Google of course is just a big computer which counts and categorizes. So lets look at how it learns about what is related? Lets use the example of learning about writing standards, which is an educational question. Lets be more specific and talk about how each state defines it.
Each state has its own educational standards. For each grade, there is specific grade level curriculum standards. One of these is writing. How does google understand the concept of teaching writing? How would google find out what links to a page about teaching writing are relevant and what aren’t? Here’s the principles.
1. Authority sites and word counting. If you are an authority site, Google treats your opinion seriously. They count the words on your site. Lets say you use the words teaching writing alot. Well, your links out with the words teaching writing will really count a lot. Lets say you talk about different ways of teaching writing. Google will understand this if the word count is high. So for instance, on your authority teaching writing site, you talk a lot about six trait writing and four trait writing. Google will consider these to be relevant to each other based on the correlation. Google does a lot of word counting and correlations.
2. Neighborhoods and links and authority sites. Google also learns about education and associations by looking at which sites link to what sites. If there are ten authority sites which link to each other a lot, google considers them a neighborhood and the content is somehow associated. How does google categorize the association? Again, they rely on keywords.
3. Does google start with any taxonomy of what associations to look for or is it purely statistical? Traditionally, they relied on DMOZ and the google directory is an exact replication of the dmoz structure and content. Is this still relevant or has google learned more from their statistics of words and links (and perhaps visits) or do they still rely on the human edited directories for some sort of organization?
I don’t know. One of the best intermediate seo sites is perhaps worth asking….
