Friday, September 15, 2006
Findory is based on an interesting concept: "The more articles you click on, the more personalized your Findory homepage will be!" They don't go into too many details on their "Personalization Technology", but it apparently works by looking at your reading history (what articles you click on) as well as what other readers look at. If this sounds similar to Amazon.com's personalized recommendations and "Customers who bought this item also bought" feature, that's because the founders used to work at Amazon. The plus side is that you can use Findory anonymously.
As I clicked through articles on the site, I could literally see the news items, videos, and podcasts shifting in real time in response. Very impressive. You can even add specific blogs to Findory on their "Add a Blog" page.
There are a lot of parallels to Yahoo's Mindset search algorithm that allows you to filter search results based their commercial or research content. In Findory, you get to train the algorithm to fit your needs. My only concern is that the algorithm isn't adaptive enough. It works well when you're only interested in a narrow area like robotics or U.S. politics, but if you mix and match your clicks on climate, genetics, Iraq, Mozilla, and MySpace, then the algorithm may get "confused". Still, given the vast amount of information that is out there, Findory is a valuable tool in filtering through it all.