Testing the domain registration myth...

"construded"?

I agree. I would have to laugh also. I would be forced to taunt them again. I would have to fart in their general direction. I would have to throw le vache over the wall. There would be no end to the vilification.

 

I think Ministrel + VS hit the nail on the head.
If G used registration time it would not make any real sense.

The only situation it could be useful is if 2 sites were exactly equal in every other factor (which is never the case anyway). Lots would just register for 10 years and it would slow growth of startups - not something any search engine wants to do.

The ownership details of the domains could prove very useful.

 

"[0099] Certain signals may be used to distinguish between
illegitimate and legitimate domains. For example, domains can
be renewed up to a period of 10 years. Valuable (legitimate)
domains are often paid for several years in advance, while
doorway (illegitimate) domains rarely are used for more than a
year. Therefore, the date when a domain expires in the future
can be used as a factor in predicting the legitimacy of a
domain and, thus, the documents associated therewith."



Patent

 

Yep. That's where the idea came from. But it is nonetheless somewhat simpleminded, don't you think? As has been said many times, a patent application is just that: It doesn't mean that Google is using/doing that thing or ever intends to.

Someone else pointed out in this thread that a porn site doesn't have to exist or to exist in the same location to make a lot of money quickly. At prices as low as $5 to $10 USD per year for domain registrations, if a fly by night webmaster thought he could make a few thousand bucks quickly, do you think $100 for a 10 year domain registration would cause that guy to blink an eye?

 

To debate about whether or not google is smart in doing something is one thing, but it appears google thinks it has some importance or they wouldn't have put it in their patent.

I personally think it is only one of many factors they might use in a filter. Do I think that you can register a domain for 10 years and automatically escape the spam filter? No. I think some of these sex sites register hundreds of doorway pages and I DO think most would not register a throw-away domain for 10 years- especially since I don't think it is an automatic in. With enough throwaway domains, you could be talking about thousands of extra dollars. I think it is very possible google uses as part of a filter, but I didn't run out and extend the registration on any domain. All my domains are on auto-renew.

I don't have any problems with my sites being flagged as spam or a problem site, but there may be some benefit to some- although I don't think it is an automatic in by any means. I also know that google has a real problem with adult sites, so it is possible this "method" may only be used on certain "high problem" sectors.

I was just posting the absract to show the actual wording and that it isn't just a theory someone invented with no basis.

 

No. This is flawed logic.

First, I'm not debating whether Google is smart -- I take that as a given, based on their history and continuing performance in the SE marketplace.

The issue of a patent application has been repeatedly confused with current or planned practice by many Google watchers. It is (1) a patent, not a claim of current or intended practice, and (2) an application which may or may not be approved. That's all. Companies patent things for all sorts of reasons, including to keep someone else from claiming the patent should they ever decide to use the process/technique or even to keep someone else from using the process/technique when they have already decided to never use it themselves.

The question is not whether they have filed a patent application or even own that patent but whether it would improve Google's ability to determine relevant search results. Using date of current domain name lapse is just too simple-minded, IMO, to be useful as a factor. It is far too easy to work around that to fool SEs. Additonally, someone recently pointed out that currently in either Russia or Romania (I forget whether it was .ru or .ro domains), when you purchase a regional domain name there you own it for life.

 

I disagree. First, no one knows exactly what google is doing at the current moment. Using your logic, there should be no discussion on anything because you can't prove it is currently being used.

The immediate reaction to the abstract on domain registration length is that it doesn't make sense because it can't possibly be perfect and there are simple ways around it. But if you try to look objectively, you should be able to realize it could play a small role in weeding out some problem sites based on statistical models. When you are dealing with millions of sites, you employ methods that are never going to be perfect out of necessity. It's all about percentages. If google has run tests and found that 80% of doorway pages are only registered for one-year, then this certainly has statistical value that can be used as part of their process.

I have a friend who works for a company called Fair Isaac. They do predictive analytics that are used to give you a score which determines the risk of loaning someone money. You might get a premium score if you have been employed for 2.2 years at the same job, while you don't if you have only worked for 2.1 years. At face value, it doesn't sound right. One month shouldn't make the difference, but that’s what you get in non-human risk management assessment.

Does it really make sense for google to say that if you have a new domain it is going to go in the sandbox and not have the same rankings as a site that is a year old? No. A new site could obviously be much more superior to an older one. But Google does this. They do it because they obviously have statistics that show it returns, in their opinion, better search results and gives them time to weed out sites.

The way around it is simple. You look ahead and create sites six months in advance. So, just like the domain registration time, google's method is flawed- but we both know for a fact that google DOES use methods that are flawed. They do it because it statistically returns better, not perfect, results- at least in their opinion - which is all that really matters.

 

Not at all. I'm not telling you it can't be discussed -- I am cautioning you against drawing conclusions based on a patent application.

And how many new legitimate sites are registered for one or two years only initially until the owner can determine whether there is a likelihood of that domain being useful or profitable? There's also the issue of non-expiring regional domains mentioned above...

I suppose it is possible that Google may include some flawed assumptions in its algorithm. However, I have my doubts that they would put much if any value in something so easily manipulated. Whatever else you can say about the Google engineers, they are not stupid.

But you are correct in saying that neither you nor I nor anyone else who doesn't work for Google knows what exactly is in that algorithm.

 

Consider this- What if google HAD being using this method of domain length of time as part of their filter, and now with the public filing of their patent abstract, it is a red herring. Now all new sites that register for 10 years are automatically flagged as a possible cloaking page because they have statistical analysis that shows 80% of ten year registrations turn out to be problem sites?

I think google will use anything that they thinks gives them an advantage or returns better search results. As you say, they aren't stupid.

 

LOL - good point!

I think devaluing 10-year registrations is at least as likely a possibility as giving them extra weight...

 

Geez...you sound like a...oh, my bad, you are Canadian.

 

Philistine.

That's a quote from Monty Python and The Holy Grail... "eek: