Special Section

Delivering Information Services through Coolaboration

Michael Schrage and Collaboration

by Steve R. Hardin

MIT research associate and Merrill Lynch Fellow Michael Schrage believes collaboration provides exciting opportunities for research and understanding which could never be realized by individuals working alone. Collaboration, he says, also presents issues which must be addressed by the information science community. A contributing editor to WIRED, Schrage is the author of Shared Minds: The New Technologies of Collaboration (Random House, 1990) and its revised edition, No More Teams (Doubleday Currency, 1995). He discussed Collaborating around Collaboration: Shaping the Future(s) of Information Science at the 1998 Mid-Year Meeting of the American Society for Information Science in Orlando.

He told his audience he has been interested in collaboration since he worked as a reporter at the Washington Post. He wanted to know how great collaborations materialize. There are not as many brilliant individuals as brilliant collaborations. But when you deal with collaboration, you also have to deal with the issues involved in any relationship. For example, how do you reduce friction between two persons?

Schrage quoted 19th century engineer-turned-sociologist Frederic LePlay: "The most important product of the mines is the miner." LePlay, he said, was ahead of his time in understanding human capital issues. It is important to understand there is an ecology between the technologies we have and the way people adapt to them to increase their own value. Schrage modified LePlay's dictum to read: "The most important product of the network is the networker." He is disturbed by how often that truth is ignored. The value added, he said, is not the replication of information, but the type of people who use it. The kind of networks we want to design depends on what kind of networkers we want our people to be. Schrage said this is one of the biggest issues facing information scientists. He drew an analogy: you can have a doctor who treats people as bundles of symptoms and diseases or one who treats people like people. People are information processors, but if you define them as information processors, you have a warped idea of what people really are.

When Schrage was at the Post, he was a chief proponent of the idea that we live in the information age. He now believes that when we talk about the information age, we profoundly misunderstand the impact of the technologies and media in which we work. The "Big Lie" of the information age is

D i = D b
which means that as we change the quantity and quality of people with information, we change the quality of their behavior. It is a rational idea, he said, until you think about it. If information were really the most valuable and important thing, the people who run our organizations would be the smartest. Since that is demonstrably not true, some other variables must be making an impact. He noted that people smoke, even though they know they shouldn't. People drink and drive, even though they know better. We've invested billions of dollars in helping our managers have more information to make better, more informed decisions. But changing behavior is as important as changing minds.

To draw an example from his own life, Schrage said he could stand to lose a few pounds. He can stand on a scale and get digital information on how much he weighs. He can get analog information by looking in a mirror. He also has good information on what he can do to correct the situation: he can exercise more and eat less. But he does not. Information isn't his problem. But suppose his scale is networked, and it sends information about his weight to the five people who most like him and the five who most despise him. Wouldn't that have a greater impact on his behavior than his individual knowledge of how much he weighs? The information aspect, Schrage said, matters far less than the relationship aspect.

One of the most important design shifts, he said, is that we have gone from asking how we structure information to asking how we structure information to affect relationships. We're moving from structuring information to structuring relationships. We need to consider the relationship aspect as carefully as the information side. It's no accident that companies like Yahoo have sought to create a community/relationship infrastructure to complement their information.

We are also moving from "creative individuals" to "creative relationships" as the sources of our information. "The real value of a medium lies less in the information it carries than in the communities it creates," Schrage said. The Internet is as much a medium of community as it is a medium of information retrieval.

Consider Gutenberg in the 15th century: at the time, the new medium was movable type. The Bible, not an almanac, was the first thing published. It was the medium of the community. One of the arguments during the Reformation was about alternatives in Biblical understanding and interpretation opened up by wider distribution of the Bible. What are the key relationships involved here?

Schrage discussed the concepts of automation vs. augmentation. He said we are moving from using technology to automate what we do to using it to augment what we do. It is an important distinction, he said. He showed a slide presenting a box: Augmentation Information Relationships Automation

It is an interesting matrix, Schrage said. The overwhelming majority of research and investments in computer science and management has been done on the left side of the box. Different people have different definitions of these concepts. Since these technologies require us to be explicit to work, we must define these terms more rigorously.

The results of successful collaborations are many: the airplane, quantum physics, the atomic bomb, the double helix and personal computers. (He added he knows Steve Jobs very well; there is no way he could have come up with the personal computer without Steve Wozniak.) Even the Internet was originally a tool to help physicists collaborate.

