by Clay Shirky
The White Bicycle program would have been just another footnote in the Age of Aquarius, but for one detail: it was an almost instant failure. Within a month all the bicycles had either been stolen or thrown in the canals. Undaunted, many urban visionaries have resurrected the basic idea of the While Bicycles in one way or another. The cumulative results of these experiments are unambiguous: programs that offer unrestricted access to communal bicycles have struggled with theft, and most have ended up collapsing completely, while the communal bike programs that have succeeded have placed restrictions on the use of the bicycles with things like locked sheds and ID cards for checking them in or out. Despite the Provos’ optimism, human nature has turned out to be fairly context sensitive; given the opportunity to misbehave, and little penalty for doing so, enough people’s behavior becomes antisocial enough to wreck things for everyone. (If only the Los Angeles Times had understood this before launching the Wikitorial project.)
The fate of the various White Bicycle programs illustrates, unintentionally but dramatically, a basic truth of social systems: no effort at creating group value can be successful without some form of governance. The groups now adopting social tools form the experimental wing of political philosophy, a place where hard questions of group governance are being worked out. One remarkable aspect of harnessing social value is that social groups tend to be homeostatic, which is to say resistant to external pressures. The classic illustration of individual homeostasis is body temperature. Humans have an internal temperature of 98.6°F, whether they are in the Sahara or in the Arctic. A group, once formed, can achieve homeostasis as well, finding ways to stay together even as the external environment changes.
Pierre Omidyar, a co-founder of eBay, credits the success of his business to trust in the users; he has often said that one of his founding assumptions was that people are basically good. The reality is more complex: eBay may have been founded on a basic trust in human goodness, but within a couple of months after it launched, enough of the transactions were going awry in one way or another that the company had to respond. EBay’s solution was to create a reputation system, allowing the buyer and seller in any transaction to publicly report their satisfaction with each other. This system was designed to cast the shadow of the future over both parties, giving each an incentive to maintain or improve their standing on the site; with that addition, eBay became the site we know today. Omidyar was right, with a caveat: people are basically good, when they are in circumstances that reward goodness while restraining impulses to defect. The rewards and restraints can be quite simple and small, but in big groups with relatively anonymous actors, they need to be there or behavior will decay over time.
If a group can last a year it has a good chance of lasting quite a bit longer. The fans of Buffy the Vampire Slayer who inhabited the Bronze:Beta bulletin board have remained active long after that series was canceled; Buffy provided the rationale for them to come together, but the value of the community itself became a reason to stay together. Similarly, the Howard Forums community for avid mobile phone users includes a set of off-topic discussions, because the users had been introduced around the topic of mobile phones, but they decided that they liked one another well enough to want to talk about sports, pets, and diets. In extreme cases of homeostasis, the original rationale disappears entirely but the group remains intact. The S-100 Computer Users Group, a Silicon Valley group founded in the late 1970s for users of an early type of personal computer, was still meeting as a social group under that name in the early 1990s, despite the fact that all of the members had long since abandoned their S-100s. The group survived the loss of the founding rationale because they liked one another’s company.
This is not to say that all groups become purely social—there is always a tension between the pleasure of community and the rationale for the group’s existence (between satisfaction and effectiveness, in other words), and that tension works itself out in different ways. One of my students, Marion Misilim, observed a LiveJournal group in 2002 called “I Love My Boyfriend,” populated by girls celebrating the emotion encapsulated in the group’s title. During the time Misilim was observing the group, a crisis erupted when one of the group’s members had discovered that her beloved had been cheating on her. This discovery had the predictable effect on the relationship, which left the hard question: Now that she no longer loved her boyfriend, could she still be part of the group? Various palliative rationales were offered by the others—she still loved him, even if she despised his actions, or she used to love him and that was enough. These justifications were meant to hold out the possibility of their friend remaining in the group, but to no avail—the wronged girl now loathed her ex-boyfriend, and that was that. The founding promise had not decayed for everyone, as with the S-100 group; it had simply failed for her, and so she left the group.
This is also not to say that groups don’t eventually fail. The two longest-lived groups on the internet, the SF-LOVERS and WINE-LOVERS mailing lists, both founded in the early 1970s, each lasted three decades but eventually petered out. The corollary of homeostasis, though, is that most failures happen quickly, usually because of a failure to get one of the big things right.
Many such sites fail the very first test—they offer no plausible promise. After the failure of the L.A. Times Wikitorial project, observers discussed whether the problem was with the use of a wiki or with the misbehavior of the users. The answer is, none of the above. The problem with Wikitorial was that the basic offer—“Come help improve the L.A. Times’ editorials!”—was of little interest in the first place. An editorial is not the sort of content that benefits from group editing, and the promise of working on behalf of the Times, a distinctly uncommunal entity, held little appeal except to users who were minded to take advantage of a potential soapbox. Since there was no community that wanted to defend the Wikitorial, the experiment collapsed. Michael Kinsley, then an editor and one of the proponents of the Wikitorial, correctly blamed the failure on the misbehaving users but incorrectly suggested that this condition was unusual. Any experiment of any importance will always have people who want it to fail. Only the systems that have defenses against such users can thrive; the assumption that those users won’t appear is an inadequate defense.
