by Daniel Coyle
Allen wasn’t a typical ivory tower academic; he was a middle-class kid from New Jersey who’d graduated from tiny Upsala College, then enlisted in the Marines during the Korean War. When he got out, he worked for Boeing, then went to MIT for dual graduate degrees in computer science and management, which left him perfectly positioned to pursue the government’s request for research. (“I didn’t even know they had a management degree when I got [to MIT],” he says. “I took a few classes, liked it, and some people talked me into getting a PhD.”) Allen started his research by locating what he called “twin projects,” where two or more engineering firms tackled the same complex challenge, such as figuring out how to guide an intercontinental ballistic missile or communicate with a satellite. He measured the quality of their solutions, then attempted to find the factors that successful projects had in common.
One pattern was immediately apparent: The most successful projects were those driven by sets of individuals who formed what Allen called “clusters of high communicators.” The chemistry and cohesion within these clusters resembled that between Larry Page and Jeff Dean at Google. They had a knack for navigating complex problems with dazzling speed. Allen dug into the data to find out where the people in these clusters got their knack. Had they written for the same journals? Did they possess the same levels of intelligence? Were they the same age? Had they attended the same undergraduate schools or achieved the same level of degrees? Did they possess the most experience or the best leadership skills? All these factors would seem to make sense, but Allen could find none that played a meaningful role in cohesion. Except for one.
The distance between their desks.
At first he didn’t believe it. Group chemistry is such a complex and mysterious process that he wanted the reason for it to be similarly complex and mysterious. But the more he explored the data, the clearer the answer became. What mattered most in creating a successful team had less to do with intelligence and experience and more to do with where the desks happened to be located.
“Something as simple as visual contact is very, very important, more important than you might think,” Allen says. “If you can see the other person or even the area where they work, you’re reminded of them, and that brings a whole bunch of effects.”
Allen decided to dig deeper, measuring frequency of interactions against distance. “We could look at how often people communicated and see where they were located in relation to each other,” he says. “We could see, just through the frequency, without knowing where they sat, who was on each floor. We were really surprised at how rapidly it decayed” when they moved to a different floor. “It turns out that vertical separation is a very serious thing. If you’re on a different floor in some organizations, you may as well be in a different country.”
When Allen plotted the frequency of interaction against distance, he ended up with a line that resembled a steep hill. It was nearly vertical at the top and flat at the bottom. It became known as the Allen Curve.*1
The key characteristic of the Allen Curve is the sudden steepness that happens at the eight-meter mark. At distances of less than eight meters, communication frequency rises off the charts. If our brains operated logically, we might expect the frequency and distance to change at a constant rate, producing a straight line. But as Allen shows, our brains do not operate logically. Certain proximities trigger huge changes in frequency of communication. Increase the distance to 50 meters, and communication ceases, as if a tap has been shut off. Decrease distance to 6 meters, and communication frequency skyrockets. In other words, proximity functions as a kind of connective drug. Get close, and our tendency to connect lights up.
As scientists have pointed out, the Allen Curve follows evolutionary logic. For the vast majority of human history, sustained proximity has been an indicator of belonging—after all, we don’t get consistently close to someone unless it’s mutually safe. Studies show that digital communications also obey the Allen Curve; we’re far more likely to text, email, and interact virtually with people who are physically close. (One study found that workers who shared a location emailed one another four times as often as workers who did not, and as a result they completed their projects 32 percent faster.)
All of which gives us a lens to understand what Tony Hsieh is up to. He is leveraging the Allen Curve. His projects tend to succeed for the same reason the creative cluster projects succeeded: Closeness helps create efficiencies of connection. The people in his orbit behave as if they were under the influence of some kind of drug because, in fact, they are.
During our conversations, I ask Hsieh how he goes about recruiting new people into the Downtown Project. “If someone is interested, and we’re interested in them, we invite them out here,” he says. “We sort of do it in a sneaky way. We give them a place to stay for free and don’t tell them too much. They get here and they hang out and see what’s happening, and some of them decide to join. Things just sort of happen.”
What percentage end up moving here?
He pauses for a long time. “Probably about one in twenty.” At first, this number doesn’t seem all that impressive—only 5 percent. Then you think about what’s beneath that number. One hundred strangers will visit Hsieh, and after a few conversations and a handful of interactions, five will uproot themselves from their home and join this group they have just met. Hsieh has built a machine that transforms strangers into a tribe.
“It’s funny how it happens,” Hsieh says. “I never say very much; I don’t make any big pitch. I just let them experience this place and wait for the moment to be right. Then I look at them and ask, ‘So when are you moving to Vegas?’ ” He smiles. “And then some of them do.”*2
* * *
*1 The Allen Curve echoes another famous social metric, the Dunbar Number, which reflects the cognitive limit to the number of people with whom we can have a stable social relationship (around 150). They would seem to underline the same truth: Our social brains are built to focus and respond to a relatively small number of people located within a finite distance of us. One hundred and fifty feet also happens to be the rough distance at which we can no longer recognize a face with the naked eye.
