There's some interest in rolling back the change we made during the 2014 Fundraising drive, which prevents Google from indexing CentralNotice banner contents. I haven't been able to find a bug tracker for this incident, but it was a big deal at the time. Google's cache of our site was slowly taken over as articles were crawled, and IIRC, at the peak there were at least several hundred thousand bogus pages. We took several approaches to fixing the problem, and I'm not certain which ones were eventually effective. The bad pages took a month to slowly clear out as they were replaced with good content.
One approach we took was to add <googleoff/googleon> tags around the CentralNotice div, on the off chance that this enterprise search technique would also apply to Google's web search. Another approach was to contact a FoaF at Google and ask them to set up a manual trim regex.
The argument for keeping the googleoff/on change is simply the high likelihood that we might cause the same glitch, and our search results will go crazy.
As best as I can represent it, the argument for reverting the change is that Google might be penalizing us for "cloaking" our pages by showing different content to the indexing bot, vs regular readers. Evidence so far is http://webmasters.stackexchange.com/questions/54735/can-you-use-googleon-and-googleoff-comments-to-prevent-googlebot-from-indexing-p (and earlier thread http://webmasters.stackexchange.com/questions/16390/preventing-robots-from-crawling-specific-part-of-a-page ).
Patch in question is https://gerrit.wikimedia.org/r/#/c/177611/ , please don't deploy until a fr-tech member is available to monitor Google results for 48 hours.