If promise were enough, the normal case would be the success, rather than the failure, of social tools. Sometimes the promise is entirely appealing, but the tools are inappropriate. One such example is MoveOn.org, the liberal political organization famous for using the Web to gather support. MoveOn was founded to convince Congress to censure (rather than impeach) President Clinton for lying about his relationship with Monica Lewinsky, and move on to other business; later causes included advocating campaign finance reform and supporting John Kerry’s 2004 run for the U.S. presidency. When MoveOn wants to mobilize support, it can e-mail nearly a million users, who in turn send e-mail to Congress.
This hardly sounds like the stuff of failure, but in fact e-mail is the wrong tool for lobbying Congress. Before e-mail, a rule of thumb in Congress was that one handwritten letter from a constituent indicated that something like two thousand voters in that district cared about the same issue. E-mail enormously lowers the transaction cost of sending a message while creating superdistribution, the effortless forwarding of the message from person to person and group to group. The problem with e-mail as a tool is that it is now too good—the cost of lobbying Congress by e-mail is so low that an e-mail message has become effectively meaningless. Attempts by Congress to reintroduce some value to the communications—by asking e-mail correspondents to include their mailing address, to make sure that they are a congressperson’s constituents—have failed because users can easily cut and paste addresses from the congressperson’s district, whether they live there or not. One of the reasons e-mail campaigns continue, despite their near uselessness, is as a public show of force. The individual communications have been denatured, so the battle has moved to public claims of how many mails were sent, whi
ch play out in the court of public opinion, not in the halls of Congress. MoveOn, and every other organization that lobbies Congress, would be better served by a less convenient, more expensive tool, one that took real effort to use and so communicated real commitment on the part of the people writing in.
This is happening now. Fans of the TV show Jericho were so upset when CBS canceled the show that they started mailing peanuts to CBS in protest using the NutsOnline delivery service. This effort cost the fans real money, so there was no mistaking their commitment, especially not when twenty tons of peanuts eventually arrived at CBS. (CBS relented and revived the show.) Similarly, antiwar protesters in Michigan and immigrants upset at changes in the handling of U.S. visa rules are protesting by mailing flowers to a Michigan representative and the head of U.S. Immigration Services respectively. Flowers have the dual advantage of signifying respect and being hand-delivered; it’s much harder to ignore flowers than e-mails. All these protests have what e-mail lacks, which is proof that the protesters are willing to express their opinion, even at some expense and difficulty.
A similar problem with a change in costs beset the Howard Dean campaign. The first time Dean appeared in the national consciousness was when three hundred people showed up for a Howard Dean Meetup in New York City in early 2003. This level of attendance was unprecedented, and Dean himself took note of it, coming down from Vermont to speak to his supporters. It seemed like a predictor of great success, but the size of the Dean Meetup was as much a testament to Meetup as to Dean. People were right to be excited about the Dean Meetup but wrong about the reason, because Meetup was founded to lower the coordination costs of real-world gatherings. Prior to Meetup, a turnout of three hundred people would have indicated the existence of a huge and latent population of Dean supporters; as with letters to Congress, one individual turning out would have suggested much broader support for Dean. However, because Meetup makes it easier to gather the faithful, it confused people into thinking that we were seeing an increase in Dean support, rather than a decrease in transaction costs—the 2003 Dean Meetup simply brought out a much larger percentage of Dean supporters than would have shown up previously. (We’ve seen this sort of effect before, as when written correspondence on letterhead stopped being a sign of a solvent company, thanks to the desktop-publishing revolution.)
Finally, both the promise and the tool can be effective, but the bargain kills the deal. Even Wikipedia’s harshest critics concede its popularity and the value in having users report errors. When the managers at Encarta, Microsoft’s digital encyclopedia, saw the excitement around Wikipedia, they offered Encarta users the ability to perform a similar service for Encarta. The resulting problem was with neither the promise nor the tool: Wikipedia had shown that people are more than willing to contribute to online reference works, and that the tools are available to do so at low cost and large scale. What doomed the Encarta’s effort to minor status was its bargain with users: users had to grant Microsoft permission to “use, copy, distribute, transmit, publicly display, publicly perform, reproduce, edit, modify, translate and reformat your Submission” for a product Microsoft was going to charge money for. This was hardly a bargain at all, as all the power lay with Microsoft, a fact that made the option for user contribution largely irrelevant. Encarta talked the same basic talk, but the details of the bargain, including its tone, undermined the possibility of any large-scale recruitment.
