Welcome to the Q&A Forum

http://moz.com/learn/seo/url says:

http://www.example.com/category-keyword/subcategory-keyword/primary-keyword.html

However I am wondering about structuring things this a little backwards from that:

http://www.example.com/primary-keyword/ (this would be an introduction and overview of the topic described by the primary keyword)
http://www.example.com/primary-keyword/secondary/ (this would be a category landing page with snippets from articles within the niche described by the secondary keyword, which is itself a niche of the primary keyword)
http://www.example.com/primary-keyword/secondary/article-title/ (in-depth article on a topic within the scope of the secondary, which is within the scope of the primary)

Where http://www.example.com/primary-keyword/ is the most important page targeting the most important URL.

Thoughts?

This concern is not lost on me, but it prompts two questions. 1. How big a risk is it? 2. What other technical factors could be at play here?

Here is an example of a site that uses a directory for its homepage, whose SERP snippet shows its full path:

https://www.google.com/search?q=skype

The full URL of a particular site's homepage is something like http://www.example.com/directory/.
The canonical and og URLs match.
The root domain 301 redirects to it using the absolute path.

And yet the SERP (and the cached version of the page) lists it simply as http://www.example.com/.

What gives? Could the problem be found at some deeper technical level (.htaccess or DirectoryIndex or something?)

We fiddled with things a bit this week, and while our most recent changes appear to have been crawled (and cached), I am wondering whether I should give it some more time before I proceed as if the SERP won't ever reflect the correct URL. If so, how long?

[EDIT: From the comments, see here: https://www.youtube.com/watch?v=z8QKIweOzH4#t=2838]

I am doubting it. Seems like an either/or thing. I'll probably do rel=canonical to the view-all.

Thanks everybody.

Now I just want to know if that usage of rel=canonical can coexist with rel=next & rel=prev.

"rel=”canonical” can specify the superset of content" -http://googlewebmastercentral.blogspot.com/2011/09/view-all-in-search-results.html

Consolidating for usability - I want to provide access to both formats. I hate the idea of noindexing a page that could be the target of inbound links.

Using rel=prev and rel=next would do the trick if it's still cool to point the rel=canonical to the view-all page. Anyone know?

Well, the individual articles are entries in a blog while the consolidated article is a separate page altogether. I could 301 the individual article URLs all over to the consolidated article I suppose, but I'd rather not for usability reasons. For the end users I think the ideal really is to have all the individual blog posts as they are, but provide access to the consolidated article.

Should I make the consolidated article a PDF download? Let them all be indexed without canonicals or 301s? I just don't know the best practice here.

I assume that you're using WordPress correct?

Nope.

Normally I would say use [rel=canonicals] 100% of the time. For instance if I had a blog post example.com/ABC/

I would obviously use a [rel=canonical]

Same thing if it is example.com/blog/ABC/

use rel=canonicals

To clarify: I am not talking about placing a rel=canonical on a resource located by more than one URL (eg. www.example.com/article-one/ and www.example.com/article-one/?fp=email). I am talking about placing a rel=canonical on each resource in a series, that points to a distinct resource that contains all of the content from all of the resources (eg. www.example.com/article-one/ -> www.example.com/article-all/, www.example.com/article-two/ -> www.example.com/article-all/, etc.).

I would of course index your house.

I don't know what this means.

In all honesty I would check using open site Explorer before you actually change anything. And make sure that your inbound links are not the problem that you're talking about.

What exactly should I check using OSE, and what actions should I take in response to what findings?

We've got a series of articles on the same topic and we consolidated the content and pasted it altogether on a single page. We linked from each individual article to the consolidated page. We put a noindex on the consolidated page.

The problem: Inbound links to individual articles in the series will only count toward the authority of those individual pages, and inbound links to the full article will be worthless.

I am considering removing the noindex from the consolidated article and putting rel=canonicals on each individual post pointing to the consolidated article. That should consolidate the PageRank. But I am concerned about pointing****a rel=canonical to an article that is not an exact duplicate (although it does contain the full text of the original--it's just that it contains quite a bit of additional text).

An alternative would be not to use rel=canonicals, nor to place a noindex on the consolidated article. But then my concern would be duplicate content and unconsolidated PageRank.

