August 18th, 2007

The Problem With “Friends” On The Social Graph


Think about all the different relationships in your life — parents, children, siblings, extended family, close friends, casual friends, acquaintances, closely collaborative colleagues, professional contacts, superiors, subordinates — our lives are an intricate web of relationships. Yet on the Web — with its capacity for near infinite complexity — these relationships have been reduced to a uni-dimensional descriptor: FRIEND.

Is this really the best we can do? Brad Fitzpatrick, founder of LiveJournal, has targeted social network interoperability as the BIG problem to solve for social connectedness on the Web — and indeed, the siloed nature of social networks could not be more un-Web-like.

But I think the bigger problem is the inability of the “social graph” on the web to capture the infinite variability of human relationships — and the limited nature of social applications, which don’t enable us to communicate and interact with each of the people we know in all the myriad and infinitely varied ways we do offline.

The reality is the the humble text email is still a far more powerful tool for tailoring each of my virtual social interactions to the exact nature of the relationship. On Facebook, Twitter, and even this blog, there’s only one, maybe two “settings” for my social connections.

Social networks are still in the dial-up connection phase — back in the 90s, we marveled at the connectedness of the Web, despite our utterly primitive connection to the network. Today, we marvel at the connectedness of online social networks, even though the connections are still primitive.

So what does the “broadband” phase of social networking look like (to continue the metaphor — yeah, I could just call it social networking 2.0, but whatever)?

Well, how about I share a piece of content and the network “knows” who to share it with. I send a message and the network knows who should receive it. I start a collaborative project and the network invites the right collaborators. I add a new person to my network, and the network introduces that person to other people in my network who that person isn’t connected to but should be. I post an update, and the network figures out who cares to know about it.

Everyone you know is “friend” on Facebook, but why would you ever want to communicate withe everyone you know all at once? Is there ANYTHING that you do equally with EVERYONE you know? Human relationships don’t operate on a single, always on setting.

Some of the metadata we need for the next evolution of online social networking is already part of the social graph — for college students, this captures may different types of relationships:

Facebook Stupid Network

But that metadata is just static reference information — the network doesn’t really let you DO anything with it.

If you want to create a one-size-fits all social network (which is about as useful as a one-size-fits all anything), the least you can do is let me define how I know someone using user-defined tags instead of site administrator defined check boxes.

You know, that Web 2.0 thing.

Comments (7 Responses so far)

  1. Scott, Interesting post. I also think advertisers might like splitting these groups apart better, because it leads to more opportunities for more efficient, more targeted advertising. To pick a very simple example, family-oriented sites/messaging could be good places for advertisers selling Thanksgiving travel and flowers. (Steve Boriss, The Future of News)

  2. a very astute comment and one that will prove to be influential I’m sure. In historical terms we are still at the big bang stage of social media, and I’m sure we’ll look back in a few years with embarrassment that we referred to many complete strangers on very unsophisticated networking sites as ‘friends’.

    Now a confession: on our crowd-sourced interview magazine and social network site we have the facility to add ‘friends’.

    Thanks Scott, time for a rethink on both terminology and a serious analysis of the types of relationships people want to have.

  3. You are on to something with the “primitive” idea, and Internet power users, like your self, are beyond social networks like Facebook. We use numerous tools to communicate, each with their own unique benefits. I run a handful of small private social networks, Google Groups, blogs, and more for all the different segments of my relationships out there, i.e. family, friends, network contacts, etc. Each has its own limitations. There is something to be said for face-to-face interaction between people.

    Have you tried Second Life by chance? Didn’t cut it for me but it is certain to evolve.


  4. Agreed. The current definition of “Friend” on social networking websites is useless. It reflects no reality.

    The “explicit” nature of connections in the social graph is problematic. People are connected through experiences. Not through straight lines.

    The future social graph will adapt with us. It should detect a connection implicitly by monitoring the way we connect with people. There are 100s of indicators that the social graph ignores: our endorsements, communications, events we attend, transactions, information we read, goals we share. All of these clues help us put a connection in context.

  5. Scott,
    Ultimately I think the kind of “useful” social network you’re describing will evolve only once the network itself becomes more trustworthy. But I think that, from the perspective of, do I want such and such a message going to the wrong subset of “friends,” checkboxes seem more reassuring to the casual user (i.e., everyone who doesn’t write about technology for a living) than user-generated tags.

  6. Scott. I think we all need to step back and think about what it really means to be “Friends” with someone. I wrote about this a bit in my post “Friendverse Weakens.” One true “Friend” is worth more than a thousand Facebook friends. I think people will come to realize this eventually – building social networks based on the quality of the connections rather than the quantity of the connections.

    “So perhaps we should stop looking for ways to expand and transport meaningless “Friend” lists and instead focus on building a meaningful and rich Friendverse. People we like, know, and trust.”

  7. [...] This is similar to the assumptions that Facebook makes about how we know people: [...]

Add Your Comment


Receive new posts by email