This week, Facebook started a new system called Open Graph. The premise behind Open Graph is that individual websites can place a ‘like’ button on their website, and readers can ‘like’ a post, article, or the site itself. This information gets relayed back to their Facebook profile, and their followers. Facebook announced agreements with some major websites right off the bat including the New York Times, Yelp, Pandora, ESPN, IMDB and CNN, with many many more following suit already. No doubt these widgets will soon adorn most of the major websites.

I am skeptical, however, that there will be a high percentage of ‘likes’ that will show up in a user’s stream that they will actually have any interest in visiting. Part of the problem, I suspect, is the fixed, singular nature of the Facebook social graph. I have no hard evidence on whom most people become friends with on Facebook. However, I suspect that almost all friends are people who know each other in real life, and who have likely spent a reasonable amount of time together in person (in aggregate, as there will be obvious outliers).

I believe that this type of social graph is limiting. Sure, there is the odd time where knowing what old high school friends enjoy is interesting, but it’s more the exception than the rule. However, there are other social graphs that are more relevant given any interest. Some interests might only apply to a small subset of ones Facebook friends, and some social interests might be completely independent from ones Facebook friends. In fact, for any real interest that I really care about, and specialize in, the number of Facebook friends who also cares is likely very small. As an example, I have almost no Facebook friends who are also Stewart Copeland drum fans. A social network of any given interest group is totally independent from the Facebook social network. Facebook’s social graph is only one graph. I’m not sure that putting all this information into a single static graph makes sense.

An example of something equally irrelevant is television advertising. Unfortuantely for the advertisers, it is impossible to know exactly who is watching the television program, and what they might be interested in buying. You probably wouldn’t want to show me commercials for tampons when I (probably) wouldn’t ever buy them. They’re wasting their advertising dollar on me, and because I don’t care about the ad, I dislike advertisements just a little bit more. Facebook ‘likes’ aren’t quite that bad as there is something in common between all of ones Facebook friends (they are all friends), but almost that bad. Together, they represent the lowest common denominator of ones friends with respect to that interest. Certainly if I ‘like’ Stewart Copeland, I suspect the majority of the friends on my friends list would rather not see it than to actually see it. This is a waste of their attention. And we should care about our friends attention. Contrast that to having your own website dedicated to a topic, where only people interested in that topic will read it.

Of course, I would love to see all Stewart Copeland friends, and see what they like, with respect to topics relevant to those people. I’d love to know what other bands they like. For each individual thing I care about, almost certainly, my Facebook social graph is not the appropriate graph. No doubt people I’ve conversed with in person, and perhaps online, are the most important as they can be the most trusted, so that needs to be included in a good social system. However, I doubt that Facebook is the ‘be all and end all’ of social networks, and likely a new one needs to be built from the ground up, with multiple independent networks in mind, to accommodate this task. It might be possible to “simulate” some of these ideas within Facebook, but unless a concept is completely front-and-centre in the interface, that will not be how it is used overall.