Business / research

Jam at a massive scale

February 18, 2012

We spoke with Frank Barrett of the Naval Postgraduate School to learn what jazz can teach us about collaboration. Barrett is the author of Yes to the Mess: Surprising Leadership Lessons from Jazz.

Andrew Benedict-Nelson: In your work on jazz improvisation and organizational dynamics, you argue that group leaders can foster something called “provocative competence.” The example you give is the way that Miles Davis set up the “Kind of Blue” sessions, giving his musicians difficult tasks they had never seen before but trusting that they were capable of accomplishing them. You also argue that tasks that are beyond a single individual’s ability to complete can play the same role, drawing out the best in us. How do you think one can cultivate provocative competence in a group where there is no “band leader” or manager?

Frank Barrett: When I talk about this now, “provocative competence” is actually the last point I make rather than the first. The first point I talk about is some form of “unlearning.” That’s implicit in provocative competence. You have to have a mindset that is willing to disrupt and change in general. That’s the starting point.

Most groups that have history develop routines that take on a life of their own. So you should think about how much history your group has. If a group has a long life, then routines that were originally functional may become dysfunctional.

Andrew Benedict-Nelson: It’s interesting to think of power within a group residing not so much with a “president” who directs it, but some sort of “secretary” who understands that history and monitors it.

Frank Barrett: Keeping history is good, but routines are usually implicit, not written down. That’s what makes them useful, what makes them efficient, but also makes them potentially onerous. A routine is not going to be something that someone can easily document. It’s more of a tacit awareness. There are some explicit routines, but it’s powerful, tacit routines that I’m referring to.

Andrew Benedict-Nelson: Let’s say I’m a member of a group and I want to disrupt one of those routines or encourage provocative competence. What should I do?

Frank Barrett: My sense is that the dynamics are the same whether you’re a manager or a member. The dynamics are around having connections and linkages. If you just lead with disruption, you’ll be seen as obnoxious.

One of the examples I use sometimes is Steve Nash of the Phoenix Suns basketball team. If you ever listen to the way he describes his job, it’s like listening to a mother describing how to take care of four or five kids. There’s a tacit sense of what the mood is of the players, who’s having self-esteem  issues, who should be pushed or pulled a little bit, who needs to be passed to more often. He may decide that he’s not going to throw to a guy because he’s being too egotistical, and he needs him to be a little bit more hungry. Those are all provocative moves when you don’t have an official leadership role.

But you need to have those connections with people so you’re not seen as an outsider. If you try provocative competence as an outsider, people will kill you.

Andrew Benedict-Nelson: It feels like there’s almost a “political capital” sense to it. If you challenge a group to do something really difficult without a history with them, you might be taken as challenging the group’s legitimacy. They might think you are setting them up for failure. But when you have that history of trust, they know that you think they can pull off what you’re challenging them to do.

I’m also curious about what you call “the competency trap” – the problem where you build an organization to do one thing really well and then you wake up one day and you’re irrelevant.

Frank Barrett: Sure, you don’t even know what has happened. You wake up one day and the environment has completely changed.

Andrew Benedict-Nelson: Right. So I’m curious about what this might mean given the way the Internet has been evolving. If you compare the Internet of today versus the 90s, people are spending a lot more time on a small number of sites that have dominated certain functions, like search and social. I think there might be a sort of environmental “competency trap” as people become really good at very specific types of interaction. So how should you react to the “competency trap” when you’re a participant in a system rather than its designer?

Frank Barrett: Well, first I think you have to think about whether there is real human contact in an environment like the Internet. When I think of my work on improvisation and collaboration, what I think about in particular is responsiveness. Are you alive and alert and able to respond in the moment? That’s what good musicians do, it’s what Steve Nash does, it’s what good collaborators do. It’s what good executives do if they don’t get caught in that “competency trap” and stop paying attention.

Are you alive and alert and able to respond in the moment? That’s what good musicians do, and it’s what good collaborators do.

I question how much real responsiveness can happen through information technology. You can mainly exchange explicit information and data. It may be that social media or Skype are closer to a  real connection, but I wonder.

Andrew Benedict-Nelson: Sure, on the Internet everything is very intentional. You don’t pick up on a lot of those subconscious cues unless someone is sensitive enough to consciously put them out there. That would seem to hinder the kind of deep collaboration you describe in your research.

Frank Barrett: That’s right.

