Hi LD - Thanks for that example, and I'm sorry for the delayed response. Here's the skinny on how related content works:
First, when our plugin is installed on your site, our server attempts to index the pages of your site by checking for an XML sitemap and retrieving the URLs from this list, passing them into our processing queue to be crawled as individual pages.
When the pages are crawled, we extract various pieces of data that are used in calculating the relationship of one page to another. Meta tags are definitely useful to provide more information about your content.
We have continued to work on the intelligence portion of this app so as to improve the calculated relationship between pages. This is a complicated coding process, so I am unable to provide specific details in how the machine learns, but it is the backbone to the Related Content app.
Once we have data for each page, a file is generated to summarize a list of pages so that it can be passed along to the recommendations engine, which handles the delivery of your recommendations. The app is a very complex network of systems that work together to produce this feature, and it will continue to improve as we further refine the code.
We're working on this now, so hopefully we can deliver some improvements soon!