Duplication, pagination and the canonical

[.egg]

Hi all, and thank you in advance for your assistance.

We have an issue of paginated pages being seen as duplicates by pro.moz crawlers.

The paginated pages do have duplicated by content, but are not duplicates of each other. Rather they pull through a summary of the product descriptions from other landing pages on the site.

I was planing to use rel=canonical to deal with them, however I am concerned as the paginated pages are not identical to each other, but do feature their own set of duplicate content!

We have a similar issue with pages that are not paginated but feature tabs that alter the URL parameters like so:

?st=BlueWidgets

?st=RedSocks

?st=Offers

These are being seen as duplicates of the main URL, and again all feature duplicate content pulled from elsewhere in the site, but are not duplicates of each other. Would a canonical tag be suitable here?

Many Thanks

[magicrob]

Hi

It doesn't sound like rel=canonical is the solution, as each one of your pages might feature multiple pieces of content from various other parts of your website (if I've read your question correctly) - so which would be the canonical version of the page?

You could use Parameter Handling in Webmaster Tools to ensure Google knows what to do with your various parameters. Moz doesn't matter here, as long as Search Engines are aware of how to handle your pages correctly.

There's a good overview here.

I hope that's helpful

[.egg]

Thank you Robert, and for the helpful link.

You did read my question correctly, however I failed to ask it ask entirely correctly. Just to complicate matters, I neglected to mention that there is body copy on each page, which technically will be duplicated.

It sits above the tabs and does not change, while the tabbed pages - under new URL parameters - pull in a sentence or two of product description from elsewhere (a unique landing page).

So,

?st=BlueWidgets

?st=RedSocks

?st=Offers

will all feature the same body copy and different duplicate content. For obvious reasons, we only want the SE to index the main URL.

Any ideas?

Thanks again

[CleverPhD]

If I am understanding the question - I think pulling in some body copy from each search result (and not just the whole page) would be fine.  I think Google will see that this is a search result and that you are pointing to other pages.  You are probably going to pull in text from the title too.  This is common practice in search results - heck Google does it!

If you are still concerned about the pulled in descriptions, your option is to setup the system to have an alternate description for each page.  Use the alternate description when you pull it into your main page.  It is more work, but it will eliminate this issue.

Separately, paginated pages no longer need to be canonicaled to the index page.  You can use rel next and prev.

http://googlewebmastercentral.blogspot.com/2011/09/pagination-with-relnext-and-relprev.html

https://support.google.com/webmasters/answer/1663744?hl=en

It explains to Google the relationship between P1 and P2,3,4,5,n etc.

Beyond that, you need to watch that you do not get into too many paginated pages to get to the exact same product pages.  Lets say you had 1,000 widgets that were blue, red and green and also were Free, Expensive or Cheap.  You would have several sets of paginated pages (one set for Blue, one for Red, Green, Free, Cheap, Expensive, one for Red and Expensive) etc.  It gets to be a little crazy as they all lead to the same set of widget product pages.  You need to manage how to have Google crawl all that and not have your Paginated Category pages look like duplicated. Adam Audette writes great stuff on this.  Look here for things to consider

http://www.rimmkaufman.com/blog/site-search-dynamic-content-and-seo/01032013/

[.egg]

Thanks CleverPhD,

That's a very interesting read by Adam Audette too, thanks.

I should say that there's no internal search, each tab has a series of duplicated 'blurbs' taken from the product's unique landing page, while the body copy remains the same across the slight variations in the URL. So with:

example.com/example/

example.com/example/?st=BlueWidgets

example.com/example/?st=RedSocks

all of these will feature the same body copy, while the last two will have a series of small descriptions from other landing pages in the site. Would the canonical tag be appropriate in this case? We only need to index 'example.com/example'.

Also, does the rel next prev take into account duplicate content? We want only the main URL indexed as all the paginated pages feature duplicate content, there is no view all page however.

Many thanks

[CleverPhD]

The rel next prev is not for duplicated content - it just shows google how the parts relate to the whole.

An alternative to the rel next prev is the "Classic Pagination for SEO" that uses noindex another article by Adam

http://searchengineland.com/the-latest-greatest-on-seo-pagination-114284

If you have a duplicate issue, this would solve it as you would noindex all the duplicate pages.

What you need to do (and I can't do this for you), is to look at all the crawl paths that you are providing Google.  As I mention above, you are not doing any favors to Google or to your site when you show Google an infinite number of paths to get to the same content.  It just wastes Google's time and you don't want to do that when Google also has to crawl the rest of the internet.   If you solve this issue, you will solve your duplicate issue.

AJ Kohn just posted an article on the concept of crawl budget that talks about this.  I think the article is quite good and it explains why we need to look at all the topics of noindex, nofollow, robots, canonical and rel next prev  http://www.blindfiveyearold.com/crawl-optimization