To a certain extent, google data mining gmail for ranking information makes a lot of sense. I would assume this would only be for text emails and would probably have a throttle like page juice that prevents very active and/or new emails from distorting SERPs. But each email (properly filtered) with a link will have a lot of valuable contextual information not polluted by markup nor SEO gibberish of what that link is about.
Opinions on this? Anybody hear about or see any evidence that this is the case?
Wonder what the best way to take advantage of this is then. There has to be an anti-spammer/anti-newsletter filter otherwise that would be unfair.
I suspect text emails from unique emails addresses, from unique senders, whose content doesn't get flagged by gmail spam filters and with strong contextual text is key. Tricky thing is trying to get people to send out emails with your link...
Msg#: 4406543 posted 10:32 pm on Jan 13, 2012 (gmt 0)
Obvious question that I didn't see an answer to: Does the average human (in other words, not us) realize that their mail is being read?
On some level you have to be aware of it. How else would a spam checker-- either on your own computer or in the intervening mail server-- be able to make decisions based on the content of the mail? But there's some kind of conceptual chasm between mail I get and mail I send.
Msg#: 4406543 posted 12:25 am on Jan 14, 2012 (gmt 0)
The question "Does google reward links in gmail" is an interesting choice of phrasing. It suggests that rankings result only from the earning of points.
I think a lot of signals are used more in a reinforcing or confirmation manner, rather than a signal that earns ranking in the first place - and links in gmail are more likely to be this kind of thing.
For example, if some kinds of online activity suggest a certain ranking level, but the reinforcing signals are almost completely absent, then that initial ranking score becomes questionable, or weakened - not reinforced because that possible ranking is more likely to be artificially created.
So some kinds of signals may not be able to generate a ranking on their own, but their absence may undermine a potentially good ranking. So goes the theory I'm playing around with, anyway.