Andrew Benedict-Nelson: So perhaps online collaboration could be improved if we could capture more of this intuitive-level interaction. Well, you’ve found some ways for people to improve their abilities to pick up on those sorts of cues in real life. Do you think we could adapt some of the things you’ve learned by studying those situations to enhance that intuitive level of collaboration online?

Frank Barrett: But there’s a contradiction even in the question. You say we can’t “capture” this intuitive interaction online. Well, we don’t “capture” it in real life either – that’s what makes it intuitive. You sense it. I’m not sure you could even articulate why it makes sense to make a particular intervention in the group when you do, why it makes sense to challenge Ken or Bob or Joe. You just feel it. So it doesn’t lend itself well to codification.

Maybe there’s technology out there I don’t know about yet. I think eye contact helps a lot, so it may be there with things like Skype. Having a face to see and respond to changes the game. When you can hear a voice and pick up on the tone, that helps. But it’s still not the same.

Online technology lends itself well to things like brainstorming – that’s why crowdsourcing has become so popular and successful. You just need ideas to be thrown out one at a time and you’re trying to get as many ideas as possible. But that’s not really collaboration – it’s just more input.

Andrew Benedict-Nelson: It’s good to come to a collaboration armed with all that, but it’s not a collaboration itself.

Frank Barrett: What makes it collaboration is if you can argue and go back and forth and entertain and disagree and challenge. Can that happen online? My guess is that it can if it’s part of a larger historical relationship.

When I co-author a paper with a friend of mine on the East Coast, we work exclusively with text. But we know each other so well – I know the way she thinks, the kind of things she might add. That’s a sort of sequential interdependence – Person A does some work, then hands it off to B, who then does some work and hands it off to A. It’s not reciprocal. It’s not in real time like Steve Nash or Miles Davis. But it does lead to new insight because it’s in the context of a historical relationship.

Without that relationship, there would be hurdles and obstacles. You would ask “What the hell did he mean by this?” or “Why did she put this here? This has nothing to do with this part.”

Andrew Benedict-Nelson: That makes me think of another aspect of jazz improvisation you describe in your work. You say that musicians build improvisations around minimal structures that maximize flexibility. In jazz, the examples are things like common standards or time signatures. I think that’s also important to a task like an academic paper – in addition to your shared relationship, you share an idea of how a paper is supposed to look and feel.

S0 it seems to me that the massive collaborations of the Internet are also built around these minimal structures. But I wonder if people are falsely identifying those structures with technology rather than the social category. What do you think? What do you think is the “state of play” for these minimal structures?

Frank Barrett: I think of these minimal structures as something like a pulse. They’re already there, but it may be that no one has drawn attention to them. The question is whether you can deliberately design for them.

One thing that you can design for is regular updating when people are working independently. IT can help with that. When Kodak was designing the camera that became the Funsaver, they had to do it quickly. Several teams were working in their own environments but they came together each day to see what everyone had done.

You can imagine that sort of thing with software developers now, but the project would follow the sun. You might start a project in California, then someone in Singapore would pick it up, then someone in Paris would pick it up. But there has to be some sort of regular updating of everyone.

Another important aspect of the minimal structure is that it’s impersonal. You can’t have a situation where people start asking, “Why do we always have to do what you want to do?” You shouldn’t have to waste energy negotiating it.

If you just lead with disruption, you’ll be seen as obnoxious.

Minimal structures also shouldn’t be overly constraining. I’m a civilian who works for the military. In crisis situations, you see that there is often a minimal structure with maximum autonomy. People know just a few things that are not negotiable, that they have to do. But sometimes people learn the standard operation procedures so thoroughly that they think that you have to have a standard operating procedure for every situation. Then the situation changes, and they freeze or don’t act or wait.

So the question is what designers of systems can identify as the minimal structures that are necessary so everything else can be flexible or up for grabs.

Andrew Benedict-Nelson: You know, I think that when I asked the question I made a bad assumption. I assumed that the minimal structures get designed at the beginning, and then all these things happen around them. But you seem to be suggesting a more emergent model for minimal structures.

Frank Barrett: That’s right. I think it’s both. You can build some in advance. For example, in disaster situations you have the model of the incident command system. That’s a minimal structure designed in advance. The first responder sets up the command and then people start to coordinate around it. It doesn’t mean the first responder is necessarily telling them what to do. But yes, others are tacit.

