|Massive Cross Linking Experiment|
A client of mine bought 100 domains, 90 of which are on the same class C IP, most whois info is randomized, and the sites are now cross-linking heavily. Most sites are PR 4, and, on Google, none of them rank well for a very non-competitive search term (less than 500 queries reported by Overture, few sites optimized for the KW). The sites rank fairly well on Inktomi.
I am being hired to "fix" the situation, and what I find interesting is the opportunity to have a massive linking experiment that I would otherwise not spend the money or energy doing.
I am about to do the following:
1. Move all sites to unique class C addresses (yea, it's kind of expensive, but the client is paying for it!)
2. Make sure all Whois info is unique
3. Re-architect the cross-linking
Now, #3 above is what I would like feedback on. I am operating under the assumption that Google noticed the massive cross linking on the same IP for the same KWs and is either penalizing, or simply not rewarding the sites for the links. I am also presuming (and I do believe this) that Google does not "remember" a site's penalty history (another client of mine was literally removed from the index and after I cleaned up the violations of TOS, it went right back into the pack and is performing well).
I plan on breaking the sites into 2 groups. Group A (about 35 sites) will only link to other sites in Group A, and Group B will link to sites in either group.
(Any advice on that?)
I am expecting to see the sites emerge from whatever penalties they have, and probably Group A ranking better (all other variables constant) than Group B.
How about clustering them into groups of 5 and then having them all point to one big site?
i should have mentioned the unique objective of the client is not to have 1 site at the top, but to have about the first 25 SERPs. they are trying to knock down a site a competitor made and optimized for their company name. so they want to bury that site.
I think if you control the top 25 you'll get nailed big time.
Doesn't matter what positions you control, and it doesn't matter how many different servers you use on how many different C Classes. It's the crosslinking that will sting you.
Creating multiple websites exponentially increases the link development chores. It's ambitious and probably worthwhile, but it's a LOT of work generating different backlinks to all your websites, on top of making them rank well. Nice to dream about, harder to execute.
|they are trying to knock down a site a competitor made and optimized for their company name. so they want to bury that site. |
I'm always sarcastic, so don't be offended, but:
what's the company name in question, "european-resort-hotels.com"? If someone optimizes for that term and beats you at your own name game, then you may be chasing someone's misguided wild goose.
Sometimes clients have good ideas about how thier money will best be spent optimizing a site, others don't have any good ideas.
While I think it's kinda sexy to think of myself as some hired-gun, and I do look keen in my zip-up latex cowboy suit, sometimes it's a good idea to do a reality check, like "is this hair-brained scheme actually going to benefit my client"? Sometimes I'll strongly adivse against doing something the client wants to do because of poor expected or potential ROI, and save my smoking guns for the next idiot who dares to cross my path.
|at's the company name in question, "european-resort-hotels.com"? |
No, as I mentioned, this is a completely non-competitive term. They have an antagonistic competitor. I suggested other ways, including law suits. The CEO is the most stubborn person I have met (when it comes to understanding search engines).
At this point, I am just going to do what they say and focus on the goals.
Well, what is the client's motive? You're talking about making changes to 100 domains to appease someone on a non-competitive keyterm who you admit is search engine ignorant
to me, letting someone spend bucks like that on something so trivial is to me like watching someone beat a child. Up with which I will not put.
I'm not saying you're a bad person for taking his money and allowing him to waste it on nothing more than some misguided egotistical notion, I don't even know you.
I'd love to help you with your problem but I prefer contributing to things worthwhile, on things I can be proud of
I'm not even saying you're making a mistake, but it might not hurt to look at this a little more closely, and I apologize if I caused you any offense and offer only my kindest regards
[edited by: neuron at 3:37 am (utc) on June 23, 2004]
|Well, what is the client's motive? |
As mentioned in post #7, a competitor has a site that is giving them bad PR. My client doesn't stand to make a dime off of this. Initially I encouraged them to get a lawyer involved.
To hopefully end this tangent, the client is concerned that this is a PR issue and is throwing money at it.
I think you've got it mostly covered. I'd be interested in other's people's insite as I have a similar problem I am working on.
It seems to me that the best approach is to cluster the different interlinking so that don't have 'points of failure'. Kind of like a celluar approach and just because google finds one cell, the other do not fall.
Also, you may want to try different tactics and techniques with the different cells .. take notes and see which ones fail or succeed.
Let me know how it goes. I'd be very interested in seeing how successful you are / are not.
Bad Public Relations affects revenue in direct and indirect ways.
[edited by: martinibuster at 4:36 am (utc) on June 23, 2004]
[edit reason] Edit for clarity [/edit]
Any experts care to comment on the strategy to effective crosslinking? Also, how did your client do with the 100 domains, grant?
For those interested, I thought I'd share the results to date. (I find this fascinating).
First, Martinibuster's comments were right on.
