May 18, 2010
On June 17, 2004 Google filed a patent entitled “Ranking documents based on user behavior and/or feature data” to improve the way that they value links that point from one document to another.
Here is a quick synopsis: A system generates a model based on feature data relating to different features of a link from a linking document to a linked document and user behavior data relating to navigational actions associated with the link. The system also assigns a rank to a document based on the model.
On May 11, 2010 this patent was granted and so we are going to look into a little more of what this means. In the early days of Google they created a ranking system called PageRank to assist in determining the value of links. Google explained that PageRank could be thought of as a model of user behavior.
Random Surfer Model
This model was used as a random surfer who visits a document and then would randomly click on links to other documents. This model did not make any indications of understanding how a user would be valuing links on a page that they wanted to click on, the assumption was that users would randomly click on whatever links they wanted to.
Reasonable Surfer Model
This model would use the principles that a user would visit a document and assess which links they would like to click with a higher probability over other links. This model would use the systems and methods that would correlate in how a user is analyzing the value of what links to click over other links.
Features of links to account for:
- Position of the link
- Font size of the link
- Color of the link or other attributes (ie. italisized, shaded, bolded, etc.)
- Type of link (ie. text link, image link)
- Number of words in the link
- Commercial intention of the link
- Context of words before or after the link
- Actual words of the link
- and more
User Behavior features to account for:
- How often a links is clicked
- How often a link is not clicked
- How often no links are clicked on a page
- How users interact with navigations
- Query terms used on a website search
- Forms completed or used
- Language of users
- and more
Overall Google is continually improving how to make their computers analyze documents the way that users analyze documents. Google’s sole focus is the end user because after all users go to Google to find things and if Google cannot pull back the best results that that user is expecting then there is problems.
This will continue to make website usability very important as we analyze how users interact with websites. I have always said that user behavior does affect SEO because if no one wants to use your site or they do use it but you can tell they struggle using it then that can play huge roles in how they find your site, whether they bounce off, whether they want to share it with their friends, whether they want to bookmark it, whether they want to link to you and a lot more.
This new patent that was granted to Google will give them a huge advantage as they improve the “Under The Hood” of their search engine algorithms. Having the technology and patent to be able to analyze documents based on features and user behavior is a big thing and another reason that Google’s technology will continually make Google such an asset when users are trying to find things.