What I learned from Clay Shirky about science online

The “science online” community has somehow compiled a required reading list (thanks John!), from which many ideas and quotes are mined. I recently finished reading an entry on the list: Here Comes Everybody, by Clay Shirky.

I enjoyed the book – much of it was familiar to me, but it makes good use of specific examples to convey general principles. Of more interest to me is the application of these ideas to science online. Here’s what I think we can extrapolate from the book – and this is purely my personal interpretation.

  1. Most online science communities will fail
  2. Why? Answer: almost everything on the Web fails. However, there is so much activity that a few ideas succeed. Shirky uses the notion of “failure for free”; if you have an idea, it costs little or nothing to throw it out onto the Web and see what happens. In that context, I applaud the efforts of anyone making the effort to build an online science network – but I believe that most of them will find little use and ultimately, fade away.

  3. Networks designed specifically for scientists are unlikely to succeed
  4. One section of Here Comes Everybody focuses on Meetup, a social network through which like-minded people can organise events. One of the most popular groups on Meetup is called “Stay At Home Moms” (SAHM). The point here is that if Meetup had designed its service to cater specifically for SAHMs, it would have failed. What Meetup provides is the technology for people with common interests to find one another; what happens next is up to the users.

    I found this especially relevant with regard to the success of e.g. FriendFeed, a site not designed specifically for science (or any other topic), compared with the limited or zero value that I get from the multitude of “Facebook for scientists” sites.

  5. High barriers to entry are a big problem
  6. A quote from Chapter 10, Failure For Free:

    The number of people who are willing to start something is smaller, much smaller, than than the number of people who are willing to contribute once someone else starts something.

    What this says to me is that you’d better make it easy for people to start something. In my opinion, far too many science networks expect too much effort from users simply to get started. The most work required at the outset should be the creation of a username and password. If you want more from me: employment and educational history, interests, location, publications – then make it painless for me to add that information and make it clear how the effort involved will enhance the utility of the site.

  7. The long tail requires critical mass
  8. The long tail, for those not familiar with the term, is the observation that for any given website, a small proportion of users are the most active, with many more (the “long tail”) contributing less or nothing at all. Sounds somewhat negative? Not at all – it actually explains how websites succeed. Most Wikipedia articles are edited by a small percentage of users, but many more users contribute an occasional edit. Likewise, Amazon is able to sell a wide range of media at bargain prices because a core group of people regularly buy the most popular products, whilst a “long tail” of customers buy an occasional, less-popular item.

    However, for a long tail to form, a large absolute number of potential contributors is required. This is a problem for science online because (1) there are just not many scientists in either absolute or relative terms and (2) an even smaller proportion of those are interested in online networking.

  9. Value has to be apparent from the outset
  10. The question “what can I do at this site?” should be answered with one short, concise and punchy sentence. Here, Shirky uses as an example the announcement on a mailing list of the Linux OS project, by Linus Torvalds, back in 1991. One clear aim: “I’m building a free OS.” No outlandish claims about world-changing goals. Invitation to participation: “I welcome help with my small project and suggestions are welcome.”

    Again, we see this with successful social networks. FriendFeed is one, simple idea: your internet activity in one place. Flickr: share your photos. Twitter: what are you doing right now? Social networks for science, on the other hand, often don’t sell you a simple, appealing concept. I summarise this as: “if you build it they will come – but what do they do when they get there?” Sure, I could write a blog, post some links, join a group, contribute to a forum, create some tags – but as a new user, perhaps unaccustomed to those concepts, I’d be asking myself one simple question: why?

    If you want me to use your site then tell me what it’s for, how it benefits me and what it offers that I can’t find elsewhere. As a user, I’m happy to build content but you should be telling me why I should bother – not the other way around.

9 thoughts on “What I learned from Clay Shirky about science online

  1. Linux Pauling

    Talk about your obscure jokes — I had to look up the fact that Torvalds is named after Linus Pauling to get this.

  2. While Shirky does have his faults (his idea that everyone will stop watching television and instead edit wiki articles for fun is absurd, and his concept of “filter-failure” rather than “information overload” is essentially meaningless), there are some real gems in what you’ve gleaned here.

    To be a bit contrary, I’d argue that the biggest reason sites are failing is that no one has really come up with the “killer app”. It’s not that sites’ value is poorly explained, it’s that there’s nothing going on with most of the social science sites that’s particularly compelling for most scientists. If you can hit on that activity, something that massively shifts things for the better, than most of the details won’t matter. Which calls into question point #2. Is the problem that networks designed for scientists won’t be robust, or is the problem that scientists needs differ from those of other communities. I find too many of the sites try to shoehorn things that work elsewhere into science (“Myspace for scientists”) without really addressing the needs of scientists. Kind of trying to convince scientists that an activity is worthwhile because it can be done, rather than the reverse.

    I’d also argue against the concept of “failure for free”. Do you really think the Nature Network was built and staffed for free? For most of us, our time is valuable as well.

    But the other points are spot on.

  3. Thanks for the comments David; I think we’re largely in agreement. I’d argue that it’s difficult to explain value precisely because a killer app (or more likely, apps) is missing.

    With respect to “failure for free”, I have misused the term slightly here. It should really be used in the sense that setting up an online presence is a low-risk strategy. For example if Nature Network were to “fail”, it would be unlikely to bring down NPG.

  4. Fair enough, but I do think there’s been a huge amount of money and time invested in creating these ventures without much thought on 1) an actual business model, or 2) functionality that’s actually compelling.

  5. This discussion made me think about the arXiv (http://arxiv.org/) which is very successfull in math and physics – much less so in other fields I guess. If they added some social networking features I think it could morph into a sort of killer app for scientists. The compelling functionality is that it has become _the_ place for finding out what other people are doing before their papers are officially published.

Comments are closed.