Imagine: How Creativity Works

by Jonah Lehrer

  In order to understand the wisdom of Pixar’s office design, it helps to know about the research of Tom Allen, a professor of organization studies at MIT. In the early seventies, Allen began studying the interaction of engineers in several large corporate laboratories. After several years of tracking their conversations — counting all those exchanges in the hallways and coffee room — he came up with the Allen curve, which describes the likelihood that any two people in the same office will communicate. The curve is steep; according to Allen, a person is ten times more likely to communicate with a colleague who sits at a neighboring desk than with someone who sits more than fifty meters away.

  It’s not particularly surprising, of course, that we make small talk with those who are nearby. But Allen also discovered something unexpected about all these office conversations. After analyzing the workplace data, he realized that the highest-performing employees — those with the most useful new ideas — were the ones who consistently engaged in the most interactions. “High performers consulted with anywhere from four to nine organizational colleagues [on a given project], whereas low performers contacted one or two colleagues at most,” Allen wrote in his 1984 treatise Managing the Flow of Technology. “This suggests that increasing the number of colleagues with whom an employee consults contributes independently to performance.” The key word in that sentence is independently. According to Allen’s data, office conversations are so powerful that simply increasing their quantity can dramatically increase creative production; people have more new ideas when they talk with more people. This suggests that the most important place in every office is not the boardroom, or the lab, or the library. It’s the coffee machine.

  A similar lesson emerges from a recent study led by Brian Uzzi, the sociologist who measured the Q of Broadway musicals. In 2009, Uzzi got access to a vast trove of data from a large hedge fund, giving him a complete record of every instant message sent by every trader over an eighteen-month period. The first thing Uzzi and his collaborators discovered was that these traders sent out an astonishing number of messages. They amassed more than two million exchanges, with the average trader engaging in sixteen different IM conversations at the same time.

  What Uzzi wanted to know was how this constant stream of information affected the financial performance of the traders. He was particularly interested in the flurry of messages that occurred whenever a new financial report was released. “What you often see is [that] some new information comes over the Bloomberg terminal, and there’s this sudden spike in communication,” Uzzi says. “What’s happening is that everyone is trying to figure out what the news means. Is it good news? What’s it going to do to the stock?” Uzzi refers to this as the disambiguation process, since the traders are trying to make sense of the unclear information. They’re asking one another questions and benefiting from the diverse thoughts of colleagues. “There’s a big incentive to figure this stuff out fast,” he says. “The faster you are, the more money you make.”

  By comparing the messaging habits of the traders, Uzzi was able to document the power of these electronic interactions. He discovered that Tom Allen was right: the best traders were the most connected, and people who carried on more IM conversations and sent more messages also made more money. (While typical traders generated profits on only 55 percent of their trades, those who were extremely plugged in profited on more than 70 percent of their stock trades.) “These are the guys who get embedded in multiple chats,” Uzzi says. “They’re just sucking up information from everybody else, like a vacuum. And when they start to trade, they don’t go silent. They don’t stop talking. They IM even more.” In contrast, the least successful traders tended to engage in the fewest electronic chats. They also stopped exchanging messages right before making an investment decision. Says Uzzi, “They’d get cut off from the conversation. And this meant that they weren’t able to make sense of what was happening.”

  Uzzi compares these financial conversations to the creative process. “The act of investing is like solving a difficult puzzle,” he says. “These traders are trying to connect the dots. And if you look at these IM exchanges [in the hedge fund], what you frequently find is that they lead to a good idea, a successful trade. Because the traders are listening to their network, they manage to accomplish what they could never have done by themselves.” While all the instant messages might seem like a distraction — a classic example of multitasking run amok — Uzzi argues that they’re an essential element of success. “If I had to choose between a trader who was a little smarter or one who was a little better connected, I’d definitely go for the connections,” Uzzi says. “Those conversations count for a lot.”

