by www/non-www), but with a site: and site:www search, all the results came up only for their intended domain if
1. They were consistent with their internal links
2. They had verified their canonical domain in GWT

Since they won't show up in the SERPS for their non-canonical domain (even though it's accessible), I would assume that Google can also sort out the author issues.

Of course in cases where they've done nothing to canonicalize their URLs and aren't even consistent in their internal use, it's all over the board with URLs split in the SERPs and dupe content issues that would probably also play into the author scenario you're thinking of.

[lucy24]

I'm trying to think of any examples where I've seen both www and non-www pages from the same site returned in the SERPs - ie, examples.com/pages.html and www.example.com/other-page.html. I don't think that happens

It happens if
#1 there exist links to both forms of the sitename
and
#2 the site owner has not checked the box in WMT specifying which form they want to use in indexes

On #1: all it takes is one page. Once the googlebot is on, say, example.com/widget.html instead of the preferred www.example.com/widget.html, it will then follow all internal links from that page and duly record them as example.com/all-other-pages-on-the-site

As far as anyone knows,
example.com/widget.html = www.example.com/widget.html
is just as much Duplicate Content as
www.example.com/widget.html = www.example.com/foobar.html

[Panthro]

So I'm hearing that verifying the canonical version in WMT should sort everything out (if it wouldn't on it's own, anyway). And that then it shouldn't matter if the URL used in Author verification is www or non-wwww.

Do same rules apply for Author verification on something like www.domain.com/category/index.html (versus www.domain.com/category/)? I would assume it would just get sorted out.

[rainborick]

The most important point is that best practice is to choose a canonical domain for your site by making sure your server responds with 301 redirects for all non-canonical versions of your domain name. This insures that all of the search engines understand your preferred URL, and will also (eventually) resolve non-canonical URLs that may have been previously indexed. Many people also recommend installing 301 redirects for the index-dot-whatever URLs. You're right that the search engines generally handle those URLs without any help, but I respect those who recommend the redirects so I wouldn't quibble.

As to Author Verification, once you have effective canonicalization in place, you only need to verify the canonical URLs because those are the only URLs in the index and are the only URLs users will ever see in the search results.

[ergophobe]

I wouldn't count on WMT sorting everything out.

My point was that I'm seeing a lot of amateur sites (even though usually created by a "professional web designer") that have the same pages listed with and without the www and that you see this if you do a site:example.com search and compare it to site:www.example.com

If you give Google at least a clue or two - WMT and having consistent internal linking - it will usually sort things out for you. If you don't give it some help, it will definitely NOT sort it out for you. But I think you shoudl get as many cues correct as possible.

[lucy24]

I'm seeing a lot of amateur sites (even though usually created by a "professional web designer") that have the same pages listed with and without the www and that you see this if you do a site:example.com search and compare it to site:www.example.com

Site design has nothing to do with it. The with/without www. element happens at the hosting level unless you've explicitly taken one or both of two actions:

--Tell the host to redirect to your preferred form (there's usually a click button somewhere) -- the easiest way, though it can lead to duplicate redirects since it happens before the request reaches your site

--Include a redirect in your own htaccess to canonicalize the domain name.

When you're talking amateur sites-- or even small-to-mid-sized professional ones-- you have to assume shared hosting. But your average amateur-- or even {et cetera}-- doesn't know from search engines and Duplicate Content. They're more likely to think it's a convenience to the user, letting them get to the site by whichever name they happen to prefer.