The folks at Yahoo Labs have been pretty busy lately. They recently published a paper titled “Building Enriched Document Representations using Aggregated Anchor Text” (opens a pdf file). It’s an interesting read.
The paper is all about how to rank web pages based on what the anchor text pointing to those pages say about them. The problem is, there are a lot of web pages on the Web that don’t have links pointing to them. So what the does the search engine do? Well, you look at the Web graph (a fancy term for the aggregated whole of the Web – pages, links etc.) and consider the anchor text of external pages (pages external to the website on which your page exists) that point to that page.
The paper refers to what is called anchor text sparsity, which it defines as the difference between the small number of sites that have a large number of inbound links and the large number of sites with just a few, or no, links. This sparsity is the hurdle to overcome.
Here’s the problem: If you’re one of those sites, the majority, with few inbound links to it then your site is more difficult to figure out for the search engines. Why? Because inbound anchor text helps to define a web page. Let’s take an example.
If you have a web page that has 150,000 inbound links and 85,000 of those use the anchor text “spaghetti sauce” then it’s pretty easy to see what that web page is about, isn’t it? But if your page only has 2 inbound links and they both have different anchor text then it’s more difficult to know what that page is about based on the anchor text. Now see the problem?
If you do then you should see the solution. It’s time to start building your collection of inbound links with the proper anchor text. Don’tcha think?