  Pixar tries to maximize such conversations. The studio knows that an office in which everyone is interacting is the most effective at generating new ideas, as people chat at the bathroom sink and exchange theories while waiting in line for lattes. “The secret of Pixar from the start has been its emphasis on teamwork, this belief that you can learn a lot from your coworkers,” says Alvy Ray Smith. “Ed and I were really determined to create a kind of mutual admiration society, so that the techies thought the artists were geniuses, and the artists thought the techies were magicians. We wanted people to want to learn from each other. That’s always when the best stuff happens: when someone tells you something you didn’t already know.”

  The computer scientist Christopher Langton once observed that innovative systems constantly veer toward the “edge of chaos,” to those environments that are neither fully predictable nor fully anarchic. We need structure or everything falls apart. (The classic demonstration of a space without structure is the “nonterritorial office” developed by the ad agency TBWA Chiat/Day. In 1993, the company decided to do away with every tradition of the corporate office. Employees were no longer given fixed desks or cubicles or computers. Instead, they were encouraged to assemble with their colleagues based on the task at hand; the model was the college campus, in which students were free to work anywhere and everywhere. (Time magazine hailed the Chiat/Day office as the “ forerunner of employment in the information age.”) The reality of the new space, however, failed to live up to the hype. Although the freeform interior was designed to encourage interaction, it actually derailed it. Because there were no offices, people couldn’t find one another. Productivity plummeted. The couches with the nicest views became the subject of petty turf wars; fistfights broke out over meeting spaces. By 1995, it had become clear that the new Chiat/Day model was broken. The walls were reinstalled.) But we also need spaces that surprise us. Because it is the exchanges we don’t expect, with the people we just met, that will change the way we think about everything.

  3.

  Every day at the Pixar studio begins the same way: A few dozen animators and computer scientists gather in a small screening room filled with comfy velour couches. They eat Lucky Charms and Cap’n Crunch and drink organic coffee. Then the team begins analyzing the few seconds of film produced the day before, ruthlessly shredding each frame. (There are twenty-four frames per second.) No detail is too small to tear apart: I sat in on a meeting in which the Toy Story 3 team spent thirty minutes discussing the reflective properties of the plastic lights underneath the wings of Buzz Lightyear. After that, an editor criticized the precise starting point of a Randy Newman song. The music began when Woody entered the scene, but he argued that it should start a few seconds later, when Woody began running. Someone else disagreed, and a lively debate ensued. Both alternatives were tested. (It’s not uncommon for a Pixar scene to go through more than three hundred iterations.) The team discussed the motivations of the character and the emotional connotations of the clarinet solo. By the time the meeting was over, it was almost lunch.

  These crit sessions are modeled on the early production meetings at Lucasfilm, when Alvy Ray Smith, John Lasseter, and Ed Catmull would meet with the animators to review their work. At first, the meetings were necessary because nobody knew what he was doing — computer anima
tion remained a hypothetical. But Lasseter soon realized that the meetings were incredibly efficient, since everybody was able to learn from the mistakes of everybody else. Furthermore, the crit sessions distributed responsibility across the entire group, so that the entire team felt responsible for catching mistakes. “This was a lesson I took away from the Toyota manufacturing process,” Catmull says. “In their car factories, everybody had a duty to find errors. Even the lowly guys on the assembly line could pull the red cord and stop the line if they saw a problem. It wasn’t just the job of the guys in charge. It was a group process. And so what happened at Toyota was a massive amount of incremental improvement. People on the assembly line constantly suggested lots of little fixes, and all those little fixes had a way of adding up to a quality product. That model was very influential for me as we set about figuring out how to structure the Pixar meetings. I wanted people to know that if a mistake slips through the production process, if we don’t fix something that can be fixed, then it’s everybody’s fault. We all screwed up. We all failed to pull the red cord.”