There is also a minimal structure built into every group: the structures of time and punctuated equilibrium. Whatever the length of the life of the group is, at the halfway point some sort of convergence usually happens. Let’s say the group is supposed to work together for two years. At the one-year mark, there is usually a perceptible shift in the organization. There is an increased tension and convergence. Halfway through that second half, there is usually another convergence. That’s sort of the natural rhythm of the life of a group. It’s a minimal structure that’s tacit.

Some people think that groups operate according to “stages of development” – things like “forming, norming, storming”  – you’ve heard all those. It’s supposed to be a linear pattern. But that has never been demonstrated with empirical evidence. No one has ever shown that that has actually happened.

But we do have evidence for these stages of punctuated equilibrium. There are paradigm shifts at these mid-points and opportunities for the group to change course. In some ways, it’s an invitation – it could get better or it could get worse. But in functional groups, members know that the group is going to end, and something shifts. For example, it may be that someone says explicitly something that no one has yet talked about until this point. It may be that a new idea comes to a surface.

Andrew Benedict-Nelson: It’s funny, my greatest experience of the sort of thing you’re talking about is actually working in an Insight Lab, and it’s for just the reason you say – everyone knows when the project is going to end. There’s no chance that someone will suggest we meet again at a later date. And you’re right, the result is a series of recognizable pivots, though I don’t know if they always occur at the exact midpoints.

Frank Barrett: Some laboratory models suggest that there is a natural pivot point that is roughly halfway through. People upgrade their learning, get focused, get serious, because time is running out. You’d have to look back at your own experience to see if that’s intuitively true.

Andrew Benedict-Nelson: I think it’s halfway through as long as you’re willing to assume a somewhat relativistic model of time. It usually feels halfway.

Frank Barrett: You have to go back and see what the appropriate measure is for each group. It could be halfway through the number of days the group has existed or halfway through the number of face-hours they have had together.

In one interesting study, they set up several groups to do some work, and in some groups they set the clock wrong. In some groups it was faster and in some way it was slower. And sure enough, the group had a transition point at the halfway point according to the clock, even though it was biologically false. But the group also had another shift that was biologically true and had nothing to do with the clock time. They had an innate sense that half of their energy had been spent and they needed to get something done.

Andrew Benedict-Nelson: So how do you think you could take advantage of these principles on a larger scale? If I’m meeting in a room with 20 people, I might have a good sense of how excited people are, how long they can go before they get tired – I can design around that. I might have a similar sense if you’re talking about an entire company. But it seems really hard when the collaboration you’re talking about is massive and global. You also have the problem that with collaborations like Wikipedia, it’s really difficult to define when the project ends. You wouldn’t know when you’re halfway done. So how do you design in that kind of potentially infinite environment?

Frank Barrett: I think you would have to ask whether Wikipedia really falls into this category of group collaboration. It certainly does not have reciprocal interdependence. It’s also timeless and boundary-less like you said.

Online technology lends itself well to brainstorming. But that’s not collaboration — it’s just more input.

Andrew Benedict-Nelson: I wonder if you just need to identify subgroups. Maybe you could identify the punctuated equilibrium pattern within the history of the Wikipedia page about, say, Miles Davis.

Frank Barrett: This is interesting. Now I’m making the link. There was a study that came out of Harvard on Wikipedia and the innovations that people make in the history of a page. There is something related to what we’re talking about in terms of time. They found that when a new topic gets put up, there’s lots and lots of activity, and then it trails off to just a few people. Over time, there are just a few watchdogs who keep an eye on it and make tweaks. So in a way, it does become a finished project.

Andrew Benedict-Nelson: That’s interesting, because the endpoint is defined just as much by the life-cycle of the group as much as any objective, external criterion.

So I’d like to talk about some of the other attributes of collaboration in jazz improvisation raised by your research. You mention that jazz musicians embrace errors as a source of learning. You discuss both simple errors where people know they’ve screwed up and move on, but also the more interesting situation where one musician will make a mistake and another will use it as the basis for a new improvisation.

I think that’s a neat idea, but I worry about whether we’ll continue to be able to take advantage of it. Both on the Internet and in our culture, we seem to be obsessed with the “gaffe.” We record everything so perfectly that it’s hard to escape from the errors of the past. If you have this perfect record, how do you build a culture of forgiveness?

Frank Barrett: Well, in a way the “gaffe” has already been compensated for by society. When you play the error over and over again, people get used to it. I mean, why is Newt Gingrich still around? Why is Bill Clinton still around? Eventually it just becomes part of the background. It isn’t purposeful forgiveness, but it might be purposeful forgetting.