Any thoughts?

Thanks. Ok looks like my client's individual pages, including their page-specific rel=canonicals are in fact being indexed. So far so good.

In this case, to address your question, I had had a need for a cross-domain rel=canonical. I no longer need it because the case is closed. But there are real-world use cases for it.

Wix, where can I go to edit the canonical tags for the escaped fragment versions of the pages on my client's Wix site? What you are saying makes sense to me conceptually, but I don't see where I  can edit anything but the one canonical tag. Once I do, how do I verify that it's being served to Google?

I would also be curious to verify that this crawling solution will not get sites penalized for cloaking. It does appear that Google has Ajax crawling guidelines, but I need to be assured that these are being followed and working as expected. (Sorry for my ignorance here--I'm not a developer and the client's on a budget. The crux of the value Wix brings is that it's accessible to non-developers.)

These are technically different companies with different products, all of which are in the securities industry. They are each founded by different groups of individuals, however my client is common among them and happens to be a fantastic writer. Many of the articles would add value to the readers of some of the other sites. I am hoping to develop a common command center so that in the editor for a given article he is able to just check off which of his sites the article will be published at, and which is to be considered canonical. So the sites will have different aesthetics and navigation, product pages, and other company-specific content, and not every article will show up on every site, however many will show up at multiple sites.

The idea of phasing in common articles with the cross-domain canonical strikes me as wise, and then just noindexing the non-canonical versions if I run into trouble.

Budget not an issue, although skilled labor is.

Very helpful, thank you!

There is in fact a legal reason why the sites must be distinct from each other and strong marketing reasons why we do need more than one site.

I should mention that although the pages hosting the shared articles will be 99% identical, each site will have other content distinct from the others.

I am open to dropping my idea to share an article database between the sites and just having unique content on each, although I have to wonder what the use of cross-domain canonical is, if not to support this kind of article syndication.

That's my suspicion as well. Can anyone confirm with data?

Recently discovered this:

Indicate the canonical version of a URL by responding with the Link rel="canonical" HTTP header. Addingrel="canonical" to the head section of a page is useful for HTML content, but it can't be used for PDFs and other file types indexed by Google Web Search. In these cases you can indicate a canonical URL by responding with the Link rel="canonical" HTTP header, like this (note that to use this option, you'll need to be able to configure your server).

Link: <http: www.example.com="" downloads="" white-paper.pdf="">; rel="canonical"</http:>

Google currently supports these link header elements for Web Search only.

-http://support.google.com/webmasters/bin/answer.py?hl=en&answer=139394

Hmm, ok that's helpful.

The content would be identical with the possible exceptions of a very slightly different meta title and site footer.

What's my alternative to a setup like this? One site, one brand? Noindex the non-canonical article versions?

What I dislike about noindex is that it means inbound links to the non-canonical article versions bring me no benefit.

That makes sense. I am looking into whether any portion of our content can be made public in a way that would still comply with industry regulations. I am betting against it.

Does anyone know whether a page requiring login like this could feasibly rank with a strong backlink profile or a lot of quality social mentions?

Should have read the target:

"Subscription designation, snippets only: If First Click Free isn't a feasible option for you, we will display the "subscription" tag next to the publication name of all sources that greet our users with a subscription or registration form. This signals to our users that they may be required to register or subscribe on your site in order to access the article. This setting will only apply to Google News results.

If you prefer this option, please display a snippet of your article that is at least 80 words long and includes either an excerpt or a summary of the specific article. Since we do not permit "cloaking" -- the practice of showing Googlebot a full version of your article while showing users the subscription or registration version -- we will only crawl and display your content based on the article snippets you provide. If you currently cloak for Googlebot-news but not for Googlebot, you do not need to make any changes; Google News crawls with Googlebot and automatically uses the 80-word snippet.

NOTE: If you cloak for Googlebot, your site may be subject to Google Webmaster penalties. Please review Webmaster Guidelines to learn about best practices."

"In order to successfully crawl your site, Google needs to be able to crawl your content without filling out a registration form. The easiest way to do this is to configure your webservers not to serve the registration page to our crawlers (when the user-agent is "Googlebot") so that Googlebot can crawl these pages successfully. You can choose to allow Googlebot access to some restricted pages but not others. More information about technical requirements."

