|Canonical tag and pagination|
I'm revisiting the setup, from an SEO perspective, of one of my websites.
The website is gallery-centric. For example /videos/. There are around 100 pages of videos, /videos/2/, /videos/3/ etc, etc.
Each /videos/ page, has 30 links to actual videos, example /videos/funny-video/
I've noticed that on /videos/2/ - /100/ there is a rel="canonical" href="/videos/" telling Google, that pages 2 - 100, are all duplicates of page one. Which isn't strictly true, as each page of course, links to a different 30 videos.
Are the links on canonical pages followed properly? Will the videos on pages 2-100, be followed and have link juice passed to them? and ultimately, will it look odd to Google, that the website is claiming, there are almost 100 versions of the same page.
My feelings are that this was likely a mistake and I should remove the canonical links.
I didn't want for pages 2 - 100 being indexed or competing with page 1. I think it would be a better idea, to use follow, noindex. I'm assuming Google would still follow links on the page, but not index the page itself?
On the idea of using noindex, follow. If PR is passed to the page, that's noindex, will the PR flow through the page and on to the links on that page. Or will the PR dissipate and the links will only be followed for discovery purposes?
So much to consider, my head is starting to hurt :)
Yes, noindex,follow would mean that the links still flow PR - even though the page is no longer in the index. It also means that those URLs will continue to be crawled, otherwise the meta tag and the links wouldn't be seen. In contrast, a robots.txt disallow will stop those URLs from being crawled and free up that crawl budget.
I've never been comfortable with the kind of canonical link you describe, either. That's not really a canonical URL or duplicate content. I know a lot of sites are doing it, but it just doesn't seem right to me.
[edited by: tedster at 4:46 am (utc) on Feb 4, 2011]
Thanks Tedster. You've answered my question completely and reinforced my feelings enough to remove these tags.
Thanks a lot :)
|brotherhood of LAN|
|That's not really a canonical URL or duplicate content |
And Google managed to overcome the 'session ID's in URL's' infinite crawl problem quite early on, IMO they should be good at managing to notice paginated results as they're similarly prevalent across sites... without casting doubt on the intention of the content.
Very true. Page 2 etc, were showing up as duplicate content, which is what I wanted to try to eliminate at the time. But I guess it just means, the page would struggle being indexed, which isn't an issue, as I don't really want it indexed.
Though still, I don't want PR being wasted on it. So I guess the noindex shoald fix that.
Even a page that is noindex can still accumulate PR if there are internal links pointing to it. A noindex,follow metag tag will ensure that the PR gets passed through the links on that noindexed page too.
A nofollow on the link that points to the noindex page will stop PR from being accumulated on the noindex page. However the PR that is not voted over just falls into a "black hole" - it does not get redistributed to the other links on the page. It used to and people started doing PR scuplting. but Google changed the calculation methhod about two years ago.
Yeah, I figured it was best to steer clear from rel="nofollow" now. There doesn't appear to be any benefit in using it anymore. I used it for PR sculpting purposes and had decent success but for all that link juice, to suddenly fall in to a black hole, wasn't good.
I just found a section of the website, which has canonical links to itself.
For example [mydomain.com...] actually has rel="canonical" href="http://www.mydomain.com/videos/widgets/2/"
I'm not sure what the implications are with this? Will it simply be ignored? As it's clearly not been used correctly.
edit: Just done a little research, it appears this is fine. Some WP plugins such as allinoneseo appear to do this too.
I would suggest that paginated category pages alone are not a great way to pass pagerank to your videos. You would probably do much better to implement a section on each video page called "related videos" or "you might also enjoy watching". Then you wouldn't have to rely on all the pagination pages to pass pagerank and they could be canonicalized or removed.
|I just found a section of the website, which has canonical links to itself. |
This is fine if this page is a true canonical version of "itself". We have used this with no problems.
canonical links to self had never been a problem... in the cited example, it is perfectly fine to have it that way...
Canonical links to self can even provide some protection against low-level scraping.
Tedster: can you clarify something for me regarding the effects of canonical - I've been trying to fix dupe content issues on my site over the past several weeks.
Couple of concerns: 1 - in googlebase, I link to ALL these colors separately - they require differnt URL's but I can keep the URLS and just have rel canonical redirects to the main color page, can't I?
I am currently number one for the term targeted by the area, and I always suspected that it was because I had SO MANY pages and so MUCH content - but perhaps I would be better served not to dilute the content. Maybe its just a waste of time and I'd rank well anyway?
Hi guys, I have an issue since implementing the noindex follow tags on the gallery pages >1.
The days, weeks proceeding, my adsense earnings dropped quite considerably. They are now almost half of what they were. They were previously steady for coming up to a year. I'm losing literally thousands per month and can only attribute it to these changes. Only the CPC has been effected, the CTR is stable.
I cannot figure out how or why these changes could have effected adsense.
I don't display adsense on any of the pages that were noindexed.
I've been trying to theorise why this may have happened and one thought, is that the noindexed gallery pages, lead to the download page, which is where the download ad is.
Could it all be related to some kind of quality score? The referring page is now seen as a lesser quality because it's not indexed and has no PR at all? It seems a little far fetched but it's my only lead atm.
*prays tedster pops by*
I don't really have experience using this, but this is my understanding. Correct me if I'm wrong.
If you do canonical and reference the first results page, link juice will be passed to the first results page instead of distributed over your paginated results. You obviously want to pass juice to all your videos so this is pretty bad.
You could do noindex, follow but when you do that you're still giving your results pages link juice but telling the engines to never rank them. I just don't see any purpose for this.
This type of thing is really supposed to be used for duplicate content. Pagination really isn't duplicate content since each page lists different videos and it's crucial to your information architecture. If you were indexing listings that were the same except for with all different sorting options, that would be different.
Using noindex, follow. Should basically mean that only the first gallery page is indexed. IMO it should consolidate the KW phrase to that page. The follow portion of the tag, should still pass link juice, to the "videos". Which is what I want.
Theoretically there shouldn't be an issue with this method. But adsense clearly isn't happy.
(If it's your EPC and not your CTR, it could be any number of things, including smart pricing. But I don't think it's at all likely that the changes you made caused the drop - unless your site structure is so different now that Google has to "learn" it again to target properly. I have had this happen, but only when I completely re-structured a site from the ground up. You need to do some more granular testing to figure out what exactly has dropped and where)
This is a graph of adsense. It's been stable for coming up to a year. Then the day we accidentally nofollow the entire site, it dropped as you can see, due to loss of traffic. Then we fixed it and traffic returned quickly but adsense stayed in a slump.
I'm 99% certain this isn't coincidence. I've reverted the changes and I'll monitor how adsense reacts.
I did this test one year ago, and the anchor text of this links was ranked on Google results