| 7:31 pm on Jul 11, 2008 (gmt 0)|
Pagination can be a challenge - especially paginated "results" rather than paginated articles, which you don't need to control dynamically.
Sometimes it's better not to index result sets at all, rather than to confuse things. Sometimes it's better only to make page 1 indexable. Sometimes it's better to just let Google sort it all out however they want to. It really does depend on the exact situation.
I often prefer ensuring that there is another click path to the individual items - something outside the result set - and then using meta robots to no-index page 2 and on.
| 10:25 pm on Jul 11, 2008 (gmt 0)|
My belief is that pagination of results is THE killer for the last 2 years or so.
| 10:42 pm on Jul 11, 2008 (gmt 0)|
I am maybe about to look at a site that has this problem, and my thought is to let page 1 be indexed, and exclude page 2 from being indexed. They are mostly articles where starting at page 2 would usually make no sense to the reader anyway.
The long term fix will probably be to make most of the articles one page long, but the multiple images in the articles would make the pages way too big if that were done right now.
| 9:27 am on Jul 12, 2008 (gmt 0)|
rocco, would you say that's to do with pagination not be handled carefully (i.e. you end up with duplicate titles/metas/etc) or in *all* cases, pagination is trouble?
Pagination is hard to avoid, if you've got 200 results spread across 20 pages! I'm reluctant to lose the long-tail benefits of the paginated pages being indexed, so I'm going to see what happens if I make Titles/Metas unique. If that fails, then "no index, follow".
| 9:50 am on Jul 12, 2008 (gmt 0)|
I'd say the key is that paginated result sets cannot be the only click path to your deep product pages - not if you want them to rank, even for long tail searches. In addition to the technical challenge of getting those unique titles and descriptions, there's no way enough PageRank is going to circulate down into that pit.
So in cases like this, I try to be creative and consider how else a user might be served in navigating to the individual product pages. Database result sets alone are rather lame. They may be easy to generate, but they're not very effective beyond the top few pages.
| 12:04 pm on Jul 17, 2008 (gmt 0)|
could you be please more specific about the navigation you propose?
I agree that result sets are rather lame (although Google SERPs are also a kind of result sets) but sometimes such navigation schema is the only one possible.
| 10:48 pm on Jul 17, 2008 (gmt 0)|
The need is to have alternate click paths, beyond offering just one set of paginated results . The key to creating that is to forget about databases autogenerating the entire website, and in fact forget all about technology for the moment. Just think like a visitor. How many ways can you help them find what they might want?
Is there more than one way to slice and dice your content - can you create different "categories" or different facets? For instance, let's talk about a book site within a professional specialty. Books can be organized by title. They can also be organized by author, or even by ISBN number. They can be organized by printing date, by sub-topic within the specialty, by the best sellers. Reviews and articles can be written to drive traffic to deep but valuable content. You might create a search on the Tables of Contents.
On and on it goes. One category or department system on its own always runs into the pagination barrier, so you need to make use of other ways to offer drill downs into the information.
| 11:15 pm on Jul 17, 2008 (gmt 0)|
I've just had several sub-directories on a site removed from the index in WMT because the important related pages are doing just fine and it isn't near worth the time, effort and client cost for individual titles on 3rd level pages that get clicked from the "referring" pages anyway.
Couple of reasons:
1) Not worth doing custom titles, the entire site was the identical title and metas (taken from the homepage of the sister ecom site - blech)
2) The above, and they're so far down into supplemental it's no loss to nuke them and exclude them permanently - it would take months for them to get found and the dups corrected.
3) Keyword cannibalization - major issue on a lot of sites (I'm seeing and fixing it on a few sites right now)
4) For not enough PR for a site, sculpting PR to favor pages that mean something and that I would want coming *out* of supplemental status just makes sense. For some sites, rel=nofollow and removal in WMT are good friends.
I'm dealing with "paginated" pages and similar, and being ruthless about limiting what's indexed.
[edited by: Marcia at 11:16 pm (utc) on July 17, 2008]
| 11:16 pm on Jul 17, 2008 (gmt 0)|
|rocco, would you say that's to do with pagination not be handled carefully (i.e. you end up with duplicate titles/metas/etc) or in *all* cases, pagination is trouble? |
Yes, exactly. If you have a large site and cannot handle the metas individually then only allow page 1. The possible risk is too high. I have seen many sites nuked recently (from 2 years till now). These database powered sites used to be gold, but know they are turning into pain.
| 11:22 pm on Jul 17, 2008 (gmt 0)|
BTW, the site that I'm being ruthless with to eliminate the dups and PR dilution and pages indexed:
It's only a very low PR2, only a couple interior pages with PR1, and since making the initial major alterations, the site is now getting around 1K uniques a month for the last 30 day period, with very decent pageviews and time spent on the site.
No revenue, it's just an info site - but it was perfect to try out as a test case.
| 11:40 am on Jul 18, 2008 (gmt 0)|
|tedster: ... . For instance, let's talk about a book site within a professional specialty. Books can be organized by title. They can also be organized by author, or even by ISBN number. They can be organized by printing date, by sub-topic within the specialty, by the best sellers. ... |
Currently I'm using this kind of navigation. Though I'm not able to avoid paginated result sets. As you said ... books can be organized by title, by author or by topic ... - still you end up with several kinds of paginated result sets.
