Decrease in indexed URLs from Sitemap

[ibeau]

Yesterday I noticed that my alexa rank fell about 18% so I checked a few longtail keywords (which I know I rank on page 1 for) and noticed that I'm not showing in serps for them anymore. I then had a look in WMT and noticed that indexed URLs from my sitemap dropped from over 700 down to 546 (with 1248 submitted).

I then checked duplicate meta descriptions and found a description with over 602 duplicates. These duplicates were a particular category with all the different parameters. ie category-name/?cat=138&dir=asc&limit=10&order=position&p=1&price=1%2C100.

What effect & crawl setting should I be using for the "cat" parameter and could these duplicates have caused my sudden drop?

[phranque]

you might want to read through this thread which discusses the challenges of "faceted navigation":
http://www.webmasterworld.com/google/4097582.htm [webmasterworld.com]

and Faceted Navigation Problems and Noindex:
http://www.webmasterworld.com/google/4263118.htm [webmasterworld.com]

[ibeau]

I've had a read through them both but still unsure how to handle certain parameters. For example, the following all go to the same category page. Should I setup the ?cat= parameter as a Representative URL or just stop any ?cat= content from being indexed?

essentials?cat=133
essentials?cat=134
essentials?cat=135

[phranque]

first you should determine what the canonical url should be for your content.

then prevent any non-canonical urls from being indexed.
how you do that depends on a lot of things, but your choices include 301 redirecting non-canonical requests to the canonical url, noindexing non-canonical urls, using the link rel canonical element, or ignoring non-canonical parameters by specifying these in a webmaster console.

so if http://example.com/essentials?cat=133 is the same as http://example.com/essentials?cat=134 is it also the same as http://example.com/essentials - i.e. is the entire query string non-canonical?

[ibeau]

For some reason I had 2 link rel="canonical" on each page. One had a trailing slash and the other didn't. Not sure how long it's been like that but I'm sure it's from a dodgy extension.

Question.. Do you think its a problem if canonical url's have a trailing slash but the urls in the sitemap don't?

[lucy24]

Yes, but not necessarily a huge one. If the URLs belong to real, physical directories, then the slashless version will redirect to where it belongs. Google will be slightly annoyed, but could be worse.

If on the other hand both versions are getting rewritten to serve the same content, then you will have Duplicate Content all over the place.

Calling two different things "canonical" does kinda, ahem, defeat the point of the "canonical" label ;)

[g1smd]

The canonical URL for a folder ends with a trailing slash.

The canonical URL for a page does not end with a trailing slash.

Notwithstanding the fact that a folder request without trailing slash should redirect to add a slash, when there are two URLs for the same piece of content, the canonical URL is usually the shorter one.