General Tech Talk  - 
Anyone notice some wild flux in the number of "not selected" or indexed pages reported in Google Webmaster Tools on 1/6/2013? I've noticed this on several unrelated sites. More info here:
Tim Capper's profile photoCory Howell's profile photoJohn Mueller's profile photoTrond Lyngbø's profile photo
Just checked our graph and it starts looking like a Richter Scale starting on December 9. I previously chocked it up to issues on our end.
Not your dates, but i did have some wild jumps in Nov 2012. Stable since then
Hi Jana, with regard to not selected > mainly redirects and canonicalization ;)
301 redirects make up the bulk of the "not selected" URLs for most sites, but no redirects were added or removed to the sites in question on or around 1/6, so that doesn't explain the wild fluctuation.
Regardless of when the redirects were created, and depending upon the size and popularity of the site they can take quite a while to be crawled and processed anyway.

I suggest what you (and many others) are observing is simply changes in how they are reported in WMT. It's not something I would be overly concerned about, I would instead be far more concerned about the amount and stability of the reported indexed pages.
I think a change in reporting is the only reasonable explanation due to how different all the sites are, and the fact that the data changed all on the same day.

Interestingly enough, a few sites I monitor also saw increases/decreases in the number of indexed pages also on 1/6 (although not as dramatic as the "not selected" pages).
Bear in mind that although WMT is fantastic for such information, from the horses mouth as it were, its always slightly out of date with some of the data that it provides ;)
I would - unless you're looking for traces of a site-move being recognized - perhaps go so far as to say that you may want to ignore the "not selected" counts. They reflect some processes which don't always map to specific changes on your side.

For example, if you have set up rel=canonicals or canonical redirects for URLs with specific query parameters, and our algorithms decide to check that we're not missing anything, then that can lead to a rise in "not selected." When we clean up our data internally and consider re-checking older URLs that used to be "not selected," then that can lead to a drop in these counts. Provided your site is configured well (and I'll assume that all of yours are), this isn't something you'd need to worry about -- we'll crawl, and then index the good pages, and "not select" the URLs that are redirecting, etc.
I saw this today with one if my clients, the number of not selected was reduced by over 50% for no apparent reason. I'm glad to see this happening elsewhere as a quick google search threw up no other examples.
FWIW "not selected" is now removed (it wasn't providing much value & confusing too many people), so you won't be seeing fluctuations there anymore :). 
+John Mueller I didn't see anything in your comments about the shift of "total indexed" that occurred on 1/6. **Edit: I checked and about 30% of our client portfolio experienced the shift.

For the domain I'm seeing the shift on, I believe it's a great thing that it dropped, but it would be helpful to know if it's just a reporting change (e.g. how canonicals are evaluated, etc.) or if something changed on the site. I'm leaning towards the former since that date seems to be a common trend, but definitely appreciate any feedback you may be able to provide.
I don't see a URL here, but I might have missed it. Feel free to start a thread in the help forum (with the site's URL) if you want me to take a look! 
Add a comment...