A newly released site, known as Smatchy (www.smatchy.com), presents an interesting take on recommendations and could become more of the standard than the exception.
Smatchy’s premise revolves around users answering questions and then calculating ‘People Like Me’ for users in fourteen different categories. This ‘People Like Me’ statistic drives almost every feature on the site, linking people together based on how similar they are and not any user-defined criteria. While the math behind this seems to be fairly in-depth, the site has produced a slick interface to navigate all of the options, making everything from figuring out your ‘Smatches’ to querying the dataset fun.
So how do the recommendations fare? Well, pretty well, to be honest. Especially when you consider a lot of the questions are addictive and fun (unlike a strict ranking system). My favorites ones to answer included:
- I’d rather have a root canal than live in Texas.
- It doesn’t get any better than Where the Wild Things Are.
- I would rather be friends with John Cusack from ‘Say Anything’ than Matthew Broderick from ‘Ferris Bueller’s Day Off.’
For recommendations, you get the choice of either ‘Normal’ or ‘Obscure’ methodology. While the Normal are good, the real gems are in the Obscure, where unique books, movies, and music are pulled from user’s profiles based upon your matching. Finally, the Forum makes good use of the system with how well individual authors match up to you being displayed alongside the post.
Ruby on Rails is the power behind Smatchy, and does a good job on handling the extensive database while keeping the navigation fun and easy. Rails development being what it is made it fairly quick and painless to put Smatchy on Rails!
As more and more people latch on to this concept, you could just see a better way to get recommendations and interact with other searchers.
[Note: I was personally involved with the site’s Rails development.]