is it me, or is G going crazy slapping that damn duplicate content filter these days? funny thing is, these pages aren't even duplicate ... or even close to duplicate. it's liek once that filter starts up, it doesn't stop until you have nothing left. what's the deal? i have several sites that are getting hammered ... google hates me
they are on a mission to kill all my sites ... granted ... mostly AWS ... but i used to be able to revive them, i'm getting my a** handed to me this week. updates are slow, not helping matters.
I have AWS sites, quite a few of them, with stable page counts in the 40 and 50ks. I don't even think a dupe content filter exists except if the source codes are exactly the same, or like 90%. With the right amount of PR pointing in I was able to index a dmoz clone so it now has 165k pages indexed. I would say your problem is you don't have enough PR pointing in. Buy a few PR5 links and point them into the main category pages and they should return back into Google's index.
I've seen big page drops on my sites as well. Google was showing far too many pages to begin with. Using the site: command, Google was saying my site had around 20,000 pages, when I know it only has around 10,000. It's now showing around 9000 pages with the site: command. The API shows a very different story however. I appear to have dropped from about 3800 pages down to 10. Not too happy over that. While I may have some pages that are very similiar, there's no way I have that many similiar pages.
It's nothing to do with similarity, if you ask me, but mostly about not enough inbounds to the category pages. Google needs to be convinced that the site is worth keeping in the index. Pointing a bit of co-op, say... 1,000, into the inner pages will be helpful. It's like a vicious circle.. you either point the weight to it, or soon enough you have no weight to point at all!
i have 2 sites which have significant coop weight and recriprical links and google has kicked out all the pages on both sites except for 1-3 each. the sites were made of amazon bookstores one of the sites has like 100,000 backlinks, so I would say link pop will not save you