There is so much talk about how important it is to put your sites on separate servers at data centers far, far away from each other. Link builders who have huge networks of sites do it all the time. They have huge lists of reseller packages with dozens if not hundreds of hosting companies. The question is what do we know about how Google devalues links?
A topic over at Search Engine Watch Forums named Links from Same IP Blocks, mentions "we lack evidence" that having a network of sites on the same C IP block (i.e. 192.168.1.x) will devalue or have a negative impact on ones linking campaign. I have selected a few posts by recognized link experts to help clarify.
rcjordan writes "if I were doing, say, 3 unique content sites for a company and the content warranted each having its own domain I wouldn't hesitate putting them on the same server and c-class."
Ammon Johns writes "Google were known to aggressively penalize cross-linked domains before LocalRank was ever discussed."
GoogleGuy in msg # 7 writes, "Virtual hosting means that innocent sites and spam sites can be on the same Class C block--or even on the same IP address."
It is to my understanding, that Google will not automatically devalue the links from cross-linked sites based on an IP address or C Class. However, Google will, to my understanding, devalue links from sites where you have 20 or more links on a domain name pointing to the same page. For example; if you have a site named www.cars.com and on the footer of this 5,000 page site is a link to www.mysite.com/page1.html. Google will count up to about 20 of those links and devalue the remaining 4,980. So if you have five, ten, thirty huge content sites, all on different C blocks the links can still be devalued based on the number of them pointing to a site from the same domain name. I believe this has little to do with IP addresses but more to do with the volume of links within the same domain name.
This post is based on discussions with a wide range of SEOs and non SEOs, and from hundreds of forum posts. As I said here, this is not based on scientific data.