That doesn’t really relate to the issue of collaboration, though.

Andrew Benedict-Nelson: I’m still concerned with this visibility problem. Let’s go back to the metaphor of the Wikipedia page, where in the page history you can often see all the discussions that went into the form of the current page. Do you think any group can really make a decision if they know all of that will be recorded for all the world to see?

Frank Barrett: It’s like watching sausage get made. And yet it works. It does converge somehow.

Andrew Benedict-Nelson: My fear is that it works for Wikipedia, but it doesn’t work for Congress. Who knows if there was ever effective group collaboration there, but you get the sense that every vote and every decision is now so much under the spotlight that they can’t go away and actually negotiate anything.

Frank Barrett: When negotiations are too public, people show up as stakeholders and they hold on to their stakes. In the process of negotiation, you have to be willing to be changed. One of the reasons Congress is so broken now is that people will do things like signing a pledge right after they’re elected in which they promise never to do something, like raise taxes. So by the time you get into negotiation with another member of Congress, they are no longer a work in progress. They’re not there to learn. They’re just there to assert, to change others but not to be changed.

Andrew Benedict-Nelson: Do you think there are versions of that that we encounter in collaborations that are less dysfunctional than Congress?

Frank Barrett: It happens a lot in organizations. There are stories like when a hospital gets a big grant to build a new building and different departments like surgery or anesthesia or research all want it. They may also show up intending to change others and not to change themselves. They caucus with their colleagues before going in to the collaboration. They get their ducks in a row and say, “Make sure you don’t lose this! Make sure you don’t do that!” I’d say it happens wherever there are multi-party collaborations where participants represent group interests.

Andrew Benedict-Nelson: I’m curious about the cognitive aspect of this too. It feels like because of the information environment we live in today, you often hear about not only the positions that groups actually took in  a negotiation, but the various ones they were considering. You hear stories like, “Obama was ready to give away such and such an amount.” That seems like a barrier not just to collaboration, but to even thinking about collaboration.

Frank Barrett: That information can also be part of a signaling strategy. You can let people know that you are considering giving up A, B, and C, when you already know you will have to give up all of that plus D, E, F, G, H, and I. You can use those strategies to make yourself look a certain way in a negotiation.

Andrew Benedict-Nelson: So to get back to the errors problem, it seems to me that the common problem we have here is about information. We have a culture where information is becoming more and more visible. But what’s strange is that it’s getting harder to actually use that information for actual collaboration. Instead it’s being misspent trying to game the negotiation or the collaboration.

Frank Barrett: It’s being hijacked, in a way.

Andrew Benedict-Nelson: But we can’t help living in an information culture. So what do we do to make sure information actually informs the collaboration instead of being hijacked for things like signaling?

Frank Barrett: It’s hard to talk at such a generic level – part of my answer would depend on the task. What kind of group interdependence do you require? Are we working together toward one big goal? Or are we interdependent in a way where we all have different stakes? I’m not sure you can actually call that collaboration.

Andrew Benedict-Nelson: I’d like to ask you about another aspect of improvisation you discuss in your work: retrospective sense-making. The point is that an improvisation is not something that is simply planned then executed. There may be a strategy, but musicians often have to think about what they just did (or, eventually, listen to a past recording) to figure out what they were actually up to.

Story and narrative are powerful things. They are resources that should be listed along with other kinds of capital.

So, I get that we’re all doing this kind of retrospective sense-making all the time. But it seems like it’s really pronounced in our culture right now. If you think about the three biggest social movements of the past few years, for example – the Tea Party, the Occupy movement, and the Arab Spring – they all seemed to start with smaller events that other people later made into a larger phenomenon. When Mohamed Bouazizi self-immolated in Tunisia, for example, he didn’t necessarily intend to start the Arab Spring. Do you think it’s always been like this?

Frank Barrett: I think that’s the way things have always worked. There’s no way to predict where something is going to go. If you’re, you hold an event here, you hold an event there, you send out e-mails every two weeks regardless. You plan actions, but you don’t know what the outcome will be, especially if you’re talking about a second-order change like the Arab Spring. You don’t know the meaning of the event until people start telling stories, and it probably didn’t mean that at the time it happened.

What it means for practical purposes is that story and narrative are incredibly powerful things. They are resources that should be listed along with all other kinds of capital. They can be used to direct energy or direct attention, to kill energy or kill attention.

