Welcome to WebmasterWorld Guest from 220.127.116.11
It would make sense that if an engine wanted to detect link buying a 'red flag' might be identical anchor text to a site/page from thousands of pages or sites. [That also might be an indication that the site/page is highly relevant to that anchor text phrase.] Which is it? What is 'natural variation' - does it exist?
If a site has internal links pointing to a page, I'd imagine that variation in the anchor text going to a single page may indicate deception to some degree by an engine. [Shouldn't there be consistency in anchor text on a site going to a single page?]
I track the deep link ratio of my sites, but have not been able to determine if there is a correlation with rankings. Is there a consensus on the importance of links going to many pages in a site versus just the homepage?
I am interested in your tests and experiences regarding variation in anchor text for both internal and external links as well as thoughts on deep link ratio. I am also interested in knowing more about penalties related to anchor text in general.
This would be just as plausible as the notion that uniform anchor text in IBL might indicate an attempt to manipulate the results.
In the end, I really doubt that google cares about these issues, other than in clear cases of link schemes.
Regarding internal links, why should google care what a webmaster uses as anchor text? It's the webmaster's own site and he/she can describe the pages any way he/she wants. Deception would only exist if the pages were something other than what the internal links "claimed them" to be. Now, if internal anchor text unduly influences serps, this is really google's issue and I think they view it as such. The appropriate response would NOT be to penalize websites, but rather alter their search algos to minimize the value of internal links.
Regarding external links, let's use the example of advertising. Let's say you buy a run of text links on a well known site for advertising and this is for a site about red cars. You would naturally expect the links to include the words "red" and "cars". Why would the links be varied? It wouldn't even make sense, other than an attempt to avoid "perceived" negative treatment from google. But if that's the only reason for wanting IBL anchor text to be varied, then I submit that there's no valid reason for wanting IBL anchor text varied----because google is not that stupid or simplistic.
I have about 50 subsections on the website and each subsection index page links back to the home page with that subject's keywords.
The result is I get two search results back - the subsection index and the homepage on most searches.
I also employ a similar technique for pages and subsubsections further down the tree.
All kind of factors would/should generate natural deviation from identical source text. So the benchmark would vary in a granular way. It would vary by the kind of domain (keyword or branded, multiple keyword, etc) the kind of link (to the domain root, deep link, etc.), how many links are on the source domain, the positioning of the links on the source domain (in content, in footer, in menu, etc), the history of both source adn target domains, the age of the linking profile overall -- whatever G wants to take into account. I'm sure they take a few factors in.
A domain would need to get far afield from the benchmark, and again, to a statistically significant degree, to throw a flag.
the search engine is a war.
each keyword (internal page) is a battle.
the more battles you win ( more search engine placments and higher),
you end up winning the SE war!
you have to look at ranking a page, and not a site. the more pages on more sites you have, the more keywords you're able to target. more keywords being targted means more $$
the second question. i've heard 60 percent homepage, 40 percent inner pages.