Redux Discovers Friends So You Don't Have To

redux_logo.pngYou’re probably familiar with the “recommended items” lists on sites like Amazon or Barnes and Noble. They help discover things you’ll probably like based on what you buy and what you view. Redux founder Darian Shirazi wants to apply the same power of discovery to growing communities of people online. The site is launching into beta today.

This site is part social network and part quiz show with the central goal being to learn the most about what you’re like and who you’d like. You fill out a profile with all the standard questions (music, age, sex, books) with the added bonus of an optional Myers-Briggs personality test. The site also lets users post photos, list their location, or chat with each other.

redux2small.pngHowever, the heart of the site is training their algorithm on what people you like. Similar to “I’m In Like With You”, users are periodically asked quiz questions by the site, such as “Do you like sports?” or “Do you have a teddy bear?”. Your answers to the questions coupled with your profile info and whose profiles you view help Redux recommend people like yourself with on a percentage compatibility score.

The algorithm isn’t “dumb” or based purely on matching up people based on answering questions the same. It actually learns what properties signify compatibility based on how people use the site and takes special care to match people up with niche interests (something Shirazi calls the “Anomaly Filter”). For instance, people who play sports will probably get along with people who watch sports. From there, the system could discover that people who play sports get along well with people who enjoy action films or any number of other attributes.

Finally, Redux closes the loop by encouraging compatible people to hang out at any of the thousands of public events they’ve pulled from sites like Upcoming.

While I have yet to make a friend through a random conversation over a website, there are clearly plenty of people who do. Members of social networking sites frequently drop a line to someone they might like to strike up a conversation. A service like Redux that matches based on personality makes the initial impetus behind the introduction more than superficial.

But Darian doesn’t think that Redux is necessarily about finding a life long friend. There has been a growing interest in personalizing the web and recommendations from real people are often the ones users trust. That’s why Redux will be distributing its service as a platform by the end of this year. Websites will be able to integrate with Redux to power all kinds of people driven recommendations. It seems like a smart move, because I can’t see most people logging on every day just to chat with friends and answer some questions. The context of another site can make the process more compelling.

Redux was created by the team behind Flick.IM and is funded by $1.65 million from investors including Peter Thiel.