Welcome to the Q&A Forum

This is happening because Google thinks the product is more relevant to the keyword than the category. In short, your products are more optimized. This most likely has to do with more external links with the keywords as anchor text going to your product pages than go to your category pages. But it could be any number of reasons related to SEO, all of them detailed somewhere or other on this website.

Run a search with Open Site Explorer and select "links from external pages" and "to all pages on the subdomain", then export the .csv. That way you have all the links the Roger Mozbot has found linking to your site with PA and DA metrics. Or use Yahoo! Sitexplorer, that gives you the opportunity to export the report too.

Thanks for the suggestions, Darren - we did discuss this as a possible reason and we plan on testing this theory out.

No problem Everett.  Thanks for the information.

What happens when you look at non-branded organic search by engine? I want to make sure that we're talking about a Google dropoff here, and not that suddenly you've been de-indexed from Yahoo and Bing and we're looking at the wrong issues. Were there any code changes to your site recently where something might have gone wrong? Did you run Adwords or some other type of paid search advertising that wasn't correctly tagged (thus showing up as organic) that ended and is now showing a huge drop?

Yep you sure can.  (Won't be a HUGE issue if you link the two together since there are only two sites) That's actually what the search engines frown upon.  Depending on your main focus (Which one you want to rank higher) you can use canonical tags. You can use the canonical tag to point the lesser important one to the more important page.  This will let the search engines know that the content has a parent site.

Hi Steven, thanks for your fast answer!

I'd just add that if the solution chosen is noindex, to do the noindex, follow method, just to give the extra cue if there are links on those pages.

Just wanted to leave a quick note saying that SEOmoz has upgraded the Link Directory! You can view a post Cyrus wrote with more information about the directory updates and link building via directories at http://www.seomoz.org/blog/seo-link-directory-best-practices

Thanks. I've used Link Assistant and it works well. I was thinking more along the lines of something that would track the progress of links I had requested rather than the status of links already acquired. Stuff like date, URL of potential link, action I took, link text used/requested, status of request, person I had requested the link from, etc. If anyone has a good resource please let me know

The answer depends on a few factors. Step back and take a fresh look at each page. Don't think of it as the page's author or webmaster. Think about it from the view of a user who searched for a phone and landed on the page. How much information is what the user was hoping to find? How much of each page is repeated keywords that have nothing to do with the page's content? How much of each page is duplicated information that will be found on every other page in your site? It comes down to percentages. I do like your pages. They are clean, easy to navigate and well-presented. The only concern I have is the sidebars with the points mentioned. I do not see how Google could have any concern with your content unless you have other duplicate pages on your site of the information. I just took another look at the links you offered. While there is some duplication, it really is not much.

Thank you for your quality comments. I think we are going to go for Evaretts suggestion since I do feel like hundreds of thousands of 301's might be a bad idea. We are also discussing the possibility of keeping the old pages (do index, do follow) up with the old ads, just removing the seller information. We believe that users might well be interested in information about past items sold or expired. This wouldn't require any 301's, just dynamically adding the message suggested by Chris and a search engine/index to help find interesting sold/expired items. This way we could generate hundreds of thousands of content pages which might over time bring plenty of quality traffic to the website. These pages would have dynamically generated fresh content (new suggestions about relevant ads) so I don't feel too worried about old content in Googles index. We could also consider scraping some details of the sold item from the manufacturer website. I've never dealt with this many pages combined with this thin content before so The wrath of the Pand worries me a bit. However the root domain is quite strong and these ads seem to draw quite a few links. The question that requires some further thought is whether or not having this many old, thin pages in Google's index will prove problematic. Thoughts? ps. Since I didn't mention this before, the website is about used cars

Your point is also valid Yousaf, however it's up to automated algorithms to determine what's valid, what's not, what's acceptable use and what's not.  That's the concern.  If you feel truly confident that your use could in no way be perceived algorithmically as being a problem, then feel free to proceed. Personally I live in the world of best practices and recommend to clients that they avoid such potential pitfalls whenever possible.  Of course, that does limit their ability to make the most out of web technology, but that is sometimes a price that has to be paid when such risks could harm their revenue stream.