SEO Class in Chicago, IL
Are you a Google Analytics enthusiast?
More SEO Content
Too Many Links From The Same Server?
Posted 02 June 2004 - 12:31 PM
When I finish a site for a client, the final thing I do is add a small 'Calgary Web Design that links to my personal web design site in the footer of some pages site (usually the home page and a couple others) to increase my inbound links and PR. As I am lazy, and don't change what currently works well for me, I use the same server for all of my clients when they request hosting as the hosting company I use is super fast and solid. Because of this, I may have 400+ inbound links from different domains, but from the same server. For example:
40 links from site X on 207.34.112.001
25 links from site Y on 207.34.112.002
100 links from site Z on 207.34.112.003
So my question is, does link weight get 'discounted' by the search engines if they are all coming from the same server?
Posted 03 June 2004 - 06:16 AM
Come on guys give us some help on this one....... we don't know everything
Posted 03 June 2004 - 06:30 AM
Taken From A WPG2 PAge Critic Report
"Purchasing and setting up multiple domains can be a great way to increase your visibility on the Web and if you cross-link your domains, you can improve your link popularity for each site. This can also enhance your "theme" rating. However, it's best if you host the cross linked domains at separate Web hosts so that the first 3 blocks of the site's IP address (xxx.xxx.xxx) are different. Some engines may ignore link popularity if the IP addresses of the sites are too close together numerically. Warning: Make sure each new domain has unique content to avoid spamming!
The bit about the last octet also seems reasonable as hosts tend to get IP's assigned in blocks so servers from the same host are more likely to differ only in the last octet.
Posted 03 June 2004 - 06:46 AM
I've just had a quick look around on this subject and it does seem that the idea of linking from the same server does do more harm than good.
It seems if you wish to link to other sites back and forth then different sites, different servers and different content is the way to go ........
meaning what Geordie said is spot on
Posted 03 June 2004 - 07:04 AM
Even between different domains and servers excessive crosslinking can be penalised. As always search engines never say what is excessive....
Posted 03 June 2004 - 07:38 AM
Posted 03 June 2004 - 10:21 AM
Posted 03 June 2004 - 10:57 AM
That clearly is out of the realm of chance therefore the link text shouldn't be exactly the same. Variations including the same KW is likely a better workaround and less of a maintenance headache than different hosts. Most importantly this provides more relevant terms you can be found for. I understand in your case that term would be the biggy but as Mr. Cutts pointed out that is easily identified and discounted. IMO, if the host is local it makes sense that a lot of sites on that server could be pointing at another local business they are recommending.
Posted 04 June 2004 - 12:38 PM
The IP issue is also tricky -- I don't think they check IPs, I think they have a very clever way of discerning fake and real links.
We have hundreds of design sites which we also host and all the (identical) links show up and appear to be transferring PR. Some 'fake' links we put up are disallowed while others don't seem to be and I can't see the difference, except if the fake links are too blatant they do catch them.
I have heard that Revenue Canada has a mathematical model which can tell if you are making up numbers on your income tax because 'real' amounts have certain mathematical properties. I suspect Google has something similar which analyzes linking structure.
This makes sense for their culture as well -- coming from Stanford and using a topographical analysis of linking structure as the basis for their algorithm.
my two cents worth anyway...
Posted 04 June 2004 - 01:50 PM
Posted 04 June 2004 - 09:29 PM
There are times when I also think about this "multiple sites hosted on same server" issue. But I will take reliability over PR anytime
Just me personal opinion though.
BTW, must be only lucky but never experienced a PR drop. In fact, some pages increased their PR.
Posted 04 June 2004 - 11:04 PM
Posted 05 June 2004 - 02:56 AM
Spot on Skarraman, and it is for this reason that we constantly say that you are not penalised, your just not benefiting. Google understands that it is good business practice for you to link back to your own site from a client site, it fully understands this. It also understands, that a 'vote' from a site you control is not really a vote, as it is a vote for your , by you, so it devalues that vote. No penalty, just no or very little value.
We pick up quite a bit of work coming from links in the footer of sites that we develop. The links are definitely not there as part of any link building campaign, it's just plain and simply a good marketing angle , the fact that it does help with Link popularity is an added benefit .
Castlebooks, again you are on the right lines IMO. We spoke about this a wile back during the raging links debate. Google own applied semantics, and, with the acquisition acquired an incredible tool for the understanding of words and their relationships. I said at the time, that if I had that technology, I would be running probability scores for the appearance of words on a page, and the likelihood of those words appearing within that group or word count. ie. 2,000 words on a page, what is the likelihood of this 4 word phrase appearing 6 times, when broken down it appear 8 time separately etc.
I am sure that G has the ability to know instantly if text is being abused, be it in link or on page. Eventually it probably will do.
Posted 05 June 2004 - 01:43 PM
It also understands, that a 'vote' from a site you control is not really a vote, as it is a vote for your , by you, so it devalues that vote. No penalty, just no or very little value.
So then, probably the same value as an internal site link??
Posted 05 June 2004 - 02:15 PM
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users