  The harsh atmosphere of Pixar’s morning meetings — the emphasis on finding imperfections and mistakes — may at first seem to contradict one of the basic rules of group creativity, which is to always be positive. In the late 1940s, Alex Osborn, a founding partner of the advertising firm BBDO, came up with a catchy term for what he considered the ideal form of group creativity: brainstorming. In a series of bestselling books, Osborn outlined the basic principles of a successful brainstorming session, which he said could double the creative output of a group. The most important principle, he said, was the absence of criticism. According to Osborn, if people were worried about negative feedback, if they were concerned that their new ideas might get ridiculed by the group, then the brainstorming process would fail. “Creativity is so delicate a flower that praise tends to make it bloom, while discouragement often nips it in the bud,” Osborn wrote in Your Creative Power. “In order to increase our imaginative potential, we should focus only on quantity. Quality will come later.”

  Brainstorming is the most popular creativity technique of all time. It’s used in advertising offices and design firms, the classroom and the boardroom. When people want to extract the best ideas from a group, they obey Osborn’s instructions; criticism is censored, and the most “freewheeling” associations are encouraged. The underlying assumption is simple: if people are scared of saying the wrong things, they’ll end up saying nothing at all.

  There is, of course, something very appealing about brainstorming. It’s always nice to be saturated in positive feedback, which is why most participants leave a brainstorming session proud of their contributions to the group. The whiteboard has been filled with free associations, the output of the unchained imagination. At such moments, brainstorming can seem like an ideal mental technique, a feel-good way to boost productivity.

  There’s just one problem with brainstorming: it doesn’t work. Keith Sawyer, a psychologist at Washington University, summarizes the science: “Decades of research have consistently shown that brainstorming groups think of far fewer ideas than the same number of people who work alone and later pool their ideas.” In fact, the very first empirical test of Osborn’s technique, which was performed at Yale in 1958, soundly refuted the premise. The experiment was simple: Forty-eight male undergraduates were divided into twelve groups and given a series of creative puzzles. The groups were instructed to carefully follow Osborn’s brainstorming guidelines. As a control sample, forty-eight students working by themselves were each given the same puzzles. The results were a sobering refutation of brainstorming. Not only did the solo students come up with twice as many solutions as the brainstorming groups but their solutions were deemed more “feasible” and “effective” by a panel of judges. In other words, brainstorming didn’t unleash the potential of the group. Instead, the technique suppressed it, making each individual less creative.

  The reason brainstorming is so ineffective returns us to the importance of criticism and debate, the very elements that define the Pixar morning meeting. (Steve Jobs has implemented a similar approach at Apple. Jonathan Ives, the lead designer at the company, describes the tenor of group meetings as “brutally critical.”) The only way to maximize group creativity — to make the whole more than the sum of its parts — is to encourage a freewheeling discussion of mistakes. In part, this is because the acceptance of error reduces its cost. When you believe that your flaws will be quickly corrected by the group, you’re less worried about perfecting your contribution, which leads to a more candid conversation. We can only get it right when we talk about what we got wrong.

  Consider this clever study led by Charlan Nemeth, a psychologist at UC-Berkeley. She divided 265 female undergraduates into five-person teams. Every team was given the same difficult problem: How can traffic congestion be reduced in the San Francisco Bay Area? The teams had twenty minutes to invent as many solutions as possible.

  At this point, each of the teams was randomly assigned to one of three different conditions. In the minimal condition, the teams received no further instructions; they were free to work together however they wanted. In the brainstorming condition, the teams got the standard brainstorming guidelines, which emphasized the importance of refraining from criticism. Finally, there was the debate condition, in which the teams were given the following instructions:

  “Most research and advice suggest that the best way to come up with good solutions is to come up with many solutions. Freewheeling is welcome; don’t be afraid to say anything that comes to mind. However, in addition, most studies suggest that you should debate and even criticize each other’s ideas.”

