Strange Logic Diary
The Scientific Business Diary

Today's Quote:
-- Anon
Logic Diary Info Panel

Logic Diary Syndication
Syndicate using Atom
Syndicate using RSS

Site Map

Keyword Resources
Keywords for Internet Success

Dollar & Euro Banking
Guaranteed Multicurrency Bank Accounts

Market Reports
The internet market reports

Our archives

  • 02/29/2004 - 03/06/2004
  • 03/07/2004 - 03/13/2004
  • 03/14/2004 - 03/20/2004
  • 03/21/2004 - 03/27/2004
  • 03/28/2004 - 04/03/2004
  • 05/09/2004 - 05/15/2004
  • 05/30/2004 - 06/05/2004
  • 06/06/2004 - 06/12/2004
  • 08/08/2004 - 08/14/2004
  • 12/12/2004 - 12/18/2004


    Wednesday, March 03, 2004

    Website families and their death in the SERPs 

    "Website families and their death in the SERPs"
    Or how the Search Engines will commit genocide!

    I believe that the search engines have no option but to commit genocide but before I go on I should explain my understanding of what genocide means:

    Genocide is the murder of an ethnicity or the extinction of any group sharing a genetic or ancestral affinity

    Searchers and search egnines have a symbiotic relationship. Since time (or the web at least) began search engines have tried to deliver relevant search results to their audience of searchers, aiming to build a long term relationship so they'll keep coming back to search and search again, rather than have their searchers go and have an "affair" with another search engine.

    The searchers will occasionally click on an advert and the search engines will earn an income while the searchers get the valuable information they need and want.

    Webmasters learnt that getting high rankings in the search engine position pages meant extra, targetted traffic to their shop, store or informational resource allowing them to convert them into money according to their own business model. Many of these webmasters learnt the art of Search Engine Optimisation and over time many of the results pages became skewed with results the search engines and many searchers found irrelevant to their wants and needs.

    The search engines had to act. They NEED the searchers, whereas the searchers don't NEED the one specific search engine and will happily leave their previously monogomous search engine relationship and will dally with all manner of other search engines until they find a new relationship they are happy committing to (for the time being at least)

    Then along came Google and the roles were reversed. No longer did the search engine need the searchers as much as the searchers needed Google! The results were so much better than those they received with every other search affair they had previously been involved with. They were relevant, they were clean, they weren't bombarded with advertising and most of all all their friends and acquantances said "Have an affair with Google as G is happy to partner with you and - I won't tell if you don't!"

    The SEOers at first weren't quite sure what to make of Google but in time they worked her out and decided en masse that there is little point in SEOing for the other engines as Google has all the searchers. And then the SEO game really started :)

    Over time Google's results changed from being extremely relevant to becoming more and more affected by the SEOers. It was like the movies where "If you build it they will come" except this wasn't about baseball stadiums but inward links.

    It became clear that the more links you had pointing to your website that the higher up the SERPs you would be and people built links like links had never been built before.

    But Google wised up and started employing a new algo to diminish the importance of links for links sake.

    Hilltop came to the rescue!!

    But I can hear you saying, "What the hell does this have to do with website genocide ?"

    Read on, dear reader and I shall explain all :)

    Websites, like people have parentage, but unlike people I believe that it takes 4 to tango!

    The 4 parents of a website are:

    The Domain name
    The IP Address
    The Whois Information
    The Content

    These 4 parents come together and all leave a genetic fingerprint on a website and all of them can be analysed and traced back to form a family tree.

    I am going to get a little geeky now and apologise in advance for doing so, but hope it is in plain enough English for you to understand

    Every website has a domain associated with it and every domain name points to an I.P. address. There are 2 ways that a domain name can point to an I.P. address :

    Either using the http 1.0 or http 1.1 protocols.

    The main difference is that http 1.0 domain names all have their own IP address. This means there is a direct, 1 to 1 relationship, between domain name and IP address.

    Under http 1.1 many web sites may live under 1 single web address. This means there is a many to 1 relationship, between domain name and IP address.

    With the massive growth of the web and other services on the internet it was thought IP addresses would become scarse and so because it is so much more frugal with IP addresses http1.1 became the norm in web hosting.

    It is more likely than not that your web sites share their IP addresses with many other web sites without you even realising it.

    Now onto the other 2 parents :)

    Every time a domain name is registered, at least under .com, .net and .org) information is recorded about the domain owner, technical contact and person to bill. This information is publicly available and is stored in a variety of databases the world over.

    Finally there is content. The meat and potatos of a website, the very words that make up its being.