What do these accomplishments have in common? One attribute is better communication among the collaborators than persons in other relationships. This concept is where most organizations and individuals run aground, Schrage said. Collaboration is not communication. It is easy, he said, to fall into the trap of thinking we could have better collaboration if only we could increase the bandwidth. Collaboration requires communication, but it is not communication. Collaboration is not about buy-in, either. Another confusion is participation with collaboration. Participation is a means, not an end. It is needed for collaboration, but it is not the same thing, he said.

So, then, what is collaboration? Schrage said it is the process of shared creation or shared discovery that individuals realize they could not have done on their own. Watson and Crick both said they could not have come up with the double helix working alone. They are geniuses individually, all right, but the problem was bigger than either of them could solve working in isolation, Schrage said. The value comes from the interaction.

Technology has allowed more people with more information to interact more. There are huge problems of scale, Schrage said. Going back to the matrix, he stated scaling information is relatively easy, but scaling relationships is unbelievably hard. You can broadcast, but shut up and listen is not really collaboration. By the same token, scaling automation is relatively easy. But scaling augmentation is astonishingly hard, he said. The challenge is to go from "dialogues" to "kilologues" to "megalogues." What does a megalogue look like? Schrage admitted he does not know. How can an organization manage thousands of dialogs in a meaningful way? He said this question is one of the most important faced by information scientists.

With no exceptions, Schrage said, the key element of all successful collaborations is shared space. We have all had the experience of sketching on a napkin together. The collaboration depended on the napkin. "It takes shared space to create shared understandings," he said. And how does technology fit into collaboration?" The properties of the shared space shape the quality of the collaboration," he said. Chalkboards have limits. Computers represent many more possibilities. We can scale shared space by making the shared space smarter, he asserted.

He stated we already have an abundance of bandwidth and processing power. Constraints are a wonderful way of structuring ideas, Schrage said. What are the best constraints? For example, why not differentiate between e-mail that asks the receiver to act, and e-mail that merely passes on information? What sort of filters, agents and knowbots make the most sense? How do you moderate and facilitate interactions? These are design dimensions critical for information scientists.

Consider the "bozo filter." Schrage said participants in The Well discussion groups noticed some people made stupid contributions. So the operators gave people the ability to filter out contributions from certain people. What does the existence of bozo filters in an organization say about the values of the organization? Why not give the organization's network the ability to create shared space(s)? Suppose you are running a library facility, and you notice people from Germany are doing searches similar to those being done by people from Taiwan. Should your organization let them know that? There are privacy issues, but Schrage said there is a simple solution: just ask people whether they want the fact that they are making particular searches known publicly. And if you are in a Fortune 100 company in which 80% of persons do not want others to know about their searches, what does that say about the company?

The creation and management of shared spaces, Schrage said, will become the dominant design issue as network technology proliferates. What kind of communities will our constraints help create? There is the question of anonymity vs. attribution. In collaboration, people attribute contributions to their originators. One company that doesn't attribute has discovered it gets fewer contributions. Then there is the issue of specifications vs. prototypes. The traditional organization does a requirement analysis and then tries to build a prototype to fit the specs. Schrage said he has had success building the prototypes first and eliciting specs from them. Innovated prototypes generate innovative teams, contrary to conventional wisdom that innovated teams generate innovative prototypes.

Designing the incentive to collaborate, he said, is just as important as designing the technology for collaboration. It is ludicrous for companies to say "we want to build collaboration, but we're going to reward only individual performance." But they do. If you are not rewarded for collaboration, what is the incentive to collaborate?

In closing, Schrage noted it wasn't Watson or Crick who won the Nobel Prize; Watson and Crick won it together. It is important to develop collaboration to the benefit of ourselves and our communities.

In response to a question, Schrage acknowledged collaboration is not necessary for every task in an organization. But organizations still must decide whether collaboration has any value to them. He added that in research and academe, collaboration is usually very important. Asked "when does collaboration make sense?" Schrage stated he believes more fields are becoming collaborative. He'd like to see anonymous peer review eliminated; people should put their names on their comments, he said. Multidisciplinary peer review in grant applications is needed, too, he said. We should not suffer from hardening of the arteries. Maybe we should have a "megalogue" review process, with hundreds of people on the Web reading a proposal.


Steve R. Hardin is associate librarian at the Cunningham Memorial Library at Indiana State University. He can be reached by phone at 812/237-7685 or by e-mail at libhard@cml.indstate.edu

Bulletin of the American Society for Information Science