*2 Shortly after my reporting was completed, Downtown Project leaders embarked on a controversial series of belt-tightening moves, which resulted in the layoff of thirty staffers and Hsieh’s pulling back from his leadership role. It remains to be seen whether this experiment can succeed in the long run.
Building safety isn’t the kind of skill you can learn in a robotic, paint-by-numbers sort of way. It’s a fluid, improvisational skill—sort of like learning to pass a soccer ball to a teammate during a game. It requires you to recognize patterns, react quickly, and deliver the right signal at the right time. And like any skill, it comes with a learning curve.
This learning curve applies even to the scientists who study belonging. For example, Will Felps, who did the bad apple study (see Chapter 1), described how insights from his research affected the way he communicated in his personal life. “I used to like to try to make a lot of small clever remarks in conversation, trying to be funny, sometimes in a cutting way,” he says. “Now I see how negatively those signals can impact the group. So I try to show that I’m listening. When they’re talking, I’m looking at their face, nodding, saying ‘What do you mean by that,’ ‘Could you tell me more about this,’ or asking their opinions about what we should do, drawing people out.”
Amy Edmondson (whom we also met in Chapter 1) has studied psychological safety in a wide variety of workplaces. “I used to not think about whether I was making people safe at all,” she says. “Now I think about it all the time, especially at the beginning of any interaction, and then I constantly check, especially if there’s any change or tension. I bend over backward to make sure people are safe.”
Felps and Edmondson are speaking to the same truth: Creating safety is ab
out dialing in to small, subtle moments and delivering targeted signals at key points. The goal of this chapter is to provide a few tips on doing that.
Overcommunicate Your Listening: When I visited the successful cultures, I kept seeing the same expression on the faces of listeners. It looked like this: head tilted slightly forward, eyes unblinking, and eyebrows arched up. Their bodies were still, and they leaned toward the speaker with intent. The only sound they made was a steady stream of affirmations—yes, uh-huh, gotcha—that encouraged the speaker to keep going, to give them more. “Posture and expression are incredibly important,” said Ben Waber, a former PhD student of Alex Pentland’s who founded Humanyze, a social analytics consulting firm. “It’s the way we prove that we’re in sync with someone.”
Relatedly, it’s important to avoid interruptions. The smoothness of turn taking, as we’ve seen, is a powerful indicator of cohesive group performance. Interruptions shatter the smooth interactions at the core of belonging. They are so discohesive, in fact, that Waber uses interruption metrics as sales training tools. “When you can show someone numbers that the top salespeople hardly ever interrupt people, and then rate them on that scale, you can deliver a powerful message,” he says. Of course, not all interruptions are negative: Creative sessions, for example, often contain bursts of interruptions. The key is to draw a distinction between interruptions born of mutual excitement and those rooted in lack of awareness and connection.
Spotlight Your Fallibility Early On—Especially If You’re a Leader: In any interaction, we have a natural tendency to try to hide our weaknesses and appear competent. If you want to create safety, this is exactly the wrong move. Instead, you should open up, show you make mistakes, and invite input with simple phrases like “This is just my two cents.” “Of course, I could be wrong here.” “What am I missing?” “What do you think?”
R. C. Buford, general manager of the San Antonio Spurs, is one of the most successful executives in the history of sports. But if you watch him operate, you might mistake him for an assistant. He’s a quiet, affable hound-dog Kansan who asks questions, listens keenly, and radiates humility. Early in our conversations, he brought up the looming retirements of several star players and said, “I’m absolutely terrified of the future.” He could have talked about the organization’s vaunted player selection and development systems, or the progress of the young players, or the smart trades they’d made, or the power of the culture they’d built. But he didn’t do that—he said he was terrified. This kind of signal is not just an admission of weakness; it’s also an invitation to create a deeper connection, because it sparks a response in the listener: How can I help?
“To create safety, leaders need to actively invite input,” Edmondson says. “It’s really hard for people to raise their hand and say, ‘I have something tentative to say.’ And it’s equally hard for people not to answer a genuine question from a leader who asks for their opinion or their help.”
Embrace the Messenger: One of the most vital moments for creating safety is when a group shares bad news or gives tough feedback. In these moments, it’s important not simply to tolerate the difficult news but to embrace it. “You know the phrase ‘Don’t shoot the messenger’?” Edmondson says. “In fact, it’s not enough to not shoot them. You have to hug the messenger and let them know how much you need that feedback. That way you can be sure that they feel safe enough to tell you the truth next time.”*1
Preview Future Connection: One habit I saw in successful groups was that of sneak-previewing future relationships, making small but telling connections between now and a vision of the future. The St. Louis Cardinals baseball team, for example, is renowned for their culture and their ability to develop young players into big-league talent. The Johnson City (Tennessee) Cardinals are St. Louis’s lowest-level minor-league club. One day on a bus belonging to the Tennessee team, one of the Cardinals coaches, sitting in the front row, gestured up toward the television on which the big-league team was playing.
