With a song in his heart, and a statistical model of network self-assembly mechanisms, Brian Uzzi reveals how the Broadway musical can teach important lessons to business leaders By Matt Golosinski
10/1/2005 - Manhattan is home to both Broadway and Wall Street. And while geographically close, the worlds of premier theatrical productions and high-flying financial transactions couldn't, on the surface, seem farther apart.
But art, science and business converge in Brian Uzzi's latest research on creative enterprises and the collaborative networks that form in fields such as social psychology, economics, astronomy and professional musical theater.
If the Kellogg School professor of management and organizations is offering a kind of song-and-dance routine, rest assured that it's not the same old song and dance. In fact, anyone with a serious interest in understanding the dynamics behind building winning teams should pay attention to the findings published this year in his paper, “Team Assembly Mechanisms Determine Collaborative Network Structure and Team Performance” (co-authored with Northwestern University chemistry professors Roger Guimerá and Luís A. Nunes Amaral and Stanford Graduate School of Business student Jarrett Spiro). The research appeared in the April 29 edition of Science and is forthcoming in the December issue of the American Journal of Sociology (with Jarrett Spiro).
This investigation uncovered how creative teams, such as those involved in Broadway musicals and scientific innovation, arise and evolve to have the optimal number of experienced (or “incumbent”) players and newcomers. The researchers examined both scholarly and artistic teams. Publication in top peer-reviewed journals served as a criterion for assessing the academic teams.
In the case of the Broadway musical, Uzzi and his colleagues studied the interactions of key figures such as directors, choreographers and librettists, but not actors. The researchers scrutinized old playbills dating from 1877 to 1990 (2,258 productions in all) looking for successful collaborations to understand the networks responsible for strong performances as measured in terms of artistic impact and financial success.
In addition, the researchers made some inferences about diversity's role in the team assembly process and, further, what individual team makeup implies about the overall creative network.
“We developed a model by which if you know how people assemble their local teams, you are then able to estimate what the larger, systemic-level network structure looks like,” says Uzzi, adding that this fact can be important for analysts, investors or artists seeking to estimate the best arena to focus their energies or investments because different systemic level networks partly determine how likely it is that breakthrough innovation will emerge from the network independent of the talent of individuals within the network.
Since the systemic level is difficult to measure accurately, Uzzi says that understanding the local level can, by implication, suggest qualities about the overall landscape.
What Uzzi and his co-authors discovered by examining the local network is that success came more readily when a mix of incumbents and newcomers collaborated. While incumbents could be expected to have some previous connection with one another — and certainly enjoyed relatively greater prominence within the industry network, serving as network “hubs” connecting many individuals to one another — Uzzi's investigations suggest that having too many incumbents repeatedly working with the same people can lead to substandard results, since this kind of homogeneity can inhibit fresh thinking.
More eclectic teams can offer a creative jolt, but Uzzi also cautions against facile notions of diversity.
“This is an important point: Gender, race and ethnicity are proxies for the kind of diversity we're talking about — diversity of background, training, experience, the way people make inferences,” he says. “We're trying to get at the underlying issue of diversity that is important when structuring groups. You could have a team composed of an African-American, an Asian-American, a white American, men, women, etc., but if they all have similar training and experiences, you're not getting real diversity.”
Part of the challenge of establishing real diversity, he says, probably stems from social-psychological reasons — most people enjoy feeling comfortable around familiar faces. But such teams usually under-perform.
For these reasons, Uzzi says, teams in corporate America are often deliberately dissembled and rearranged with regularly — every five years, say, in the case of some R&D groups. But disruptive social events can also scramble teams, resulting in a creative churning that, ultimately, benefits the participants and industry.
“If you're on Broadway and trying to put a show together, there are serious economic pressures for you to work with people you've previously worked with, because the financiers see short cuts they believe will lower their costs, minimize risk and get them closer to the product they expect,” explains Uzzi. So the system is being pushed toward the creation of homogeneous project teams that, as a consequence of their past experience, come to think like one another. “The only time we saw the market change this behavior was when there is an external shock like the Great Depression, World War II, the advent of Rock ‘n' Roll or other factors that shook up the tendency to build teams of people who come to know each other well.”
For example, rock music was one of the most damaging things to hit Broadway because “so much talent flew out the door to find their fame in the new arena,” Uzzi says.
With each disruption, newcomers have greater opportunity to enter a field, bringing with them new ideas as they mix with the incumbents.
Uzzi's interest in musical theater has both personal and scholarly components. Growing up in New York City , he was a fan of Broadway, and has fond memories of attending shows there with his family. (“Hair” had a transformative effect on him as a kid…though a very different effect on his parents who didn't expect the now-famous “nudity” scene at the end of the second act). Equally important, variables associated with the musical as an experimental subject are more easily controlled than other creative enterprises, such as Hollywood film productions involving hundreds of people.
“There's a lot of noise in the Hollywood data. With Broadway, you may have only six people working on a musical,” Uzzi says. “It's kind of like studying the Galapagos Island 's finches — a simple case that helps us to understand more complex phenomena.”
The researchers note that the optimal team size for Broadway musicals evolved to number about seven people by 1930 — a figure that has remained relatively stable since. This appears to be an arrangement large enough to enable specialization and labor division, but “small enough to avoid overwhelming costs of group coordination.”
The curtain isn't likely to go down on Uzzi's pursuit of the theater or creative networks, particularly since he is also involved with the ambitious American Musical Theatre Project ( amtp.northwestern.edu ) launched this year by Northwestern University . The project will have several elements to it, he notes, including serving as an incubator to develop original musicals and sell them as intellectual property. He is also working with Professor Daniel Diermeier, in the Kellogg MEDS Department, and Stefan Wuchty, a physicist and post-doctoral fellow at Northwestern's new center on complex systems (NICO), on how the networks of movie goers (whether they go alone, as a couple, with friends, or family members) affects why and how fast some movies become weekend blockbusters, while others flop.
“I always loved Broadway theatre; it's one of the great American exports to the world,” says Uzzi. “I grew up wanting to do something related to this industry, and now I'm doing that.”