Inveni claims that robot “recommendation algorithms” are dead. Where an algorithm can take certain categories of a film and make recommendations based on your past history, ratings and viewing patterns it is their assertion that old-fashioned ‘friend’ recommendations are best. Officially launched at TechCrunch Disrupt this year, Inveni aims to solve this problem.
When you sign up for Inveni – using a normal username/password or via Facebook Connect, you will first import any past ‘ratings’ you’ve made on other sites including Netflix, Blockbuster, IMDB and others. These ratings form a basis for future recommendations.
You then proceed to let the service know what some of your favorite movies and categories are. Starting from this point Inveni begins to recommend movies and television shows based on your interest.
To further personalize your movies recommendations Inveni encourages you to invite your friends to the service or to find friends already using it.
The ‘find friends’ feature seemed to need some work to me. How it is currently set up is not only a security problem (since it will search the list of all users, not just people you already know, you can harvest lists of names off of the service) but it seems like there should be a better way to find friends, either via Twitter or Google Contacts import or some other method.
Integration with Netflix is a dream. It works with Netflix’s API to import your existing recommendations as well as display what is currently in your Netflix DVD and Instant Streaming queue. You can effectively switch out of the Netflix recommendation engine and start using Inveni to find movies on the Netflix platform.
Also of note is the Firefox toolbar integration. After you install the Plugin it will pop up (as a bottom bar) on dozens of sites which list movies such as Amazon, Netflix, Best Buy and others. This gives you an integrated way to find out what movies have been rated by the Inveni community and also continue to add your ratings to movies – therefore making it more accurate in its own recommendations.
All of that being said, I think Inventi is on to something. Although in its current setup, with most ratings pulled in from Netflix or other services, I’m not sure how it differs from any other “robot algorithm” services but they are making the claim that once you have a network set up the recommendation engine is more heavily influenced by friends than algorithms.
Only once you add in the social factor does the service start to differentiate itself, and with the lack of a wide-base of users at the moment, it is hard to argue that it does any better than Netflix itself. If I were to make a recommendation to the company, it would be to get better integration with Twitter and other social networks to ‘widen the net’ so to speak and get as many of your friends onto the service as possible. Until then, services like [NO LONGER WORKS] Fflick are simply easier to use.
Do you use Inveni? How do you find the recommendations versus what you would find in a “traditional” recommendation engine?