11 replies to this topic

[whelanska]

I'm sure there's a thread in here that covers my following question, but I haven't been able to get the search function to pull it up for me (probably due to my not knowing exactly what string to query). Anyways, I'm curious to know how many sites within one industry one can have on a single hosting account before penalties set in from the SEs. I've got a client who has created about 15 different sites that relate in some way to his main site and are linked to his main site. If all these sites were on different servers, and they weren't cross-linked 100%, could he be ok? What if he had three sites per hosting account? Is there some generally accepted threshold where penalties kick in?

~SEO on the Brain

[yannis]

The search engines like Google tend to frown upon linking from the same IP address. Some hosts might be able to give you different IP's for the websites, so you might not have to move them to different hosts. Hostland used to have this I am not sure though if they still offer it. Unless the links offer value it is pointless to do so though. After all links from 15 websites are not difficult to get. Rather convince your Client to obtain links naturally. It may take longer but it will produce better results in the long term and without the risk of penalties.

Yannis

[JohnMu]

For what it's worth, I have seen over 700'000 sites on a single IP. It's no problem if the sites are small enough and do not generate too much traffic .

However, for SEO purposes, interlinking a small number of sites will be usually stand out anyway (especially if they are site-wide links). I seriously doubt if it would result in a penalty, most likely it would just be ignored or devaluated. Using different servers (or just IP addresses) does make it "harder" to track ownership, but it does not make it impossible: there are very many signs available to automated tools.

As Yannis said, 15 sites is most probably also too small a number to worry about -- I doubt the search engines would really care much about that.

However, if you want to be certain that there are no ill effects, it would make sense to put them all on unique IP addresses (which is always a good idea and not very expensive). I believe Yahoo also looks at the IP addresses when determining link value; for that it would make sense to see that the sites are on separate C-Class blocks (a c-class block includes all the IP addresses where the first 3 blocks are the same).

There's a good thread on this subject somewhere here at cre8, if only I could find it

John

[EGOL]

This is based upon observations of just a small number of sites. However, I am thinking that if you have two powerful sites competing for the same KWs and under the same ownership that google might show only one of them at a time. The sites I see in my SERPs that fit this will alternate in and out of the SERPs for the same KW one in and one out, occasionally both showing.

Although I am only watching a small number of sites this observation sticks for hundreds of keywords. So, I am starting to believe that google only wants one site per owner in the top ten for any KW.

I do see exceptions to this with organizations such as MSN, National Geographic and others holding 8 out of the top ten positions for specific queries with multiple sites and subdomains. In these situations multiple sites per owner might be allowed because of enormous trust that these sites have accumulated through fantastic inbound links.

I'd like to hear if anyone else sees anything like this.

[yannis]

This is based upon observations of just a small number of sites. However, I am thinking that if you have two powerful sites competing for the same KWs and under the same ownership that google might show only one of them at a time. The sites I see in my SERPs that fit this will alternate in and out of the SERPs for the same KW one in and one out, occasionally both showing.

Interesting observations Egol. Here is a thought experiment:

How can Google determine ownership?

1. Through Whois information, since they are also registrars not too difficult to check. however not accurate if you use a proxy name for registration. Therefore highly unlikely they use it. Perhaps they only use whois to determine the age of the domain and that's it.

2. Ip address. If multiple domains are hosted on the same IP address the probability of the sites being owed by the same entity is very high. Any false positives just tough luck.

3. Web Graph maths - Assume 15 websites and linked them together on a graph (assume also no other inbound links). The graph will look a little bit like a star of david, it will have a definite pattern. This 'artificialness' can be determined quite easily mathematically. Links obtained naturally will not have such a pattern as links will follow some type of random distribution curve. This is the most likely method of how Google determines link quality. Item (2) above can be used first without much computing power and I expect a type (3) filter if a website passes the type (2) test.

Coming back to the original question. If one wants to trick the engines a bit. Register all 15 websites using a proxy name. Position them on different C block IP addresses. Link them in a highly unlikely manner. Ensure that you have a lot of different incoming links for each website (outside the network)!

Not worth the trouble and the cost really!

Yannis

Edited by yannis, 07 August 2006 - 11:48 AM.

[Ron Carnell]

I am thinking that if you have two powerful sites competing for the same KWs and under the same ownership that google might show only one of them at a time.

I agree, Egol, and believe it's been that way for at least two years. However, to fine-tune that thought, I don't think it's so much the same keywords as it is the same "solution" (theme might be a more recognized word, but is still too broad). Naturally, keywords and theme often converge, but sometimes they don't, so it's a useful distinction (and may account for some of the exceptions you've seen?).

Any search engine, after all, wants to offer its searchers variety, not the same answer over and over again. I'm not sure ownership is even a factor, so much as a by-product. If the algorithm sees two sites that are very similarly themed and you can always get to one from the other, I suspect Google only returns the stronger of the two for any given SERP.

As for the original question, I don't think it matters one little bit how many sites are sitting on the same server or same IP address. The endemic advice to "hide" the origins of your sites is, in my opinion, utter nonsense in that (1) it's not necessary, and (2) it won't work any way.

Here's an analogy I've used before.

If you're going to the bank to cash a check, there's nothing wrong with driving your own car. If you're going to the bank to rob it, on the other hand, driving your own car is going to make it ridiculously easy for the police to find you. However, even if you steal a car before robbing the bank, the odds of you actually getting away with the crime aren't exactly favorable. There are just too many ways to get caught and eliminating the most obvious, be it your car or IP addresses, is of minimal help.

The answer? Don't rob the bank and the convenience of using your own car will never become an issue.

[EGOL]

That bank robbing analogy is a good one, Ron and thanks for letting me know that you think that if a single owner has two sites on the same SERP that one might be flitered. I was wondering if my mind was tricking me.

The sites that I am watching are in the same SERPs but one is a retail site the other is an info site. Although they share the same KW their content is entirely different. And, only one link connects them.

I am thinking that merging them with a 301 might yield more overall traffic - but that gives up landing info searchers and retail searchers on different pages.

Edited by EGOL, 07 August 2006 - 05:22 PM.

[whelanska]

Wow, this is some great information. Thanks.

[EGOL]

How can Google determine ownership?

They might also watch Adwords accountes, Adsense accounts, API queries, linkage patterns, hosting service, toolbar data. Now, none of these are proof positive but sometimes they play in concert.

[yannis]

They might also watch Adwords accounts, Adsense accounts, API queries, linkage patterns, hosting service, toolbar data. Now, none of these are proof positive but sometimes they play in concert.

Sure most of the above can be added quite easily on a whois like database.

Yannis

[Ron Carnell]

I am thinking that merging them with a 301 might yield more overall traffic - but that gives up landing info searchers and retail searchers on different pages.

It also gives up having a second basket into which you can put a few eggs.

It's nice, I think, to know that if site A loses its spot, for any one of a dozen possible reasons, site B already has the strength to slide right into its place now that it's no longer filtered by Site A being there.

[EGOL]

It also gives up having a second basket into which you can put a few eggs. smile.gif

That's a REALLY important point, Ron. Thanks!

The site that I would 301 to has never had any problems with tanking in google and has lots of .edu and .gov links so the rankings are rock solid. However, one never knows what might happen in the future.

I sometimes think that I want to give up all of my other sites just to work on my favorite. There's certainly enough work to do on that site.