Since I'm in the process of re-building the information architecture for two clients this week, I thought I'd re-visit Dr. Garcia's paper - On-Topic Analysis.

For those who aren't familiar with the concept behind the research, on-topic analysis is a system that's designed to help you organize your site's content by topic based on how search engines are likely to perceive the categorization of information. The research uses the search engine's own index to conduct tests to discover which topics are most related to one another and then advises that site builders utilize this technique to construct an outline of how their site's topics should be organized.

This topic is getting increasing attention because so many people are noticing the search engines shift away from pure external link analysis towards rating the actual quality of the content, site and structure of the SERPs. This trend is probably part of a natural pendulum swing away from pure-link based ranking models like PageRank and HITS into an era where search engines would like to have their algorithms uncovering quality pages even if they aren't the most popular ones on the web. Obviously, link spam contributes to the desirability of this type of model.

Even if you can't grasp the math and processes neccessary for on-topic analysis, I highly recommend glancing at the paper and thinking about the principles of information architecture when designing your site. This is an area where what visitors and search engines want should be remarkably linked up, and a little time spent organizing your topics and hierarchy can go a long way towards both rankings and usability.