Noindex large productpages on webshop to counter Panda

[oeroek]

A Dutch webshop with 10.000 productpages is experiencing lower rankings and indexation. Problems started last october, a little while after the panda and penguin update.

One of the problems diagnosed is the lack of unique content. Many of the productpages lack a description and some are variants of eachother. (color, size, etc). So a solution could be to write unique descriptions and use rel canonical to concentrate color/size variations to one productpage.

There is however no capacity to do this on short notice. So now I'm wondering if the following is effective.

Exclude all productpages via noindex, robots.txt. IN the same way as you can do with search pages. The only pages left for indexation are homepage and 200-300 categorypages.  We then write unique content and work on the ranking of the categorypages. When this works the product pages are rewritten and slowly reincluded, category by category.

My worry is the loss of ranking for productpages. ALthoug the ranking is minimal currently. My second worry is the high amount of links on category pages that lead to produtpages that will be excluded rom google. Thirdly, I am wondering if this works at all. using noindex on 10.000 productpages consumes crawl budget and dillutes the internal link structure.

What do you think?

[RyanPurkey]

Matt Cutts speaks to part of what you're thinking about doing here: https://www.mattcutts.com/blog/pagerank-sculpting/ and it's important to note that it's not nearly as effective.  The thing I would focus more on is the DA and quality of referrals to your site. Secondly, linking out from pages is actually a positive strength indicator when done in the right way, per Cutts in the same article, "In the same way that Google trusts sites less when they link to spammy sites or bad neighborhoods, parts of our system encourage links to good sites." Perhaps your product pages could be strengthened further by this as well.

[oeroek]

@Ryan, thx for your answer. The pagerank flow is indeed one of the things I worry about when deindexing large parts of the site. Especcialy since the category pages will be full of internal links to productpages that are excluded from indexation by robots.txt or robots meta.

The problem I am trying to solve however has nothing to do with pagerank sculpting. I suspect an algorithmic drop due to thin, duplicate and syndicated content. The drop is sitewide. Assuming that the drop is due to panda I suspect the percentage of low quality pages should  be optimized. Would outlinking  and better DA really be sufficient to counter a suspected Panda problem?  Or is it required to make the 10.000 product pages of better quality, I would think the latter. Since there is no budget to do so I wonder if it is possible to drop these low quality pages from the index (but keep them in the website). Would this strenghten the remaining pages to bounce back up, assuming these remaining pages are of good quality offcourse.

Since SEO is not the only factor to be taken into account I'd rather not delete these pages from the website.

[RyanPurkey]

I see. There's a pretty thorough discussion on a very similar situation here: http://moz.com/community/q/can-i-use-nofollow-tag-on-product-page-duplicated-content.  Everett endorsed Monica's answer with, "... you might consider putting a Robots Noindex,Follow meta tag on the product pages. You'll need to rely on category pages for rankings in that case, which makes sense for a site like this."  Monica's long term solution was to also work on getting specific user-generated content on as many product pages as possible. Cheers!