What I found is that search engines, especially Google, are fantastic at identifying link clusters.
Having access to hundreds of domains, many of which are on unique class c addresss with different whois info, you can only leverage that so many times before you are viewed as a link cluster.
To date, the results on the search engines do show all my sites, but several poorly optimized sites (virtually no backlinks, onsite factors are ok at best) remain high. My interpretation is that Google sees the many sites I have as belonging to one link cluster, and wants to serve results from numerous clusters.
I could be wrong, however, this is an ongoing experiment and I thought I'd share the recent update for those who'd like to know the results.
Will the people running the other site not notice what is going on and report you?
It's not just a matter of beating the algo.
|not notice what is going on and report you |
Our methods don't violate any terms of service. In other words, if you filled out a spam report, none of our sites would qualify for removal.
Hmmm....I wouldn't necessarily go that far. While you likely wouldn't get thrown out, cross-linking has always been a "No No" to Google and many webmasters can attest to getting nailed by G for this at one point or another.
If it is a non competitive term and the other site is SEO ignorant, it seems like a lot of effort for something that could be cleared up much more easily. Unless of course the "bad press" site has some validity to their claim and has some damn good backlinks because of it.
I'd build up 5 pyramid groups of websites and link just the home pages to the authority site. Hopefully you can get 5 PR7 links to possibly get that authority site to a PR8. With very good seo you should be able to achieve your goal.
I don't think you have a hope in doing this. Imagine if your client paid a fortune to host these near duplicate sites and Google pulled the plug on them all :o
> In other words, if you filled out a spam report, none of our sites would qualify for removal.
Maybe not removal, but another form on 'penalty'.
I'll give you a situation on my site. I own the .com and .co.uk of a particular domain. Typing in the domain name ending .co.uk into google and hovering over the 'Find pages that link to this page', Google has noticed that they should be the same site, clicking on this shows the links from the .com (according to the URL). It actually shows all links, from the .com and the .co.uk.
Google knows that these are the same site. Every search I do, one of the sites URLs ranks high, the other much, much lower.
Now, imagine that you optimise for these 100 sites, your competitor reports you (or it happens automatically), then Google merge the 100 domains into one 'site', thus having one high result - the most relevant page from one of those 100 sites with the other 99 'penalised' automatically, real time, for every search. This means you can dominate the top two spots. Where will you competitor be now? Number 3 (and possibly 4)? Will that serve as an advantage? It's not a penalty, but it is a consideration that needs careful thought.
And 100 crosslinked sites can be viewed as spam. The TOS can be changed at any time without warning and sites get dropped by violating their unwritten TOS (the algorithm!).
if you have 3 tiers of hosting on different class c ip's, link them in a a-->b-->c-->d-->a fashion not reciprocating and not linking to another site on the same class c........wouldnt it b very hard to identify that?
|At this point, I am just going to do what they say and focus on the goals. |
The goal is to bury a site optimised on the client's business name right? They need a lawyer, not an SEO.
|if you have 3 tiers of hosting on different class c ip's, link them in a a-->b-->c-->d-->a fashion not reciprocating and not linking to another site on the same class c........wouldnt it b very hard to identify that? |
Well that's the million dollar question, isn't it? =)
|wouldnt it b very hard to identify |
It would be very easy for Google to identify if you have the Toolbar installed...and if you check backlinks, as well as use the site: command.
I'm of the opinion that they can identify "link clusters" and sites that you own or work on in ways other than sites just linking to each other.
|if you have 3 tiers of hosting on different class c ip's |
Before I started this experiment, I was convinced that the same class c address would be the problem. However, as martinibuster accurately predicted, it is all about unique links -- the link work becomes exponential. Class c addresses seem have nothing to do with it, it is all link neighborhoods.
does whois info really affect it?
I would be amazed if this were the case.
Think about how much work it would be for Google to get all of the whois info, class c, and make wild guesses and indictments based on that info.
It's all about unique links, they trust links first and foremost.
Before I started this, I was concerned with whois and class c addresses -- I wasted so much time with paranoia.
I have seen what you think as bad actually work. I bunch of sites is only worthwhile if they each have there own independent non paid for old backlinks. Ditch any sites that only have links from your network. They are completely worthless. There is no problem for them to be on the same subnet. I would give them their own IP because that is really easy to do an not expensive. There is no problem having the same whois info. You are going to be spending a lot of time and effort for something that has no real value. You will not see any improvement by doing what you say. You will still need to go out and do other SEO work for them to be happy with your work unless they are hiring you specifically to do this and they don't care if it works or not. If they want it done on your recommendation don't do it. I have experience in this area and I know what I am talking about. There are tons of things that your time can be better spent to improve your ranking. Any site that has their own backlinks have them all point to one main site and get it's PR up. This will take a while. Any kind of linking strat takes a while nowdays. Also you need to tell them that being the top 25 is the best way to get knocked out completly.