QUOTE(snipermilk @ Sep 13 2005, 10:08 PM)
Originally, I thought it was because there was no variation in the anchor text used so I changed the "Powered By ---" link to an image, but still no substantial results. I also removed all questionable onpage elements the previous SEO applied on the site. Same results. I think it might just be a matter of increasing the "uniqueness percentage" so I'm now tempted to remove the "Powered By ---" link altogether. It's that, or move to different domain :goodjob:
Am I being penalized for having a low unique IP link count?
You would not be penalized for having a low IP link count. The significant of same C-class linkage has been misunderstood by the SEO community because people have taken references to this phenomenon out of context.
If you search for
C class links google you will find there are many people discussing their impact on rankings. See below for some comments on this phenomenon.
I can tell you from persnal experience that having tens of thousands of links from the same C class doesn't hurt my rankings in the least. Nor does it hurt the rankings of the sites I compete with (many of which are large commercial entertainment venues).
Poor performance in search results almost always results from one of two classes of factors:
1) Poor optimization (you indicate you have taken steps to improve your optimization)
2) Whizbang competition from people who are way ahead of you
You may just have to consider that your competition is way ahead of you.
About Google and tossing out groups of IP addresses. A lot of people have written various papers and analyses of the Google patents which suggest that, when ordering search results, the query tool may discard listings that come from a related group of IP addresses (they all come from the same C-block, X.Y.Z.???).
Most of the people who have gotten caught up in this overconcern regarding same-C-block addressing have disregarded the contexts set down by these patents.
For example, one patent
discusses how to refine previously performed searches identified by cookies -- this is a personalization patent. It is not a generic search patent.
QUOTE
When a search is made for information on the Internet, certain search parameters are saved in the user's system. If a subsequent related search is made, these search parameters are transferred to the internet search engine and used to reorder or otherwise modify the search results. Preferably, every time a search request is made, an entry is added to a cookie, containing the search terms and date of search. The search engine re-orders search results by dividing the results into multiple groups, including (in order of priority: (a) those URLs which are new since the last search; (b) those URLs which have been previously visited by have changed, the magnitude of change determining the ordering within this group; © those URLs which existed before the previous search, but have not been visited by the user; and (d) those URL's which the user has previously visited and have not changed. In the preferred embodiment, the user may optionally manually specify certain URLs to be included or excluded. The ordering of search results in accordance with the present invention provides the user with greater control and more meaningful information than current search technology.
(NOTE: Emphasis is mine)
Another patent discusses
how to improve ordering by relevancy within a preselected group of documents. This is the infamous and much-misunderstood "LocalRank" patent, wherein Google's engineers explain how, AFTER a group of documents is selected, their internal linkage (to each other) is used to determine a relevancy sorting that promotes (in Google's automated estimation) the most unique and relevant documents to the front of the list.
QUOTE
A search engine for searching a corpus improves the relevancy of the results by refining a standard relevancy score based on the interconnectivity of the initially returned set of documents. The search engine obtains an initial set of relevant documents by matching a user's search terms to an index of a corpus. A re-ranking component in the search engine then refines the initially returned document rankings so that documents that are frequently cited in the initial set of relevant documents are preferred over documents that are less frequently cited within the initial set.
(NOTE: Emphasis is mine)
LocalRank would not be applied in most searches (they would not return enough documents with that kind of internal connectivity). Personalized searching has been implemented through Google's (Beta) Personalized Search interface, which maintains a search history (at least one other patent delves into details relating to this service).
In many SEO forum discussions, for reasons I cannot fathom (except possibly that people haven't actually READ the thing), many people will mention
Hilltop in connection with LocalRank.
Hilltop introduced a concept whereby a subset of indexed pages deemed to be "experts" (because of their outbound links related by topic) is used to pick the most "authoritative" documents in a results set and push them to the top. A document is "authoritative" if two or more "experts" link to it for the same topic.
QUOTE
In response to a query a search engine returns a ranked list of documents. If the query is broad (i.e., it matches many documents) then the returned list is usually too long to view fully. Studies show that users usually look at only the top 10 to 20 results. In this paper, we propose a novel ranking scheme for broad queries that places the most authoritative pages on the query topic at the top of the ranking. Our algorithm operates on a special index of "expert documents." These are a subset of the pages on the WWW identified as directories of links to non-affiliated sources on specific topics. Results are ranked based on the match between the query and relevant descriptive text for hyperlinks on expert pages pointing to a given result page. We present a prototype search engine that implements our ranking scheme and discuss its performance. With a relatively small (2.5 million page) expert index, our algorithm was able to perform comparably on broad queries with the best of the mainstream search engines.
Hilltop excludes authoritative pages which are pointed to by neighboring experts. That is,
QUOTE
...We felt than an expert page needs to be objective and diverse: that is, its recommendations should be unbiased and point to numerous non-affiliated pages on the subject. Therefore, in order to find the experts, we needed to detect when two sites belong to the same or related organizations.
2.1 Detecting Host Affiliation
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.
The only connection between Hilltop, LocalRank, and Personalized Search is that all three algorithms attempt to provide the user with diverse, apparently non-affiliated content that is relevant to the user's query by excluding or minimizing the value of documents from "related" IP addresses.
In most searches, you're not going to get many documents from the same family of IP addresses. Natural content establishes relevance to a random number of topics. So, if you have 500 pages of natural content on your site, the chances that more than a handful of those pages will be deemed relevant enough for inclusion in the initial set of results (that is to be sorted before presentation to the user) are slim to none.
If you control 200 IP addresses and each has natural content spread across 100 pages (that is, you have 20,000 pages of natural content), the odds that any significant number of unique pages will be deemed relevant enough for inclusion in the same results set are still very slim.
The assumption behind discarding related IP addresses is that some UNNATURAL relationships may have been established, or that some duplication or mirroring of content may be in place.
University sites, for example, can and do mirror content on related IP addresses. So do search engines and directories. Therefore, it behooves Google to discard duplicate content from such sources by selecting only one IP address from a C class block. Is it possible that legitimate content will also be discarded? Yes. But the likelihood of that content being buried for every possible query is extremely small.
The SEO community is well aware of popular spamming techniques that have been used to game the search engines. Yes, there are (or have been) services that take 32,000 consecutive IP addresses and place almost identical content on them. These algorithms are designed to minimize the impact of that kind of unnatural diversification and replication. If you are not doing that, you don't need to worry.
In your particular situation, all you have are a few hundred links from sites carried within the same C-block of IP addresses. You imply (in your reference to client sites) that the links are embedded in unique content. It is highly unlikely that you are being impacted by any of those patented algorithms' discarding features. We don't even know for sure that Google is currently using any of them.
At this point, I think you need to look at what your competitors are doing and see if you need to play catchup with them in one or more areas. They may simply have buried you through linkage (on client sites).
Edited by Michael Martinez, 14 September 2005 - 02:18 AM.