Google SERPs are also paginated results - why google consider paginated result sets as bad?
| 7:22 pm on Jul 18, 2008 (gmt 0)|
They're not considered "bad", it's just that the content on deeper pagination pages looks less important. That's true for Google's search results, too.
| 8:12 pm on Jul 18, 2008 (gmt 0)|
So if the results further down the line of pagination look less important what about periodically changing the sort order? Much like Google does from time to time.
| 8:52 pm on Jul 18, 2008 (gmt 0)|
I had a similar problem. I used to list 10 product per page and several resutls ran over to the second page. I have increased the results sets to 20 now and I don't have many, if any at all running to page 2 now. If it does go to page to I decided to noindex, follow. This was almost a year ago. So far, so good.
| 12:43 am on Jul 19, 2008 (gmt 0)|
How about a forum that has many hundreds of pages of pagination where it lists threads, do those deserve noindex,follow?
| 9:18 am on Jul 19, 2008 (gmt 0)|
There's an alternative route through user profile pages, recent post lists, etc.
| 9:38 am on Jul 19, 2008 (gmt 0)|
I'd like to mention that the duplication problem with regard to page titles and meta descriptions isn't the same thing as "keyword cannibalization." They're different things, done in a different way.
| 10:14 am on Jul 19, 2008 (gmt 0)|
Good point - if pages have different content but you use the same title, then that heads you toward keyword cannibalization. Different urls for the same content also gives you duplicate urls (naturally) but it's not the same issue. WMT will report it either way - it's up to you to intangle the two.
With pagination, you do have different content under potentially the same title - and the question is do you need to differentiate those titles? Do you want all those pages indexed? The can't all rank for the keywords in the title, so to really do right by those deeper pages in the pagination set, how will you handle it?
| 12:12 pm on Jul 21, 2008 (gmt 0)|
I'm not sure if "noindex, follow" in deeper pagination pages would work for me. When people are searching for some long tail (let's say for exact product name) quite often Google serves them deeper pagination pages containing the product name because the product detail page is not strong enough to show up in SERPs.
Does it really worth to set deeper pagination pages as "noindex, follow" and lose traffic from long tail searches?
| 7:34 pm on Jul 21, 2008 (gmt 0)|
|When people are searching for some long tail (let's say for exact product name) quite often Google serves them deeper pagination pages containing the product name because the product detail page is not strong enough to show up in SERPs. |
I think you've identified a mjor task to take on. Getting those product pages ranked would probably increase your conversion rate nicely. I should think optimizing those pages would payoff in the long term.
| 7:58 pm on Jul 21, 2008 (gmt 0)|
If people are already getting your deep pages for product name searches, then it sounds like you're already in pretty good shape.
| 8:21 am on Jul 22, 2008 (gmt 0)|
Though I'm getting some traffic from long tail searches to deeper pagination pages I would be more happy if Google served product detail pages for such long tail searches. As MadeWills said: "Getting those product pages ranked would probably increase your conversion rate nicely." I think that also CTR in SERPs would increase because product page titles are more appealing regarding the long tail searches.
Though I'm using pagination results sets there are about 2 or 3 different click-paths to each product page. Maybe this is not enough.
I will add another 2 click-paths and set deeper pagination pages to "noindex, follow" hoping that product detail pages will show up in SERPs more often.
| 3:07 pm on Jul 22, 2008 (gmt 0)|
|With pagination, you do have different content under potentially the same title - and the question is do you need to differentiate those titles? |
I have been fighting with pagination issues now for over a month since WMT started reporting a high number of dupe titles for pages from the same category. It would be simpler to fix if my app allowed individual titles for catgeory pages, but it doesn't. I had to change the code to append "Category X - Page X" to the titles. I'm still not sure if this is having the necessary effect though as WMT is painfully slow to update. I have seen about a 10% reduction in reported dupe content in WMT two weeks after making changes to my site. It seems that whenever Google starts finding issues like this it also slows down the indexing of your site considerably
| 3:38 pm on Jul 22, 2008 (gmt 0)|
I've gone the " - Page X" route as well. In both the Title and the Description. In the Description, I also added the first 2 or 3 words of each paginated result, so that the Descriptions are definitely NOT duplicated (although perhaps not short either!). Like you, waiting for WMT to show if I've been successful or not.
| 11:43 pm on Jul 22, 2008 (gmt 0)|
Including page numbers in the title and description helped quite a bit on a couple of my sites.
Another thing that I think helped is verifying the URL if you're usinh parameters on dynamic pages. page.php?a=1&b=2 and page.php?b=2&a=1 would render the same content but I think Google looks at them as two different pages. Only one of these should be the 'correct' format of the URL but there's a lot of lazy developers out there, myself included, that might not always use the same ordering of values after the question mark. Put some logic in the top of the page to verify the URL and redirect to the 'correct' one if the values are out of order.
| 2:20 am on Jul 23, 2008 (gmt 0)|
Differing parameter order is one classic example of Duplicate Content.
Allow only one way to directly serve content, and redirect all of the others. I often add that code in .htaccess but sometimes it goes in the script.
| 3:24 am on Jul 23, 2008 (gmt 0)|
Noindex, follow has worked pretty well on one of the sites I work on. I'm a bit concerned about flowing PageRank to our product pages, but with some creativity (rotating featured products on the home page and/or landing pages, html sitemaps, links from other sections of the site, etc) you can get your product pages to rank for the less competitive long tail queries.
| 11:58 am on Jul 23, 2008 (gmt 0)|
Pages with noindex can still accumulate PageRank.
| 9:29 pm on Jul 26, 2008 (gmt 0)|
|Including page numbers in the title and description helped quite a bit on a couple of my sites. |
I added page numbers to my category pages three weeks ago, as well as banning certain redundant URLs in robots.txt. I am finally seeing a reduction (by 50% so far) of reported dupe titles. Unfortunately traffic still keeps its steady 10 week downward trend and I see now that my PR has been reduced on most pages on my site for the second time despite gaining new links and not losing many old ones. Very frustrating indeed!
| This 39 message thread spans 2 pages: 39 (  2 ) > > |