-http://support.google.com/webmasters/bin/answer.py?hl=en&answer=74536

Any harm in doing this while not implementing the rest of First Click Free??

What would you guys think about programming the login requirement behavior in such a way that only Google can't execute it--so Google wouldn't know that it is the only one getting through?

Not sure whether this is technically possible, but if it were, would it be theoretically likely to incur a penalty? Or is it foolish for other reasons?

My plan: 3 sites with identical content, yet--wait for it--for every article whose topic is A, the pages on all three sites posting that article will have a rel=canonical tag pointing to Site A. For every article whose topic is B, the pages on all three sites posting that article will have a rel=canonical tag pointing to Site B.

So Site A will have some articles about topics A, B, and C. And for pages with articles about A, the rel=canonical will point to the page it's on. Yet for pages with articles about B, the rel=canonical will point to the version of that article on site B. Etc.

I have my reasons for planning this, but you can see more or less that I want each site to rank for its niche, yet I want the users at each site to have access to the full spectrum of articles in the shared articles database without having to leave a given site.

These would be distinct brands with distinct Whois, directory listings, etc. etc.

The content is quality and unique to our company.

Is this a real thing?

Good idea--I'll have to determine precisely what I can and cannot show publicly and see if there isn't something I can do to leverage that.

I've heard about staying away from agent-specific content, but I wonder what the data are and whether there are any successful attempts?

First click free unfortunately won't work for us.

How might I go about determining how adult content sites handle this issue?

Understood. The login requirement is necessary for compliance with industry regulations. My questions is whether I will be penalized for serving agent-specific content and/or whether there is a better way to get these pages in the index.

Somehow certain newspapers' webpages show up in the index but require login. My client has a whole section of the site that requires a login (registration is free), and we'd love to get that content indexed. The developer offered to remove the login requirement for specific user agents (eg Googlebot, et al.). I am afraid this might get us penalized.

Any insight?

Didn't see this post. Interesting.

It depends on what you are using the rel canonical for. In our case we were changing domains entirely and I could not 301 the old URLs to the new (both domains located the same content), so I really needed the rel canonicals to work. In the long run I figured out 1. how to 301 the old URLs and 2. how to update the rel canonical on the homepage of the Wix site. All the pages on the Wix site I was working with were dynamic, so they didn't have unique, static URLs. Between the domain change, the 301s, the one rel canonical I was able to update, a little bit of extra keyword rich text on the home page, some directory submittals, and a handful of links, I was able to get a few #1 rankings, as well as some top 3 and top 10 rankings for keywords among those we were targeting. The client was on a budget so the targets weren't too ambitious, and we didn't have the resources to overhaul the site entirely. Also for some reason they were emotionally committed to hosting and managing their site via WIX. It turned out ok given their budget and goals.

Ceteris paribus, if you have a choice, and inbound traffic matters to you, you ought to stay away from Wix. The convenience Wix offers can be had without paying for it with the lack of flexibility and control tech-minded inbound marketers value.

Have you tried WordTracker? They offer "Search Count":

“Search Count (new tool):
For the Wordtracker data, the Search count is the number of times each keyword appears in our database of searches over the past 365 days. This constitutes just under 1% of all US search, and the data is gathered from metacrawler.com and dogpile.com.

The database is updated every day, and new data is between 15 and 30 hours old when it hits the live servers. If you’re searching using the Google data, the information presented is from the Google AdWords API. It’s Exact Match data by default, and the search volumes are from the last available month (in real terms this normally means the last calendar month).

There’s currently no indication from Google about their sample size for this data, or what kind of extrapolation may be applied to the data before it’s presented. There is a limit of 1,000 searches per month on the Google data – this is to ensure we can give an even service to all of our users.”

-https://keywords.wordtracker.com/help/metrics_explained

They offer a pretty good tutorial on how to use their tool to find profitable niches: http://www.wordtracker.com/academy/finding-profitable-keywords-just-got-easier

You can get the basic "WordTracker Count" (WT) metric for free from several different tools online, SEOBook.com's Keyword Tool among them: http://tools.seobook.com/keyword-tools/seobook/ 

Thanks. Figured out a workaround.