“You know that pitcher?”
Players looked up. On the screen, wearing a perfect white uniform, stood the heroic figure of Trevor Rosenthal, a young star who had become a dominant relief pitcher for the Cardinals; he had pitched in the previous year’s World Series.
“Three years ago,” the coach said, “he was sitting right in that seat where you are.”
That’s all he said. It wasn’t much—it took about five seconds to deliver. But it was powerful, because it connected the dots between where the players were and where they were headed. Three years ago he was sitting right in that seat where you are.
Overdo Thank-Yous: When you enter highly successful cultures, the number of thank-yous you hear seems slightly over the top. At the end of each basketball season, for example, Spurs coach Gregg Popovich takes each of his star players aside and thanks them for allowing him to coach them. Those are his exact words: Thank you for allowing me to coach you. It makes little logical sense—after all, both Popovich and the player are amply compensated, and it’s not like the player had a choice whether to be coached. But this kind of moment happens all the time in highly successful groups, because it has less to do with thanks than affirming the relationship.
For example, when I visited KIPP Infinity, a remarkable charter school in Harlem, New York, I witnessed teachers thanking one another over and over. The math teachers received T-shirts marking Pi Day as a surprise present from the administrative assistant. Then Jeff Li, who teaches eighth-grade math, sent the following email to the other math teachers in the department:
Dear math teachers I love,
On Assessment #7, a mid-unit test on linear functions (part of the foundational major work of the grade), the class of 2024 has outperformed the previous two classes on essentially the same test. See below for the data.
Class of [2022]: 84.5
Class of 2023: 87.2
Class of 2024: 88.7
I know this is a result of better teaching at every grade level from 5th grade on…so thanks for being great teachers who are pushing to get better each year. It’s working!
—Jeff
While all this thanking seems over the top, there’s a strong scientific support that it ignites cooperative behavior. In a study by Adam Grant and Francesco Gino, subjects were asked to help a fictitious student named “Eric” write a cover letter for a job application. After helping him, half of the participants received a thankful response from Eric; half received a neutral response. The subjects then received a request for help from “Steve,” a different student. Those who had received thanks from Eric chose to help Steve more than twice as often as those who had received the neutral response. In other words, a small thank-you caused people to behave far more generously to a completely different person. This is because thank-yous aren’t only expressions of gratitude; they’re crucial belonging cues that generate a contagious sense of safety, connection, and motivation.
In my research, I sometimes saw the most powerful person in a group publicly express gratitude for one of the group’s least powerful members. For example, the chef Thomas Keller, who runs French Laundry, Per Se, and other world-class restaurants, has a habit of thanking the dishwasher at his restaurant openings, highlighting the fact that the performance of the restaurant depends on the person who performs the humblest task. Urban Meyer, who coached Ohio State football to a national championship in 2015, used this same method at the team’s post-title celebration at Ohio Stadium, which was attended by tens of thousands of students and fans. Everyone presumed he would begin the celebration by introducing the star players who had led the team to success. Instead, Meyer introduced an unheralded player named Nik Sarac, a reserve defensive back who, at the beginning of the season, had voluntarily given up his scholarship so that Meyer could give it to a player who could help the team more. Meyer spotlighted Sarac for the same reason Keller spotlighted the dishwashers—Here is the unheralded per
son who makes our success possible.
Be Painstaking in the Hiring Process: Deciding who’s in and who’s out is the most powerful signal any group sends, and successful groups approach their hiring accordingly. Most have built lengthy, demanding processes that seek to assess fit, contribution (through deep background research and extensive interactions with a large number of people in the group), and performance (increasingly measured by tests). Some groups, like Zappos, have added an extra layer of belonging cues: after training is complete, they offer trainees a $2,000 bonus if they quit (about 10 percent of trainees accept the offer).
Eliminate Bad Apples: The groups I studied had extremely low tolerance for bad apple behavior and, perhaps more important, were skilled at naming those behaviors. The leaders of the New Zealand All-Blacks, the rugby squad that ranks as one of the most successful teams on the planet, achieve this through a rule that simply states “No Dickheads.” It’s simple, and that’s why it’s effective.
Create Safe, Collision-Rich Spaces: The groups I visited were uniformly obsessed with design as a lever for cohesion and interaction. I saw it in Pixar’s Steve Jobs–designed atrium, and in the U.S. Navy’s SEAL Team Six’s expansive team rooms, which resemble hotel conference areas (albeit filled with extremely fit men with guns). I also saw it in smaller, simpler levers like coffee machines.