Andrew Benedict-Nelson: I wonder if what has changed is the way that we relate to the historical events in the narrative. I go back to this problem of all text or all video in an online collaboration being recorded. I think that in many collaborations people remember stories about something someone said or did that aren’t quite true, but they’re true at the level of the greater story or narrative the group was trying to produce. But if we perfectly record all the facts, it might be harder to move from the level of fact to story. Do you think that’s a good thing or a bad thing?

Frank Barrett: It’s a neutral thing. Just because you can go back to a recording and say, “Oh look, he didn’t actually say that,” it doesn’t mean you’ve gotten any closer to nailing down a clear meaning to what was said. It just means you have privileged access to the event. It’s just one more input. It still needs an interpretive move.

Andrew Benedict-Nelson: That’s interesting, because if you have access to the facts in that way and they don’t help, it makes it much clearer that interpretation is a separate task.

So let’s talk about being “in the groove.” In a paper you wrote on this subject, you describe this feeling among jazz musicians that they are engaged in a task that transcends their individual self. Do you think that can be achieved in larger groups?

Frank Barrett: I don’t know. This isn’t an empirical answer, but I think if it were a group of 200 or something, I could imagine it would be possible if the groups broke down around sub-tasks. I’ve heard a story of a community that built its own playground in one day. They had some help from professionals, but mainly they had those 200 volunteers coming together throughout the day. I think you could argue that there were in a sort of groove, but if it could be transcendent, I don’t know. To be “in the groove” you need immediate feedback and access to the whole. It depends on what “the whole” looks like. But I could see it.

Andrew Benedict-Nelson: Another quality of improvisation you’ve investigated is communities of practice – things like having jam sessions with other musicians or just hanging out with them. When you look at our culture, what do you think is the state of hanging out?

Frank Barrett: Well, in a way it’s happening all the time. The value of hanging out is learning what it means to be an “insider.” You learn how to behave like a person does when connected to a certain type of activity. You go from being a neophyte to a specialist. There’s identity change – you go from being a marginal outsider to someone who is accepted in that identity category. That’s what happens in inner-city gangs; as you get closer to the center, you realize how you’re supposed to dress, how you’re supposed to talk, which stories are worth paying attention to. But it also can happen constructively in organizational life.

Andrew Benedict-Nelson: I get that you can’t imagine being human without some degree of hanging out. But it seems to me that in a lot of these more digital relationships, it’s much more difficult than it used to be. I don’t just mean people working online. I’m also thinking of people who collaborate at work and then all go home to three different suburbs. It makes it harder for people to learn those social norms.

Frank Barrett: You’re learning norms – you’re always learning norms – but the norms you’re learning may be what kind of computer screen to get, what sites you’re supposed to look at. We also learn norms by reading. Don Quixote learned how to be a knight by learning a lot of chivalry books before he went out and tried it. But it’s not the same as going out and doing it.

Andrew Benedict-Nelson: Something that’s been talked about a lot lately is the importance of “weak ties.” In the social media world, it’s those people who aren’t really your friends but who you are still connected with – you might interact every once in a while. And then you might be able to activate them years later when you’re looking for a job. What effect do you think that has on hanging out?

Any task where there is interdependence naturally lends itself to soloing and supporting.

Frank Barrett: There is a wide distribution of information about people now, and many places where you could link up with them that you couldn’t before. I’m sure you could learn social norms that way, but it seems very light.

Andrew Benedict-Nelson: I think it occurs, but maybe not in the way we’ve been talking about it. What interests me is when I have a really intense, face-to-face collaborative experience with someone, and then we connect on social media. Once every few weeks they might “like” a photo I post or comment on something. It’s not very intense, but the relationship is continuing in a way it might not have before, and I can learn a lot from the relationship if I want to.

Frank Barrett: I don’t know. I know it should be possible that that is hanging out, but I can’t quite see it.

Andrew Benedict-Nelson: You’ve written about the importance of alternating between soloing and accompaniment in a group. I think the interesting question here again is scale. So you might be talking in a room with 15 people, but it’s not as if you’re doing a solo and the others are silent. They are often doing an accompaniment – they may nod at what you’re saying or ask a question. How do you achieve the same kind of accompaniment with groups of thousands or tens of thousands?

Frank Barrett: Again, it’s all dependent on the task. What kind of task could thousands of people do together?

