Length of Domain Registration

[Larry]

I just read that Google is (may be) considering the length of time a domain has been registered in their ranking (SitePoint Tribune #325):

A recent patent by Google indicates that the search giant now considers the duration of your domain name registration in calculating your site's ranking. Google does this because many spam sites have short registrations: a longer registration indicates to the search engine that you're building a site.

I maintain domain name registration for many of my clients as a service along with their hosting. Obviously, I'm not going to spring for a multi-year registration without some long-term commitment from the client. Does anyone (Ross?) have any idea how significant a factor this is? Also, wouldn't the length of time the domain has already been registered also be a factor?

Thanks,

[Annette]

That seems rather ridiculous; in our experience, people lean toward domain registrations of a period of one year far more than any other length of time.

[Andy Z]

I believe they actually look at the length of time it has been registered, not just if you registered it for one year (i.e. they look at original registration date vs. today's date). At least that is how I interpreted it. :)

[IamBen]

to add to Annette's reply:

A lot of people renew on a yearly basis simply because the hosting industry is based on frequent renewals (monthly, quarterly, etc). I myself have never purchased or renewed a domain for more than one year at a time. That's mainly because I let HM deal with all of that as part of their billing cycles. But it still applies because a lot of people do that.

On the other hand, I've noticed a bunch of those annoying "search" spam sites sit on domains for years at a time.

So it seems that this would be totally ridiculous because it would have the opposite effect...

my 2 cents.

[rmjvol]

It's a piece of straw on the camel's back now that the patent's public. Was probably a good stand-alone quality indicator before, not so much any more.

It sounds like you're thinking about this in too simplistic terms. Consider that this could be a factor that only comes into play if a domain has displayed over-aggressive link acquisition of >100 inbound links from documents that are < X months old and are contained within the same C class. Consider there could be another 50 variables involved. Who knows?

BTW, what would be ridiculous would be to assume that Google didn't consider a freaking HUGE amount of user/domain data before adding some factor into their algo. I hear they've got some fairly bright people (http://www.vnunet.com/vnunet/news/2142071/google-recruits-farther) working over at the Googleplex.

On the other hand, just because they may consider a couple petabytes of data doesn't mean they always make good decisons.

Ross