Google Pagerank and "Class-C Addresses"

Hello Nanog, I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets. The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy. I googled a bit and found pages after pages of FUD and such great things as the "Class C Checker": "This free Class C Checker tool allows you to check if some sites are hosted on the same Class C IP Range." My question is: Is there any proof that Google does differentiate between /24s, or even better is there any proof that this isn't the case? I will not give a customer space from different address blocks just because he read it in a SEO magazine. Perhaps someone from Google itself can answer this question? Also how do you handle such requests? I expect I'm not the only one who gets them. Regards, Sebastian -- New GPG Key: 0x93A0B9CE (F4F6 B1A3 866B 26E9 450A 9D82 58A2 D94A 93A0 B9CE) Old GPG Key-ID: 0x76B79F20 (0x1B6034F476B79F20) 'Are you Death?' ... IT'S THE SCYTHE, ISN'T IT? PEOPLE ALWAYS NOTICE THE SCYTHE. -- Terry Pratchett, The Fifth Elephant

Matt Cutts, who works for Google, has iterated over and over that this is absolutely, positively, not the case and does not help page rank: http://www.mattcutts.com/blog/myth-busting-virtual-hosts-vs-dedicated-ip-add... So there's an answer from someone at Google. How do I answer such requests? I tell them they if they want to waste their money, that's what this service is for: http://www.seohosting.com/ Regards, Mike Gazzerro www.ubiquityservers.com 866-438-8247 x 500 -----Original Message----- From: Sebastian Wiesinger [mailto:nanog@ml.karotte.org] Sent: Monday, September 21, 2009 11:18 AM To: NANOG Subject: Google Pagerank and "Class-C Addresses" Hello Nanog, I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets. The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy. I googled a bit and found pages after pages of FUD and such great things as the "Class C Checker": "This free Class C Checker tool allows you to check if some sites are hosted on the same Class C IP Range." My question is: Is there any proof that Google does differentiate between /24s, or even better is there any proof that this isn't the case? I will not give a customer space from different address blocks just because he read it in a SEO magazine. Perhaps someone from Google itself can answer this question? Also how do you handle such requests? I expect I'm not the only one who gets them. Regards, Sebastian -- New GPG Key: 0x93A0B9CE (F4F6 B1A3 866B 26E9 450A 9D82 58A2 D94A 93A0 B9CE) Old GPG Key-ID: 0x76B79F20 (0x1B6034F476B79F20) 'Are you Death?' ... IT'S THE SCYTHE, ISN'T IT? PEOPLE ALWAYS NOTICE THE SCYTHE. -- Terry Pratchett, The Fifth Elephant

Got to stop using classful addressing terminology... It's only been 16 or so years and you're not referring to: 192.0.0.0/5 Snake-oil salesmen abound in this space. More to the point, any technique used to sculpt pank-rank scores on a systematic basis is likely to result in a countervailing adjustment by search engine operators. http://en.wikipedia.org/wiki/Search_engine_optimization http://en.wikipedia.org/wiki/PageRank Sebastian Wiesinger wrote:
Hello Nanog,
I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets.
The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy.
I googled a bit and found pages after pages of FUD and such great things as the "Class C Checker": "This free Class C Checker tool allows you to check if some sites are hosted on the same Class C IP Range."
My question is: Is there any proof that Google does differentiate between /24s, or even better is there any proof that this isn't the case? I will not give a customer space from different address blocks just because he read it in a SEO magazine.
Perhaps someone from Google itself can answer this question?
Also how do you handle such requests? I expect I'm not the only one who gets them.
Regards,
Sebastian

On Mon, 2009-09-21 at 18:18 +0200, Sebastian Wiesinger wrote:
Hello Nanog,
I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets.
The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy.
They are wrong. Unfortunately, this is a rumour that is being cashed in greatly by companies like GotWebHost.com, which offer "SEO hosting". They may honestly believe that this is true, it is not. Infact, IPs have nothing to do at all, with PageRank, and don't let any of these SEO crackheads tell you otherwise. A google employee blogged about this topic at: http://www.mattcutts.com/blog/myth-busting-virtual-hosts-vs-dedicated-ip-add...
I googled a bit and found pages after pages of FUD and such great things as the "Class C Checker": "This free Class C Checker tool allows you to check if some sites are hosted on the same Class C IP Range."
My question is: Is there any proof that Google does differentiate between /24s, or even better is there any proof that this isn't the case? I will not give a customer space from different address blocks just because he read it in a SEO magazine.
As said above: No, it is not true. Further, SEO is mostly a load of bullshit that only delivers temporary results, as the search engines will change their algorithms, etcetera.
Perhaps someone from Google itself can answer this question?
Also how do you handle such requests? I expect I'm not the only one who gets them.
It depends on how much money they pay me. If they pay me a lot of money, then I will likely give them what they want. If not, well, that's too bad for them. It doesn't matter to me, regardless, provided that they aren't violating my AUP by you know, spamming or something along those lines. In those cases, well, they probably wouldn't be asking for more IPs, because they would be offline. William -- William Pitcock SystemInPlace - Simple Hosting Solutions 1-866-519-6149 http://www.systeminplace.net/ Follow us on Twitter: http://www.twitter.com/systeminplace