  Which teams did the best? The results weren’t even close: while the brainstorming groups slightly outperformed the groups given no instructions, people in the debate condition were far more creative. On average, they generated nearly 25 percent more ideas. The most telling part of the study, however, came after the groups had been disbanded. That’s when researchers asked each of the subjects if he or she had any more ideas about traffic that had been triggered by the earlier conversation. While people in the minimal and brainstorming conditions produced, on average, two additional ideas, those in the debate condition produced more than seven. Nemeth summarizes her results: “While the instruction ‘Do not criticize’ is often cited as the [most] important instruction in brainstorming, this appears to be a counterproductive strategy. Our findings show that debate and criticism do not inhibit ideas but, rather, stimulate them relative to every other condition.”

  There is something counterintuitive about this research. We naturally assume, like Osborn, that negative feedback stifles the sensitive imagination. But it turns out we’re tougher than we thought. The imagination is not meek — it doesn’t wilt in the face of conflict. Instead, it is drawn out, pulled from its usual hiding place.

  According to Nemeth, the reason criticism leads to more new ideas is that it encourages us to fully engage with the work of others. We think about their concepts because we want to improve them; it’s the imperfection that leads us to really listen. (And isn’t that the point of a group? If we’re not here to make one another better, then why are we here? Just look at the Beatles: Lennon and McCartney had a famously combative and competitive relationship. But that turned out to be blessing in disguise, since all the internal disagreements inspired the songwriters.) In contrast, when everybody is “right” — when all new ideas are equally useful, as in a brainstorming session — we stay within ourselves. There is no incentive to think about someone else’s thoughts or embrace unfamiliar possibilities. And so the problem remains impossible. The absence of criticism has kept us all in the same place. (The emotion of anger also seems to have short-term creative benefits. That, at least, is the take-away message of a 2011 series of studies led by Matthijs Baas, Carsten De Dreu, and Bernard Nijstad. In their first experiment, they demonstrated that anger was better than a neutral mood for promoting creativity. I
n their second experiment, they elicited anger directly in some of the subjects, and then asked all of the study participants to brainstorm on ways to improve the environment. Once again, people who felt angry generated more ideas than nonangry people. These ideas were also deemed more original, as they were thought of by less than 1 percent of the subjects.) Of course, this doesn’t mean that anger is a cure-all or that nastiness is always wise. For one thing, anger is resource depleting: although angry subjects generated more ideas initially, their performance quickly declined.

  To better understand the power of criticism — why it acts like a multiplier for the imagination — it’s worth looking at another experiment led by Nemeth. While the typical brainstorming session begins with an instruction to free-associate — to express the very first thoughts that enter the mind — that’s probably an ineffective strategy. In study after study, psychologists have found that people just aren’t very good at free-associating. For instance, if I ask you to free-associate on the word blue, there’s a 45 percent chance that your first answer will be sky. Your next answer will probably be ocean, followed by green, and, if you’re feeling creative, a noun like jeans. Our associations are shaped by language, and language is full of clichés.

  How do we escape these clichés? Nemeth found a simple fix. Her experiment went like this: A lab assistant surreptitiously sat in on a group of subjects being shown a variety of color slides. The subjects were asked to identify each of the colors. Most of the slides were obvious, and the group quickly settled into a tedious routine. However, in some groups, Nemeth instructed her lab assistant to occasionally shout out the wrong answer, so that a red slide would trigger a response of “Pink,” or a blue slide would lead to a reply of “Turquoise.” After a few minutes, the group was asked to free-associate on these same colors. The results were impressive: people in the dissent condition — they were exposed to inaccurate descriptions — came up with far more original and varied associations. Instead of saying that blue reminded them of sky, they were able to expand their loom of associations, so that the color triggered thoughts of Miles Davis, Smurfs, and berry pie. The obvious answer had stopped being the only answer. More recently, Nemeth has found that the same strategy can lead to improved problem solving on a variety of creative tasks. It doesn’t matter if you’re trying to invent a new brand name or decipher a hard insight puzzle. Beginning a group session with a moment of dissent — even when the dissent is wrong — can dramatically expand creative potential.