Stanford Summit: Beyond Search

First up is a panel about going "beyond search." Moderator is James Jaoquin of Bridescale. Panelists are Paul Martino of Aggregate Knowledge, Joe Greenstein of Flixter, David Hyman of MOG, Bambi Francisco of Vator.tv. These are my notes (a few photos are available here).

Paul's company, Aggregate Knowledge, provides "discovery services" for the retail and media space.

Joe comments that Flixter is a site where people rate movies and share them with friends. For Flixter, discovery can be made easier along the way and is not only about the "magic search box". It's about that fundamental urge people have to share when they discovery something new that they think is great.

David describes MOG as a music resource and discovery tool. It helps people express themselves as musical beings. You describe yourself by the music you love. Different from a blog or myspace page. They generate about 1500 music blog posts a week of just people sharing the shows they saw, the music they heard. When you have that much content on the site, MOG lets you use the data sets to find the people who are most like you musically, and then browse their music collections.

Bambi describe Vator as a platform to let professionals network and create a marketplace for ideas and business. It's a vertically focused site, premised on innovation and entrepreneurship. The site organizes and aggregates ideas and innovation around particular topics. Discovery is about organizing the information - most vertical sites need to be set up that way. Another way they enhance discovery is to create networks around ideas. Analogous to the way social networks work, people group around ideas. They are experiencing a lot of international interest - 40% of users are from outside the US.

Paul & Joe agree that sometimes discovering from friends is often the worst way to discover. Interestingly, at Flixter, Joe says they are not trying to be perfect recommenders of films, the focus is on being fun. The result of focusing on the fun, is that they actually do get better recommendations out of it.

David talks about discovering things in the physical world, and how that's different in the virtual world. In physical world, you discover music by reading magazines, going to the record store, shows etc. At MOG, they have tried to replicate those same forms of discovery in the virtual world. MOG's technology actually tracks the music-listening behavior of the users, not just what you have purchased. This evidently is a much better way to match users tastes together and drive their system. David reminds us that "with any of these social systems, the more you put in, the more you get out."

Joe mentioned that discovery is different between music, movies, other things. The object of what you're trying to discover can drastically change how you discovery. If it's a $2 purchase, your habits are different from if it's a $1000 purchase. In movies, for example, nobody is going to watch a movie they've never heard of. But in music, people often listen to bands they've never heard of.

Paul's company encourages its customers to engage the user in a number of ways. Feedback, comments, "forward to a friend" links, rather than a static "we recommend" block somewhere on the site.

Question to Bambi - her service is the newest, just launched on June 6th - how has her thinking about adding more discovery? Vator doesn't have any algorithmic systems for discovery. They are using a social networking platform to the idea network. It's off the shelf social networking platforms. They've created a fairly simple platform rather than going super complex to start.

David's company never gives recommendations about what music the user might like. You'll never get a "MOG recommends..." from Davids company.

Paul shares a theory about the generations of search/discovery:
Generation 1 is about algorithmic discovery
Generation 2 is about putting users on top of algorithms
Generation 3 is the social realm.

Q: what is search for and how does it interface with discovery?
Search is the starting point for discovery. In MOG it's the most popular feature. Most people mostly want to find what they do know - which David explains is why pop music radio stations do well, and less people listen to college radio. People in music want to find what the know.

{One observation I have on this topic, is that my privacy radar is alerted, of course, with all the discussion about the data sets and how much better these services work the more data is included. Who owns the user's data? Do users care about this? Should user's have control over what companies do with their data/attention/click steam movements? The Attention Trust looks at these issues. I know of others who care as well. I question whether any consumer protection concerns are going to be addressed, depending on how the data is used. If it's aggregate, non-personal information, I see nothing legally wrong with the collection (provided these companies are giving their users proper notice and accurate statements regarding their use of the data), but I do see problems with being inundated by marketing that is not useful to the user.}

The privacy concern was sort of raised by question received from the web briefly at the end of the panel. David reported that for MOG, people's identity is protected through their avatar, so long as the user chooses a non-real screen name & email address. At Flixter, Joe explained that they created a movie compatibility test that let's people connect with strangers, but somewhat anonymously. Vator has some protection about harassment by limiting a user's contact with someone else to just once. Sounds like a good way to limit pestering from other users.