The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy.
Just in case anyone cares, from personal experience, I can see that Google's priority is indeed 'rank by content'. Everything else is fluff. I've chosen a key phrase or two, and incorporated them multiple times into a blog entry. Looking at Google a couple of days later for those key words, and I can get a top three ranking quite easily. Ray -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

On Mon, Sep 21, 2009 at 11:12 AM, Ray Burkholder <ray@oneunified.net> wrote:
Just in case anyone cares, from personal experience, I can see that Google's priority is indeed 'rank by content'. Everything else is fluff.
This is not true. It's been well documented that PageRank uses a number of metrics, probably the most important of them (in terms of ranking) being the number of links to a page or site (and I believe, the PageRank of the pages/websites those links come from). One of my websites has consistently been in the top 10 or at worst top 20 results when searching for the word "megapixel" despite the word only appearing on the resulting page about 4 times - if it was simply content based there's no way that site would be ranked so highly. Scott.

On Sep 21, 2009, at 2:01 PM, William Pitcock wrote:
On Mon, 2009-09-21 at 18:18 +0200, Sebastian Wiesinger wrote:
Hello Nanog,
I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets.
The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy.
They are wrong. Unfortunately, this is a rumour that is being cashed in greatly by companies like GotWebHost.com, which offer "SEO hosting". They may honestly believe that this is true, it is not. Infact, IPs have nothing to do at all, with PageRank, and don't let any of these SEO crackheads tell you otherwise.
A google employee blogged about this topic at: http://www.mattcutts.com/blog/myth-busting-virtual-hosts-vs-dedicated-ip-add...
Yes, and I'll second this -- PageRank does not in any way get improved by hosting on multiple IPs (or different ranges or Class A's or Class- C's[0] or swamp space or space from different RIRs or "premium addresses" (?!) or anything like that...).
I googled a bit and found pages after pages of FUD and such great things as the "Class C Checker": "This free Class C Checker tool allows you to check if some sites are hosted on the same Class C IP Range."
My question is: Is there any proof that Google does differentiate between /24s, or even better is there any proof that this isn't the case?
There's Matt's word and Craig Silverstein's word and (not that it count for as much) my word -- PageRank does NOT differentiate between / 24's. Google has stated this multiple times and we have nothing to gain by lying or making things up -- the SEO folks on the other hand have a large incentive to claim that IPs *do* make a difference as they sell this as a service... W [0]: Yes, yes, I know, settle down....
I will not give a customer space from different address blocks just because he read it in a SEO magazine.
As said above: No, it is not true. Further, SEO is mostly a load of bullshit that only delivers temporary results, as the search engines will change their algorithms, etcetera.
Perhaps someone from Google itself can answer this question?
Also how do you handle such requests? I expect I'm not the only one who gets them.
It depends on how much money they pay me.
If they pay me a lot of money, then I will likely give them what they want. If not, well, that's too bad for them.
It doesn't matter to me, regardless, provided that they aren't violating my AUP by you know, spamming or something along those lines. In those cases, well, they probably wouldn't be asking for more IPs, because they would be offline.
William -- William Pitcock SystemInPlace - Simple Hosting Solutions 1-866-519-6149 http://www.systeminplace.net/ Follow us on Twitter: http://www.twitter.com/systeminplace
-- No man is an island, But if you take a bunch of dead guys and tie them together, they make a pretty good raft. --Anon.

