Forum Moderators: open
We are about to do a huge update on our site and I think I see an opportunity to make good use of canonicals to create a site structure that will help the user as well as our PR.
We are a commercial site so there would be the obvious choices: bigwidgets.domain.com, littlewidgets.domain.com, questions.widgets.com, etc... I only have a few broad ones in mind and I am definitley not going to overdo it and create a canonical for each keyword or anything like that.
The main reason I am considering using the canonicals is this. We are a niche market with limited similar sites but there are many high PR non-commercial sites that would link to our content if it didn't look as commercial.
What I am thinking about doing is creating a quasi-independant site on a canonical using our vast educational content. By quasi I mean this:
1.Every page links home and page structure/look/feel is consistent throughout the site.
2.Commercial navigation would be trimmed down substantially and in favor of structure to help the user learn more from our non-commercial content.
3.Where it would be welcome and appropiate for the user I would include links into the commercial pages.
The quasi idea would allow us to generate a huge amount of non-commercial looking content that would have buried the products in our shopping pages. Within this content we could work great link placement for our reciprocal link partners and further increase our traffic and popularity with quality links from them.
A structure like this seems to make perfect sense to me as a human. We are showing that we are not just on the web to sell something. We are here to educate as well. The canonicals serve as suggestions of what the site regions are about as well as seperating them structuraly and lend PR to our main root.
I am not really looking for a rubber stamp for this idea, just a general discussion of using canonicals this way. Does this sound like an intelligent use of canonicals? Does anyone see ways this could be abused and an innocent site banned in the cleanup?
For clarification I am not talking about doing this:
1.Duplicate content. There would be zero duplicate content.
2.Keyword canonicals.
If the focus is entirely different, as you've described, and the cross-linking is done prudently, it seems that there's very valid reason for the separation.
I am a huge fan of canonicals, more for structure and themes than anything else, although I may enjoy a few benefits with the linking.
Your plan from what you’ve shared sounds perfectly acceptable to me.
When I am structuring a site I equally look at the themes and the content. In using canonicals it’s really dependent on what the topic is. Even for small sites I can usually benefit from a directory canonical, a ‘housekeeping’ canonical and the news, articles, archive, self-promotion canonical. From there it’s really dependent on how I’m structuring the rest of the site. I may need a shopping canonical if I have a whole storefront set up, I may need a community canonical if I’m building a community. If my whole site is a shopping mall I may then divide my canonicals for the audience or the product or the manufacture. For me it’s about what makes sense to group together and what can hold it’s own.
It really doesn’t take a dozen pages either to make a canonical. If all the content fits on a page or two for that canonical than so be it. That’s when the content becomes so important.
As you’ve suggested here thunderpaste, it’s risky to set up a canonical just to promote a keyword and then dump mirrored content in.
Is is a nominalized form of the adjective canonical [hyperdictionary.com]? I´m just wondering since it seem to be a strange way of using the word. When searching for canonical in Google it finds mostly pages where it is used in the canonical sense ;).
An explanation would be very much appreciated.
Andreas
I always assumed it was risky with search engines.
But we would like to do it with say.. our duscussion board
Paynt and others...
Would it all be ok if...
1. The directory www.mysite.com/forums/ was robotted texted out
2. We create a subdomaincanonical forums.mysite.com
3. Some Internal links and external links still go to www.mysite.com/forums (there are a lot already
i.e the concern is that using www.mysite.com/forums and forums.mysite.com would go to the same content. Would that not be seen as duplicate content?
We have always had them at www.domain.com/forums/
but we are considering to moving them to forums.domain.com to more clearly separate the gossip/yakking from the published and reviewed content.
Basically we want the discussion/forum as a subbrand, not part of the core brand.
chiyo, you asked a question quite a while back about the forums, I guess you didn't make the move. As toolman pointed out in a recent thread on the subject, there are technical considerations.
Actually, a subdomain is considered a whole separate domain - some hosts even require additional accounts with separate IP numbers for them. So it's linking between separate sites rather than within the same site, so keep that in mind. There are also implications about tracking and logging.
You can always use 301 Permanent redirect, but about mod_rewrite from a directory to a subdomain - I'd start a thread and ask specifically about that in the Website Technology Forum - we've got some experts around on that.
As far as search engines go, some people don't like to use subdomains and wouldn't at all, while others love them. There's been a lot of search engine spamming that they've been used for, and there is a search engine that really got down on them for that reason. There have been some posts about it, some not too long ago. Probably Inktomi, if I remember correctly. The current concern would be with Google, about cross-linking for those who lean toward "paranoia" about avoiding penalties.
Here are some past threads for reference:
What are Wildcard Domains [webmasterworld.com]
Redirecting to Subdirectories [webmasterworld.com]
A recent discussion:
Subdomains? is there any advantage? [webmasterworld.com]
We have had some very good discussions where the topic of canonicals has come up. A site search is a great way to bring that information to the top. It would be useful to anyone that is interested if you read those. A few years back there was quite a fuss about canonicals and spamming. As with anything it’s the not the use of the technique itself that’s spam but the way in which people thought to use it that was of course spam.
