Today while administrating an e-mail account I discovered that Google is nofollowing all of the links on GMail for Business:
Every single link: Privacy Policy, other Google Apps, the whole page. The question is why? There aren’t many reasons that someone might link to their mail login page (that lives on the Google top level domain), but the only place that the possible link value would pass is back into Google, and the examples of their applications.
I have a shiny quarter for anyone that has a plausible explanation for Google damming PageRank on individual login pages for their web-mail app. I wonder where else they are using this tactic?
7 Comments
My guess is they may have imposed a blanket nofollow across all their links because of all the criticism they have been subjected to for not always following their own rules.
At least this way they’ll ensure no one can sneak a spammy link into their CMS.
I’ve seen many bigger sites nofollowing their own links.
Additionally to this some research suggests that Google doesn’t really use nofollow as a ranking factor.
This is different than Zappos or Expedia nofollowing. Google would, in general, discourage PR sculpting; so why are they nofollowing links on a page that they control and for all visible purposes should noindex, nofollow, or both?
I am interested in seeing that data that nofollow is not considered in ranking. I have never seen anything that substantially refutes nofollow as suppressing link value.
http://mail.google.com/support/bin/answer.py?answer=6602&topic=1520&hl=en
they have noindex,nofollow in robots meta tag.
yet, their page is indexed:
http://www.google.com/search?num=100&hl=en&q=”The+more+spam+you+mark%2C+the+better+our+system+will+get+at+weeding+out+those+annoying+messages.”&btnG=Search
if you pay attention to URL’s:
http://mail.google.com/support/bin/answer.py?hl=en&answer=6602
http://mail.google.com/support/bin/answer.py?answer=6602&topic=1520&hl=en
you can see GET variables are different, and they depend how you access the page, so basically, they are preventing duplicate content, as the longer URL is noindex,nofollow
then, your example is not indexed:
http://www.google.com/search?num=100&hl=en&q=”keep+any+message+you+might+need+down+the+road”+site%3Agoogle.com&btnG=Search
is it an old page or still online? can you give URL?
Lazar,
It is a site that still exists, but I can’t disclose. However you can replicate it with Google Apps logins (e.g. Mail, Calender, etc.)
That is very interesting about the robot=”noindex,nofollow”
and rel=”nofollow” on the same content. Micheal makes a very strong point that I think your GET information supports: Google is worried about database injections damaging there search value.Still, it is disheartening that Google is doubling up on measures that are supposed to be singularly effective.huh, don’t quite understand what you mean by ‘doubling’. there is only nofollow meta tag, and there are no nofollows in links. if you are using FF addon that shows nofollows, it also detects metatags…
as for ‘database injections’… didn’t understand that either!
I must have been awake too long today. I thought I saw both robot and rel=nofollow. Thanks for correcting me.
Instituting an automatic robots=”noindex,nofollow” on the longer versions of urls could be a hedge against cross site scripting that could spoof the main site; like putting an outbound link into a help section search.
It is possible to introduce content into certain types of search systems by altering the way the search is performed. Specifically changing html forms from POST to GET can be used to insert html, and consequently, links into search queries. Example: http://www.seomoz.org/blog/xss-how-to-get-20-gov-links-in-20-minutes
Your point that GET commands are marked noindex,nofollow lend some value to Micheal’s theory that the blanket nofollow is there to discourage attempts to hijack Google pages.