Sebastian Wiesinger wrote:
Hello Nanog,
I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets.
The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy.
I've found that a lot of spammers enjoy having diverse ip's from which to mail/proxy requests. This may just be a case of ignorance/rumors on your customers part, but I might suspect some of them of being spammers... Leslie

On a similar issue, I have a debate going on in my company about SEO and links coming from IP blocks allocated from different upstream providers will improve page ranks. (So, if I have block A from provider 1 and block B from provider 2, web sites linking each other on block A & B, the rank will go up) Not just different /24, /24s reassigned from different upstream. I can't find anything to prove or dis-prove this theory. Anyone have a link or info on this issue/myth? I shared this discussion thread and was told it's only discussing different /24, not /24 allocated to different providers. As far as I am concerned, if Google used ARIN swip record or routing entry, it's going to identify us as the end provider so I can't see how who gave us the IP would matter. On Mon, Sep 21, 2009 at 12:18 PM, Sebastian Wiesinger <nanog@ml.karotte.org> wrote:
Hello Nanog,
I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets.
The apparent reason for this is that Google will rank links from different /24 higher then links from the same /24. So it's a SEO thingy.
I googled a bit and found pages after pages of FUD and such great things as the "Class C Checker": "This free Class C Checker tool allows you to check if some sites are hosted on the same Class C IP Range."
My question is: Is there any proof that Google does differentiate between /24s, or even better is there any proof that this isn't the case? I will not give a customer space from different address blocks just because he read it in a SEO magazine.
Perhaps someone from Google itself can answer this question?
Also how do you handle such requests? I expect I'm not the only one who gets them.
Regards,
Sebastian
-- New GPG Key: 0x93A0B9CE (F4F6 B1A3 866B 26E9 450A 9D82 58A2 D94A 93A0 B9CE) Old GPG Key-ID: 0x76B79F20 (0x1B6034F476B79F20) 'Are you Death?' ... IT'S THE SCYTHE, ISN'T IT? PEOPLE ALWAYS NOTICE THE SCYTHE. -- Terry Pratchett, The Fifth Elephant

On a similar issue, I have a debate going on in my company about SEO and links coming from IP blocks allocated from different upstream providers will improve page ranks. (So, if I have block A from provider 1 and block B from provider 2, web sites linking each other on block A & B, the rank will go up) Not just different /24, /24s reassigned from different upstream.
I can't find anything to prove or dis-prove this theory. Anyone have a link or info on this issue/myth?
I shared this discussion thread and was told it's only discussing different /24, not /24 allocated to different providers.
As far as I am concerned, if Google used ARIN swip record or routing entry, it's going to identify us as the end provider so I can't see how who gave us the IP would matter.
If Google's even vaguely smart, they'll know to use a variety of ways to automatically determine the closeness of IP addresses, including if they're announced by the same ASN, have the same RDNS suffix, have any commonalities in SWIP data, etc. If I were Google and I were engaged in mere link-counting, I would take into consideration the statistical figures and note how often a URL is referenced from the Internet in general. Then, I would look at how often a URL is referenced from "nearby" URL's, exempting URL's that are obviously components of the organization's web site, and then I would have some idea of whether or not someone was trying to game the system. This is relatively trivial to do, given the sort of data Google has on the web. I am not sure I would want to be on a list of sites-that-have-tried-to- game-Google. Who knows what sort of vengeful damage Google might inflict on your PageRank. :-) (Just kidding Google!) But seriously, just *how* stupid do people think Google is? They have massive resources and nutso bright people who have looked at these problems. To think that any trivially simplistic strategy that's been suggested for *years* now would have an impact on PageRank strikes me as naive. ... JG -- Joe Greco - sol.net Network Services - Milwaukee, WI - http://www.sol.net "We call it the 'one bite at the apple' rule. Give me one chance [and] then I won't contact you again." - Direct Marketing Ass'n position on e-mail spam(CNN) With 24 million small businesses in the US alone, that's way too many apples.

* Sebastian Wiesinger:
I'm looking into a weird request which more and more customers have. They want "different Class C addresses", by which they mean IPs in different /24 subnets.
It's not that weired at all. Others demand the same because it allegedly increases reliability.
My question is: Is there any proof that Google does differentiate between /24s, or even better is there any proof that this isn't the case?
Good luck. Google doesn't disclose their algorithms. There doesn't appear to be any Google statement on this matter, either. -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99

Good luck. Google doesn't disclose their algorithms. There doesn't appear to be any Google statement on this matter, either.
No, but for "honest" folks, they do provide guidelines : http://www.google.com/support/webmasters/bin/answer.py?answer=35769 One can't really fault Google for taking the attitude "do it right, then spend your money on advertising (with us). If you do it half-ass and spend your money on SEO instead, we might smite you". Cheers, Michael Holstein Cleveland State University
participants (12)
-
Florian Weimer
-
Jay Nakamura
-
Joe Greco
-
Joel Jaeggli
-
Leslie
-
Michael Holstein
-
Mike Gazzerro
-
Ray Burkholder
-
Scott Howard
-
Sebastian Wiesinger
-
Warren Kumari
-
William Pitcock