Andrew Benedict-Nelson: Or another way of putting it might be, wouldn’t that size of task inevitably break down into smaller groups and sub-tasks?

Frank Barrett: Right. It would have to be modular. The key is that you would have to be able to help the other person think well. Your contribution would be to complete their contribution, not just to be a solo author. That happens all the time on Wikipedia – someone sees something developing it, adds to it, and the original author says, “Oh, I never realized my idea could go in that direction.”

Andrew Benedict-Nelson: I was trying to think about soloing and accompaniment in some massive non-digital task, and the first thing that came to mind was the State of the Union address. Because of the structure of the chamber where that speech is given, it’s easy for the speaker to rapidly assess the reactions of many people. Though I’m not sure it rises to the level of accompaniment.

Frank Barrett: No, probably not, because that’s not a common task. Where I thought you were going was talking about the writing of the State of the Union, which usually involves about twelve people. Obama might take a speechwriter’s line and head in a new direction with it. They’re actually completing each other’s thinking. They make the whole greater. But that’s a small group, not thousands.

Andrew Benedict-Nelson: There’s an interesting contrast between the State of the Union and Wikipedia examples, though. Let’s say you’re JFK and I’m Ted Sorenson. You may have an idea for what you’ll say in the speech, I’ll modify it, and you’ll say, “Yes, that’s what I actually meant to say.” What’s interesting to me about a situation like Wikipedia is that the same thing can happen, but the editor can be a complete stranger that you’ll never hear from again. The relationship doesn’t need to have the history that JFK and Ted Sorenson had. What do you think that means?

Frank Barrett: It works because of the common task. Any task where there is interdependence naturally lends itself to soloing and supporting.

Andrew Benedict-Nelson: I’m imagining a jam session of blind-folded musicians.

Frank Barrett: That would be an interesting experiment.

Andrew Benedict-Nelson: To me, one of the interesting things about your work is that you’re recognizing things that we all already do as social creatures, but that we couldn’t necessarily recognize or articulate. At our upcoming Lab, we’re going to be talking about how we could design better massive global collaborations. I’d like the participants to do some mental work that is similar to yours. I’d like them to think about the greatest collaborative experiences they have already had, online or offline. What do you think are the questions they should ask themselves in order to better understand those experiences?

Frank Barrett: Sure. What made this possible? What led to this? What were the antecedents? What was it about the structure of the interaction that made it work? Where did you have old routines that you could rely on? Where did you have to abandon routines and try something you never tried before? Where did you notice someone else’s strength? Where did you notice someone’s weakness, and how did you continue in spite of it – how were you not overthrown by it? When did you notice that something was missing? Stuff like that.

Andrew Benedict-Nelson: When you talk with people about experiences like this, what are the aspects of collaboration that people are the worst at noticing, even when they’re thinking about their own memories and experience?

Frank Barrett: I would imagine that if you were to ask the players of the Phoenix Suns, I do not think they would have noticed Steve Nash and the way he was looking in their eyes to measure how much self-confidence they had. I think they would not have noticed his efforts toward provocative competence, the gentle nudges. I know the players on “Kind of Blue” knew what Miles Davis was doing. But I don’t if they noticed how much he was noticing their potential.

Andrew Benedict-Nelson: That’s fascinating. What you can’t possibly notice is the other person’s assessment of what you can do. It’s in their mind. And then when you pull off the task, it just feels like you did it.

Frank Barrett: When a group comes up with a creative idea, there’s a self-serving bias to notice how much your contribution made a difference. You may not notice how others’ indirect actions – tacit looks, clever questions – led to breakthroughs.

Andrew Benedict-Nelson: It also strikes me that that is the quality that’s most difficult to design for.

Frank Barrett: That’s right.

Andrew Benedict-Nelson: So later this month we’ll be putting together this group to design for better massive collaborations. You’ve been working on this topic for a long time. What are the fallacies or obvious conclusions you would really want us to avoid?

Frank Barrett: One that’s tired and overrated is “We have to build trust.” Trust isn’t always necessary, especially if the task is designed well.

I also think it’s important that your group know the skills and talents they have available to begin with. I think people often run to a task without taking stock.

When Don Shula was coaching the Miami Dolphins, he would use those videos players watch after a game to improve their performance next time. But he would mix up the groups. So the quarterbacks and the linebackers would watch the defense play. They would say things like, “Oh, I never realized how well they did that before.” I think that’s a good thing to do in groups, to see what you might not have seen before.