Even when a group offers an obvious promise and provides a stable service, the bargain continues to evolve. In 2004 Flickr was acquired by Yahoo, and the following year Flickr asked its users to switch to using a username and password that were general to Yahoo rather than specific to the Flickr service. It was a technical change, occasioning a minor inconvenience, or so the people running Yahoo thought. Instead, a small but very vocal part of the Flickr population went very publicly berserk, castigating Flickr and threatening to move to another photo service. Little came of the threats; Flickr has millions of users, so it would not have missed the loss of even hundreds, but the fight was never really about the details. What the most vociferous users objected to was the incontrovertible evidence that they were not in control. Although they had contributed the photos that make Flickr what it is, they were not in a position to say no to a unilateral change from Yahoo. The fact that the change was relatively minor didn’t matter, because even a minor change exposed the users’ relative powerlessness. The incident passed without creating a serious challenge to Flickr, but the amount of public agita is indicative of how seriously users take the implicit bargain, even when (and perhaps especially when) it is not explicitly supported by contract.
A similar revolt from users beset Digg, the user-edited news site, when its staff members found themselves negotiating a bargain with their users that they didn’t even know they’d made. The revolt concerned DVDs. All DVDs rely on a secret digital key that is designed to prevent users from copying the contents, and in early 2007 that key was uncovered. People protesting digital restrictions on DVDs began posting the secret number to Digg, and Digg, responding to a request by a DVD industry group, began removing the posts that included it. Digg was not merely within its rights, it was in fact required by law to do so, but the Digg users didn’t care. Thousands of them flooded the site with posts containing the key, or instructions on how to find it by searching for “09 F9,” and thousands more wrote emails to Kevin Rose, Digg’s founder. Their point of view, expressed on a spectrum from polite to livid, was simple: Digg was built on user participation. The users both suggested and rated the stories that appeared on the front page, and in this case what they wanted on the front page was the DVD key. Digg’s owners, who thought they were merely complying with a law, came to realize that the users were not casually posting the key—they were engaging in an act of civil disobedience, with Digg as their chosen platform. Faced with exerting unilateral control over their users or living up to their end of the bargain, Digg relented, allowing unlimited posting of the key. As Kevin Rose said, in announcing the new policy:
[A]fter seeing hundreds of stories and reading thousands of comments, you’ve made it clear. You’d rather see Digg go down fighting than bow down to a bigger company. We hear you, and effective immediately we won’t delete stories or comments containing the code and will deal with whatever the consequences might be. If we lose, then what the hell, at least we died trying.
“At least we died trying.” This may be the high-water mark of a commercial firm being held to account by its users. What Rose recognized, and to his credit acted on, was that his business was built not on the software that ran Digg but on the implicit bargain that his users assumed they had with Digg and, by extension, with him. This bargain had nothing to do with the official rules of the site or indeed the legal requirements—the users were intentionally violating both. The bargain was implicit but deeply felt; had Digg’s management reneged, the damage to the site’s popularity could have been considerable. Once Rose recognized this fact, he took the remarkable step of allowing his company to be a site of collective action by his users, collective action that could have destroyed his business. (Digg survived.)
The Digg revolt is one example of where our social tools are going. Starting with the invention of e-mail, which first functioned to support a conversation in a group, our social tools have been increasingly giving groups the power to coalesce and act in political arenas. We are seeing these tools progress from coordination into governance, as groups gain enough power and support to be able to demand that they be deferred to. The Digg revolt was one of the broadest examples of this intersection between groups and governance; it will not be the last.
EPILOGUE
At 2:30 in the afternoon on May 12, 2008, an earthquake hit China’s Sichuan province. Hundreds of buildings collapsed, including many schools, leaving nearly 70,000 dead and another 20,000 missing, 350,000 wounded, and 5,000,000 homeless. Word of the quake spread both instantly and globally via social media. The first line of repo
rting was from Sichuan residents themselves, with messages appearing on QQ (China’s largest social network) and on Twitter, even as the ground was still shaking. Within minutes, photos and videos of the quake’s effects were being uploaded from mobile phones, with the links being further passed around by e-mail, instant messages, and text messages. The quake was being discussed on QQ and Twitter before it was on any news site; Rory Cellan-Jones of the BBC reported learning of the quake from Twitter. The Wikipedia page for the quake was created within forty minutes to host the now-customary response of sharing links to information about the disaster and its aftermath. Within hours, sites designed to aid the search for missing friends and relatives began popping up, and by the next day, donations from all over the world were being raised on behalf of the survivors.
The speed with which the world became aware of the quake was a function not just of global technological networks, but of its social ones. China and the United States are connected by undersea communications cables, but mere technical connectivity would not have been enough to carry news of the quake as quickly as it did. There is also something we might liken to a social cable running from China to the United States, an invisible bundle of connections between people on both continents. This bundle is made up of all the bonds between the two populations that have built up over the years: the graduate students from China who studied in the United States and returned home, the branch offices of U.S. firms doing business in China; in fact every bit of human contact that makes people want to stay in touch, even when they live far away.
