louisrosenfeld.com logotype

Home > Bloug Archive

Jun 22, 2004: Social Network Tools in Search of Solutions

Like so many others, I've been experimenting with social networking tools like LinkedIn, Ryze, and the latest entry, Google's Orkut. And I'm probably not alone in wondering about the utility of these tools. I mean, it's sort of neat that I'm one degree of separation (via Cameron Barrett) away from Wesley Clark, but Orkut's intervention doesn't make it any more likely that Wes and I'll be sitting down and nibbling his favorite "regular guy" snack, Cheetos, any time soon. (Though I'll admit I'm surprised: I'd always pictured the General noshing on something square and rigid, say a Triscuit or Wheat Thin.)

Image of Wesley Clark's Orkut profile
I'm one degree away from...

Image of Wesley Clark's Orkut profile
...a guy who sees Cheetos as a cuisine.

I'm disappointed that these tools rely solely on user-generated profiles to create social connections. It may be sort of useful to let the world know that one lives in Michigan, is looking for activity partners, and drinks regularly. But these metadata don't describe me in any way that's close to adequate. And those "friends" of mine? Well, some I've known for almost twenty years, some I've met just once, and some--well, honestly, I have no clue who they are or why we're connected.

Image of my Orkut friends
Some of these people are friends, other barely acquaintances. Some I've never communicated with directly, and probably never will.

Orkut's approach demands that we describe ourselves with really simple controlled vocabularies (e.g., "industry" and "relationship status") and other highly structured information (e.g., "title" and "high school"). Pull-down menus and simple fields are great ways to quickly populate a user database without taxing those users terribly. But the drawback is that we really can't say much useful about ourselves or about our relationships to others. Providing more is simply more work than we're motivated to do, rendering metadata-based profiles increasingly useless.

When the skimpy metadata in our profiles don't do the trick, and we don't find ourselves meeting all sorts of new people, we desperately start populating our lists of "friends" with people we already know, thus defeating a major goal of social network tools. We end up using these tools to simply visually map and track the relationships we already had, which gets old after a while.

Once Orkut flames out, I'll move on to the next cool social network tool and keep hoping that the problems these solutions solve will present themselves. There is something addictive about these tools--they do improve with each new generation--but, like the General's Cheetos, they're just not filling.

I'll also keep hoping that these social network tools will take a new direction: that their developers begin looking to alternative ways of depicting social connections besides profiles based on explicit, self-assigned metadata. Specifically, it might be worth leveraging existing, implicit connections to unearth the social connections that we never knew we had.

We've already seen these implicit relationships driving recommender systems, most famously Amazon's "Customers who bought this book also bought...". And for decades in the research world these interpersonal links have been inferred by analyzing co-citation patterns. For example, if two authors are frequently cited together in other papers' bibliographies, perhaps they share a meaningful connection. You can see relationships this in action in CiteSeer. Can social network tools do something similar?

Image of CiteSeer page
These articles have been co-cited with the polar bear book, which might suggest mutual relevance.

For example, if Orkut could track links between or "co-linkedness" (like co-citation) among members' web sites, we might be able to discover people with whom we share common interests. Blogstreet already infers similarity between blogs, and there's the old search engine standby, "Similar pages," on which to draw some implicit connectedness between Web content and, potentially, content owners. There are probably plenty of other ways to unearth connections between people that might serve to greatly extend the value of social network tools. As Google is already armed with many of these techniques, it's not surprising that they're funding Orkut's development.

Of course, the relationships in Amazon and CiteSeer are relatively easy to track because they take place within fairly controlled, centralized and relatively structured information environments. Doing something similar on the open Web is another story altogether. But certainly not impossible.

If we keep seeding social networks with mere morsels of information about ourselves, those social network tools will just keep connecting us to people we already know. And this dubious level of performance will continue dropping as we let our personal profiles go to pot. Why bother to update something that didn't deliver much value in its most up-to-date state?

The first social networking tool that uses self-assigned metadata and takes advantage of the trails we leave as we create and use content on the open Web will leap far ahead of its competitors. And it may actually deliver useful new social connections. The challenge will be to arrive at just the right combination of user-supplied metadata (and not so much to be a burden) with metadata and relationships extrapolated from our usage of the Web.

email this entry

Comment: Amy (Jun 23, 2004)

I'm glad someone's raised this issue, because the actual utility of these tools has always escaped me. For example, I found Friendster moderately entertaining for the first three days after I was invited to join, but I've only looked at it once or twice since then. And I recently signed up for Monster Networking, but so far, all I've gotten from it is some completely inappropriate recommendations for people to network with, and some related spam from Monster.

Of all the social networking tools out there, I think http://www.dogster.com may have the most real utility. Any dog owner who's passionate enough about her/his dog to post the dog's profile on Dogster is probably also someone who would welcome meeting like-minded people (and their dogs). This probably also explains why there's no Catster yet.

Comment: Lou (Jun 23, 2004)

It's interesting you mention Dogster. I've been thinking that these social networking tools are just too generic; they probably work best as infrastructure for more specific domains and communities, like dating (high level of motivation to develop and maintain one's profile) and dog fanciers (but not cats).

Comment: Victor (Jun 26, 2004)

Recently, a friend who is graduating from the MIT Media Lab wanted to contact someone at Bose in MA about employment. He found the person on LinkedIn and was four degrees away, connected through me. So he used the system to make the connection, essentially asking me to ask my friend to ask the person at Bose for an informational interview (Linked In handles the routing).

"Ahhhh, that's how these are useful" said my brain. It's obvious, but it didn't really click until I had become part of an introduction, a personal connection faciliated by the technology. Until then it was simply a neat thing to see my friends all on one screen.

Comment: Lou (Jun 27, 2004)

OK, so we've got job contacts, dating possibly, dog fans... Still pretty limited utility, IMHO.

But add this technology to a large corporation's staff directory, and then I can see huge value. Again, specific domains need to be better identified.

Comment: Lou (Jul 9, 2004)

Just stumbled on this excellent post from Jeff Veen's blog:

Makes many of my points, only better and about three months earlier... ;-)

Add a Comment:



URL (optional, but must include http://)

Required: Name, email, and comment.
Want to mention a linked URL? Include http:// before the address.
Want to include bold or italics? Sorry; just use *asterisks* instead.

DAYENU ); } else { // so comments are closed on this entry... print(<<< I_SAID_DAYENU
Comments are now closed for this entry.

Comment spam has forced me to close comment functionality for older entries. However, if you have something vital to add concerning this entry (or its associated comments), please email your sage insights to me (lou [at] louisrosenfeld dot com). I'll make sure your comments are added to the conversation. Sorry for the inconvenience.