Thanks to Gen Kanai, I was able to join Gree, the new Japanese-language social networking service, and see how it compares to some of the more well-known English-language social networking services. In this analysis I will compare it to Orkut.

The first difference I noted was the distinction made between “links” and “friends.” Like Orkut, Gree.jp encourages users to invite their friends by entering their first and last names and email addresses. Orkut also allows you to browse to someone’s profile and add them as a friend. Pending their acceptance, they will be listed as your friend. This is where Gree has made an interesting design decision. In Gree, you is no option to “add as a friend” a profile to which you have browsed. Instead there is an option to “link to” (リンクする) the profile. By default, attempts to link to profiles to which you have browsed are rejected. This is a setting which members can change, so that they can allow links from friends of friends, or from anyone.

This is such a departure from the established rules and conventions of social networking sites like Orkut, Friendster, and Tribe that it took me a while to understand that this was happening. This suggests that genres are already forming in this space, and provides an example of how these genres are self-sustaining: if designers want their sites to be usable, they may opt to stick to genre conventions in order to make the use of their site immediately intelligible. Perhaps the Gree designers thought that most of their users would not be familiar with the existing (English) social networking sites and thus felt free to flaunt convention.

Gree’s design is a nice solution to the issue of unwanted contacts that often arises in Orkut. Orkut makes it so easy to “add friends” that people are faced with the dilemma of choosing to accept or reject friendship requests from people whom they do not know. By default, users of Gree must know the first and last name and email address of anyone to whom they want to link. Not absolute proof of a close relationship, but better than “I saw your profile and think you’re cool.”

Unfortunately, Gree almost negates this advantage by listing the users with the most friends, a mistake Orkut has avoided. Looking at this from an information theoretical perspective shows why this is a mistake: making this sort of list visible turns the site into a collecting game and encourages spurious “friendships.” Hopefully the Gree designers will see the light like Friendster did.

The second major difference involves the placement of testimonials, or “introductions” (紹介) as they are called in Gree. In Orkut, all the testimonials written by a particular user’s friends appear below that user’s profile. In Gree, all the introductions written by a particular user for his or her friends appear on his profile (each introduction appears next to the picture of that friend).

This simple twist completely changes the concept of the testimonial. In Orkut, testimonials serve as ego-boosters, or maybe as selling points for potential suitors. In Gree, however, they give people a chance to qualify or explain their relationships. Looked at this way, the introductions in Gree serve a purpose similar to Orkut’s ill-conceived “rainbow scale” for rating friends, except that it allows far more nuance.

We can look at the two approaches as different ways of building identity online, using Goffman’s concept of identity as something that is socially constructed via feedback loops in particular social contexts. The Orkut approach asks the user to construct his identity from the words of his friends by selecting the testimonials that flatter his self-image. The Gree approach asks the user to construct an identity by selecting a group of friends and weaving together a story about how the friends relate and the role which the user plays in that story.

Gree also incorporates some interesting innovations. Users can register a blog or web journal with Gree, and syndicated updates will appear with their profile. Users can see an aggregated page of recent updates from all of their friends’ journals. The effect is that Gree seems far more alive than Orkut—it seems like a place where things are actually happening, instead of a static gallery. It also adds yet another redundant channel for expressing one’s identity, reducing the uncertainty inherent in the online medium.

Gree also inverts the “interests” feature common to many social networking services. Orkut users can list books and music that they like, presumably on the assumption that other users can use this information to gauge compatibility. Gree, on the other hand, asks users for “recommendations” (オススメ). Rather than using common interests as a basis for judging the possibility of friendship ties, Gree encourages the use of friendship ties as a way to find new interests. The latter seems a lot more useful, and is a good example of peer-to-peer search structured by social connections. The recommended items are linked to the Amazon.co.jp. It would be interesting to compare how satisfied users are with purchases made due to Gree recommendations as opposed to Amazon’s collaborative filtering recommendations.

With the exception of the popularity lists, Gree also puts much tighter limits on visibility than Orkut. For example, search is constrained to first name/last name, where Orkut allows search by gender, location, relationship status, etc. This seems to fit with Gree’s general design philosophy of reflecting real-world friendship networks rather than creating new ones: search is meant to find people you know, not people you’d like to know. Also, whereas Orkut allows a user to see the path (if one exists) from him/herself to the profile currently being viewed, Gree shows no paths. I’m not sure if this a conscious design decision or just something that isn’t implemented yet.

One thing that I think would be quite interesting to study is the effect of the fact that Gree is implemented in Japanese on the social networks articulated within it. Unlike English-language social networking services, which can reasonably expect to spread around the world (witness the popularity of Friendster in Southeast Asia), Gree’s users are likely to be very concentrated geographically. Of course there are Japanese expats around the world, but most users will live in Japan, a country smaller than California, within which most of the population is concentrated in a few urban areas. I would be interested to know if the greater possibility of physical proximity strengthens the Gree networks as compared to Orkut. My brief experience with the site suggests that this is possible. For example, the “events” pages of most Orkut “communities” seem to be unused, probably because the community members live thousands of miles away from each other. Gree’s “groups” (グループ) on the other hand, seem to utilize event announcements far more, probably because the majority of users live in or near the same city (Tokyo).

In general, Gree seems to be well designed and well thought-out. It is certainly refreshing to see some innovation in this space. In retrospect, it is disappointing that Google didn’t take similar risks with their service.