I've got a client with a Wix site and I need to edit their rel canonical tags--does anyone on here happen to know whether that's possible?

This was recently mentioned on another Q&A post; you'll want to check it out. You may not get "credit" for the backlinks to your competitors domain, depending on the situation.

I assume since it is from a competitor it topically related to your site? Why not figure out which specific URLs on the domain have the most backlinks and 301 them over to URLs on your domain with highly similar content? If you lack the content, write fresh content and pass it through the LDA tool a couple times to make sure it's highly relevant for your closest related KW target. 301 all other links to your homepage or something.

If the site has a lot of standalone merit, why not just capitalize on it and treat it like an independent lead or revenue source? Maybe put some well-placed links to your existing site throughout?

Anybody know of one?

What I am looking for is a measure of the desire to transact business that a given keyword reflects. So I want to know whether people are searching for information or to find a place to buy my product.

Yes, Xenu Link Sleuth is supposed to be able do this. Here is a helpful post that includes a section on redirects.

It looks like the Yahoo Site Explorer and Open Site Explorer also report the 301s they are aware of as backlinks.

I'd love for someone else on here to confirm.

Like you I've also seen 301's significantly positively impact rankings.

Good luck with your research!

...besides the MS AdLab tool (which seems to always be down)?

Thanks! I think I grasp the concept of an "exact match"—this is where the entirety of the anchor text exactly matches the entirety of the keyword, right?

EXACT MATCH:
Keyword: tennis shoes ireland Anchor: tennis shoes ireland

Now, in the example you give, two of the words from the anchor text match two of the words in the keyword. However there are words in the keyword that are not in the anchor text, and there are words in the anchor text that are not in the keyword.

PARTIAL MATCH:
Keyword:tennis shoes ireland Anchor: buy shoes in ireland

I am seeking clarification on other cases. For example, what about when the entirety of the anchor text matches part of the keyword?

PARTIAL MATCH?
Keyword:tennis shoes ireland Anchor: tennis shoes

Out of all the examples this one seems like it is most clearly a partial match, which makes me wonder about the example you give. Are you sure that it would qualify as a partial match, and if so, why? Do you have some experiment you could point me to—either its results or, better yet, its method (so that I could duplicate it)? Or do you have a citation from an authority somewhere (preferably "the horse's mouth").

Now, how about when only one word out of a multi-word anchor matches only one word out of a multi-word keyword?

PARTIAL MATCH?Keyword:tennis shoes ireland Anchor:** mythology in ireland**

I would love to find out which of these, if any, count as partial matches, and which, if any, don't count as matches at all.

This page says "...'partial match' means that at least one word from the query matches the anchor text.". Does this mean that at least one word from the query matches at least one word from the anchor text, or that at least one word from the query matches the entire anchor text?

Here is what Google says about the InAnchor operator. I am curious to know if this will show results with exact match or partial match anchor text. Does anybody know?

Thanks Keri; looks awesome.

So I had JS 302s and meta refreshes up for a long time, and eventually some of the links to the old site are now showing up in my Google Webmaster Tools dash for the new site. Just thought anyone else with the same question would like to know. I am going to look more into the cross-domain rel canonical thing too.

BTW Optify just released a new study on CTR: http://www.optify.net/guides/organic-click-through-rate-curve

This makes me wonder whether one can ever get extra mileage out of word merging, for SEO or branding purposes. keyphraSEOlogy.com sticks out in my mind...

You're getting a piece of this company, right?

So one of my client's wants to know why he shouldn't just hire GoDaddy to "do SEO" for him! He found this page: http://www.godaddy.com/search-engine/seo-services.aspx

Without actually trying it, It looks like a bunch of stuff I already do (rank tracking, PR tracking, traffic tracking, ROI projections), stuff I shouldn't do (site submissions to search engines), and stuff I probably don't need ("keyword wizard").

That said, I am intrigued by: the dedicated phone number with track and monitor (is it better than Hosted Numbers?), the business listing (is it a link worth the $6/mo?), site analysis and optimization, and ROI dashboard.

Does anyone have any experience with this? Does any piece of it have any value?