Search Engine Ranking Factors

wOOt, first post.

Quite an exhaustive list. At first glance I’d question how many of those would really be factored into search engine ranking, as they (some of them at least) would seem to require some serious crunching per page indexed.

However I suppose if you’re a certain search engine with tens of thousands of servers, who’s often notable for innovating rather than copying…

I noticed he lists having cross links between sites on the same “C-BLOCK” as being bad. So what do you guys with multiple blogs do? Host them all on the same ISP? Or use multiple ISP’s so your link to your own site counts more? Not sure if its worth that much trouble for one link.

Jeff

Or actually, for those who use services such as blogger or typepad, etc. I guess that means that links between blogs on the same system are not as useful. Since if site A is on blogger, and so is site B. Then how valuable is a link from A to B.

I don’t know if my understanding is correct, Linking from different ISPs though is troublesome, and then you will have to raise the rankings of all your web sites separately.

Whereas, I am only conjecturing since I am no expert, that if you have various web sites in one domain, it might take longer to raise up, but at least your eyes is just focus on building content for that one site.

The document has been updated with input from about a dozen people (including me), so it reflects some broader opinions than originally.

Don’t be concerned by the C-Block reference. Google doesn’t actually toss out documents on the basis of C-Blocks (according to a couple of engineers who have commented on the subject). The C-Block exclusion is found in a few technical papers that have been overdiscussed and poorly understood in the SEO communities.

One of the papers proposes a methodology called “Hilltop” for determining which sites are “authoritative” by looking at how many “hubs” (sites with multiple outbound links to topically similar or related documents) link to them. The paper proposes a minimum of two inbound links from hubs would establish some authority for a document. The authors were concerned about unintentionally favoring hub documents which came from the source providers (such as large directories, universities, document archives, etc.), so they proposed a system for classifying sources as “affiliated” on the basis of IP address (where if the first three numbers of two IP addresses are the same, then they are assumed to be related).

Another such paper proposes a methodology called “LocalRank”, where the results for a query (including up to 1,000 items) are used to rank themselves on the basis of how they link to each other. Again, the C-Block exclusion was proposed to minimize favoritism among “affiliated sites”.

Since Google’s engineers understand that shared-IP hosting is very common on the World Wide Web, they say they have not implemented the C-Block exclusion process.