    As I said above the search engine's had a problem and some clever guys thought of a way to help them solve it using an algothey thought of called Hilltop. Hilltop is a very vey very clever idea and many people (myself included) believe that it is active right now in the world's most dominant search engine.

    In essence The Hilltop algo extends the previous Google algos by still using links to ascertain the relevancy of a web page and it's authority on the subject matter it purports to but it does it in a special way.

    It looks for genetic traces between web sites and their extended web site families.

    Pre Hilltop it was a standard SEO technique was to build lots of web sites and link from one website to another to another to another to another so that link popularity would build and a websites' ranking in the SERPs would raise. Hilltop said "Woooahhhhhhhhhhh there SEOer. I don't think this is fair or right for us, our bank account or our searchers" so we are going to look at the genetics of a website linking to another and throw away those that are geneticly similar.

    The original Hilltop Algorithm states that:

    "Two pages are non-affiliated conceptually if they are authored by authors from non-affiliated organizations."

    It goes on to state that:

    "We define two hosts as affiliated if one or both of the following is true:
    They share the same first 3 octets of the IP address.
    The rightmost non-generic token in the hostname is the same.

    and further says:

    "The affiliation relation is transitive: if A and B are affiliated and B and C are affiliated then we take A and C to be affiliated even if there is no direct evidence of the fact."

    They also recognise that:

    "In practice some non-affiliated hosts may be classified as affiliated, but that is acceptable since this relation is intended to be conservative."

    I can hear the SEOers jumping up and down shouting that all their hard work on link buildinfg will become defunct unless the links come from an page that is an expert on the subject matter AND not genetically related to their own !!

    I believe it gets worse than this as Hilltop only says that a website has 2 parents:

    An IP address
    and a Hostname

    I think that for genetic cleansing to work in the SERPs that Hilltop has been adapted to take into account the other 2 parents I spoke of above, the whois information and the content itself.

    Why stop at identifying relationships at IP address and hostname when Whois info is freely available and why not check to make sure that the content a website has is unique and not a copy from somewhere else ?

    SEOers can still work to get high rankings under Hilltop by looking to build Expert pages that will link to their target site they wish to raise the SERP ranking for but they'll have to be careful when they build them.

    An Expert page will HAVE TO make sure that it does not have any siblings or distant cousins as the target site.

    I'll give you an example:

    Website 1

    IP namespace:
    Generic Hostname: foobar
    Whois Info Line 1: abc
    Whois Info Line 2: def
    Whois Info Line 3: ghi
    Whois Info Line 4: hij
    (all the way to)
    Whois Info Line N: xyz
    Content: fingerprint#12345678

    Website 2

    IP namespace:
    Generic Hostname: barfoo
    Whois Info Line 1: abc
    Whois Info Line 2: abc
    Whois Info Line 3: ghi
    Whois Info Line 4: qwe
    (all the way to)
    Whois Info Line N: xyz
    Content: fingerprint#rer6567fdg

    Website 3

    C class IP namespace:
    Generic Hostname: barfoofoobar
    Whois Info Line 1: hgfh
    Whois Info Line 2: kiuo76
    Whois Info Line 3: 343
    Whois Info Line 4: fggji
    (all the way to)
    Whois Info Line N: 56hfghg
    Content: fingerprint#656fdsfdsf

    Website 1 is a quarter sibling to Website 2 due to Whois Info
    Website 2 is a quarter sibling to Website 3 due to IP address
    which means that
    Website 1 and Website 3 are distant cousins!

    In practice this means that if, as an SEOer you wished to construct a set of Expert sites for the niche areas you operate in you MUSt do at least the following.

    Check the relationship between the web hosting company and specific class C addresses against the target SERP gaining web site

    Check the relationship between the target SERP gaining web site and the expert hostnames

    Check the relationship between the target SERP gaining web site and the expert whois info

    Check the relationship between the target SERP gaining website and the expert content

    And you MUST do this for all the websites, domain names, content matching or C class of IP address that score a "HIT" for any relationship as well.

    Although this is not impossible I forsee many SEOers saying this is too much work and the genetic cleansing of the SERPs will work!

    So what do YOU think of the genetic cleansing in the SERPS and are you checking each of your web hosts for IP relationships on an ongoing basis ?


    Jason Duke is the owner and operator of Strange Logic, the business that helps your business which incorporates the Logic Diary the industry news resource for your internet business that quickly improves traffic and
    sales by using scientific analysis of the search engines.

    Widget Words is the definitive keyword resource to assist in your search engine marketing and pay per performance campaigns and Widget the site that delivers industry specific reports for your target area of business.

    Jason can be contacted by email or telephone +44 1708 762497 to discuss any of your needs or simply just for an informal discussion about the marketplace and his unique point of view.