PR passing through to canonical urls

[cupid]

i understand that the canonical tag was designed to allow webmasters to avoid duplicate content issues. but what about campaign tracking using hard links?

one of my biggest problems is not knowing which hard link trade is profitable. what i'm thinking of doing is this.

have each trade link to separate copy of index.html: index2.html, index3.html, etc. which would all reference canonical index.html

would this really pass PR just as well as if all those links would point to index.html ?

[aristotle]

I'm not sure I understand your strategy. Are these diiferent copies (index2, index3, etc) on the same domain, and also, have you put noindex metatags on them?

Also, if you link from each copy to the original, wouldn't that pass the PR to it?

[cupid]

they are exact copies on the same domain, the only difference being that only index.html is indicated to google as original content, whereas all copies are referencing it with the canonical meta tage

copies dont link to original, they're identical to the original (other than some embedded campaign trackers)

[aristotle]

I'm not an expert, but it seems to me that you should noindex all of the copies. Also, if you link them to the original, with dofollow links, that should pass their PR to the original. I don't know if the canonical tag transfers PR, but wouldn't have thought so.

[tedster]

The canonical link does help Google (and Bing) to consolidate all manner of link equity at the canonical version of the URL. That assumes, of course, that link equity is actually being transfered by any particular link. In many cases, when the website arranges for a link, then it's not being valued as a PR transfer "vote" - no matter what the destination URL. This is because such links are no longer "freely given editorial votes."

The situation that cupid describes in this thread is pretty common. There's one solution I've often recommended but rarely seen in action. We know that query strings (index.html&type=blpartner) can cause duplicate issues and that's one reason for using a canonical link. However, if your technology allows you to script this, why not use a hash tag (#) as an identifier instead of a query string (?) or a different file name altogether?

Since hash tags are not considered a different URL (except in the case of an AJAX !# tag) there is no canonical confusion, and the identifying information could still be extracted by a script on the server or the analytics program.

[aristotle]

The canonical link does help Google (and Bing) to consolidate all manner of link equity at the canonical version of the URL

I don't understand how this works. Suppose I create two pages with the same content -- call them Page A (the original) and Page B (the copy). Also suppose that Page B has a canonical tag pointing to Page A.

1. Does all of the link "juice" coming into Page B get transferred to Page A through the canonical tag?

2. Also, what if Page B contains some dofollow external links to other pages? Does all of its link juice get transfeered to Page A in this case, or does some of it go to the other pages?

3. Also, what if Page B has some additional content that isn't on Page A? How does the canonical tag work in this case?

[tedster]

Does all of the link "juice" coming into Page B get transferred to Page A through the canonical tag?

Yes, that is what Google has said it does.

Also, what if Page B contains some dofollow external links to other pages? Does all of its link juice get transfeered to Page A in this case, or does some of it go to the other pages?

Do you mean that those links don't occur on Page A? Then I'd say you are getting into a gray area where the use of the canonical link might not be trusted as accurate.

Also, what if Page B has some additional content that isn't on Page A? How does the canonical tag work in this case?

I haven't analyzed a case like this "in the wild" - but I assume you mean that the additional content was actually ranking, under the URL for "Page B". I'd guess that some of that content's power might now be attributed to Page A.

But since the content doesn't actually appear on Page A, that would create a poor user experience, something like the old "these keywords only appear on links pointing to this page" situation. So I'd expect the ranking power of that content to be dampened way down, assuming that Google decides it can trust the canonical link.

[aristotle]

Thanks tedster. It seems to me that Google needs to provide more precise information about how to use this canonical tag.

For example, suppose I publish a 1000 word article (Page A), then later publish an expanded version (Page B) that includes all of Page A plus an additional 500 words appended to it. Should I use a canonical tag (pointing from B to A) or not? I don't think Google has provided guidelines for a case like this.

[tedster]

Is there any reason in the case you describe, that you can't publish the newly expanded version of the article at the original URL? That seems quite natural to me.

[g1smd]

In the case of an expanded article I don't think the canonical tag is the right thing to use.

In the case where URL A is canonical but URL B has extra links pointing elsewhere, that's certainly an interesting situation and is probably one where Google would not trust the canonical tag too much. It sounds like a case where using the meta robots noindex tag might be a better solution.

[aristotle]

Tedster -- What I've actuallly thought about doing is writing a much longer version of one of my old articles and putting it on a different website. But that's in the future, if I do it at all. But it gave me the idea for a hypothetical case as I was reading the thread. At any rate, I think that Google should publish more detailed guidelines on how to use the canonical tag.

[tedster]

Some of what Google has already said has confused me - especially about using the canonical link for pagination. Clearly paginated product "sorts" are radically different pages - but Matt Cutts has suggested using the canonical link element to point them all to page 1.

[g1smd]

I have a situation where users can paginate content as short or long pages or can show all the content in one long huge page. I use the canonical tag to point to the long page and that is what gets indexed. Google can't get to paginated versions from anywhere within the site. They only see those if they follow a paginated URL link from another site should someone happen to post one elsewhere. Even then, there's a box at the top of every page with a link to the non-paginated URL and the text "URL for this page" next to it. Some people still cut the URL from the browser address bar rather than take the hint.

[ThisIsOli]

This seems like a complicated way to drop link juice. Anyone going to test this?