Neither Google nor any other search engine that I have run into is going to mess with you at all for using canonicals to structure your site. Not one person has come forward here in the forums that I’ve seen, or found that has had any problem at all with canonicals. I’m getting to think it’s really silly all this fear and paranoia over a structuring issue.
Google uses canonicals and for a very good reason whenever it’s used, also universities and large corporations. It’s a very common practice it’s just that some sites, from what I can tell in years past were using canonicals instead of separate domains to spam. It was the interlinking and crosslinking issues that are where the problem comes in. People may need more training in internal linking and navigation along with crosslinking and interlinking.
It always comes down to what your whole motivation and purpose is. If you are out there to spam then you might pick up early steam with heavy cross/interlinking between canonicals or separate domains, I’ve seen that techniques used and continue to work quite successfully for many sites. For most heavy spammers it’s too time consuming to set up canonicals when they are all set for multiple domain spamming. But see what I’m saying here? Now I’m talking about spamming not canonicals, one does not lead nor should people continue to think one leads to the other. It’s silly and I’m tired of it. I’d love one person to come forward who has had any problems because of their use with canonicals. It’s not going to happen because from my experience in talking with them they are quite content with the way things are and when it’s good you may not want to talk about it, why chance rocking the boat?
I can hardly wait until we can talk calmly about structuring with canonicals without all the warning bells and whistles. I’m normally cautious and am one of the first to warn folks to be careful and think about what they are doing before they act but this is getting beyond silly now all this fear and warning blather. Though, I think we need and you will continue to see me warning folks about their linking practices, geez that’s what we need to keep educating.
I use the term canonical because it was the first term I heard used in reference to what I later heard referred to as vanity, viral, sub-domain, third-level domain, sub-directory (that one gets very confusing) etc. A canonical by any other name… Since I find the discussions interesting I like to get my use of the word canonical in so I can better search here for discussions where the subject comes up. Many hosting companies, especially larger ones use the term. You can look up - canonical host - for dozens of examples or even – canonical hosting. Hosts appear to set them up differently as well. Some charge a set up fee and some don’t. Some charge an extra monthly or yearly fee and others don’t’. There appear to be many ways to set them up, which is very interesting and worth figuring out before you jump into it. As Marcia has rightfully pointed out, there are technical issues. We have talked about those and Marcia offered a few good discussions where some of the issues were discussed.
I thought this was a great discussion:
Subdomains and Navigation [webmasterworld.com] Absolute, Relative paths and other mysteries
Anyway, I enjoy the discussion and I suppose the debate although I’m of the contention that we don’t really have anything to debate about.
For example, I have fairly graphic intensive pages. So I have all the images in one subdomain where they aren't cluttering the main logfile and I can effectively ignore them. Instead of getting ten lines in the logfile every time the page is called, I get one. Then recently I developed a new subsection of the site which isn't directly related to the other stuff. Being new it doesn't get much traffic and would be swamped in the main log file. Instead, it's in its own subdomain where I can trace it's growth and see how the pages work together and suchlike seperatly from the rest.
I know there are other ways of filtering log files but I don't usually use an analyser on them, just throw them into excel, so having them pre-sorted like this is great. It also seems to be working that the things I want seperate log files for are things that make sense seperated out for visitors also, so it works from both ends :)
For Google:
I wouldn't worry about this,.. [webmasterworld.com] as Googleguy said.
Basically those subdomains are seperate sites and for Google everyone was paranoid that too much interlinking between seperate sites could possibly lead to a penalty.
I think the conclusion is that Google sees the links between and within subdomains as "internal links" (coming from within the same site).
For myself, I see no advantage in subdomains or "canonicals". It would mean more administration, more statistics to check and less singular, clean and net branding for my url, but I'm willing to be convinced of the contrary.
Added: just to convince myself to the contrary:
Do a search for "Institute of Technology" and you will see MIT, Massachusetts Institute of Technology, with several listings of its subdomains and several different DMOZ listings as well. So there are advantages to use subdomains.
the luxury of being too well placed.
I guess I better get the reputation of the MIT first..
Same goes for multiple language tld's.
Although a bit far fetched:
Search for "connecting people" and 5 out of 10 show Nokia tld's
That same advantage could become a disadvantage if a competitor finds your various subdomains dominating the results. For that reason, I'd say it is a good idea to have distinct content in each subdomain to minimize the chances of appearing to spam. I.e., setting up bigwidgets.widgetco.com is great if you can move to the top of searches for "big widgets". But if a search for "widgets" starts off with ten subdomains from widgetco.com, the celebration is likely to be short-lived.
Dominating the top ten might be a bit scary I agree. But multiple SERPs from highly relevant canonicals isn't neccesarily bad from a user point of view.
Everyone agrees duplicate content will get you in trouble no matter what the site structure. But if the SERPs are showing too many non-duplicate, highly relevant pages from the same site it seems like an algo tweak might be in order. (May the mighty G have mercy on me for suggesting such a thing:))
