The Boss Stops Here

Illustration by Marc Boutavant

The headquarters of Menlo Innovations occupy the basement level of an office complex in downtown Ann Arbor, not far from the main campus of the University of Michigan. The space is airy and bright and capacious, as befits a successful software company, but unlike the standard-issue corporate hive, there are no cubicles at Menlo, and very few walls, and only a couple of doors, one of which leads to a closet full of sensitive legal documents and another that opens into a conference room—a concession to clients unaccustomed to (or unwilling to fully participate in the spirit of) what is known around here as “the Menlo Way.”

At ten in the morning on a recent Wednesday, the 50-odd employees of Menlo, most of whom are young and appealingly tousled and predisposed toward navy or forest-green hoodies, rose from their desks and formed a large circle in the middle of the room. Menlo developers practice something called “pair programming”—a technique whereby two coders work simultaneously on a single machine, with one actually manning the keyboard and the other backseat driving from an adjacent chair. The groupings typically remain intact for a few days or a week, at which point they are scuttled and reassigned, the hope being that the constant mutation in team structure will help encourage creativity and prevent frustration.

Since there are no bosses at Menlo (at least not in the traditional sense) and no middle managers (ditto), all that reassigning and fluctuation falls to the team as a whole, a process that requires a lot of air-traffic control. Every morning, the entire staff circles up to discuss strategy.

On the day I attended, the meeting moved with martial efficiency. Two by two, the pairs stepped forward and, with each employee gripping one of the horns of a fat-lady-sings-style ­helmet, the unofficial symbol of Menlo Innovations, they laid out their plan for the eight hours ahead. The bulk of the speakers were developers and designers (or “high-tech anthropologists,” a term Menlo has registered with the U.S. Patent and Trademark Office), but there were also quality-assurance and support staffers on hand, and the conversation ran from the jocular to the mundane to the technical:

“We’re going to be calling this project ‘Gobstopper,’ because ‘Jawbreaker’ just sounded a little too hard-core.”

“Right now, I’m doing some organizing of the storage closet.”

“Every day, I’m reminded of how much human suffering in the world is related to technology. I’m talking about the new Gmail design.”

“I’m Andrea, and I’m working on client invoices.”

There was some clapping and some backslapping, of the kind you might observe at a particularly boisterous rec-league indoor soccer game, then all the employees were filing back to their desks, past a poster of Frank Zappa and a bust of Thomas Edison, the patron saint of Menlo and the guy whose famed laboratory, in Menlo Park, New Jersey, inspired this Michigan company’s name. Later, one of the quality advocates at Menlo, Joe Rock (real name), explained that the a.m. meeting helps keep morale on the team high and, more important, encourages a feeling of camaraderie—a sense that every one of the staffers is working together toward a common goal. Hence the circle, which, in a nod to King Arthur’s court, ensures that no one gets a seat at the head of the table.

As I soon discovered, basically everything at Menlo Innovations is so open and transparent and flat that the average office worker, upon entering the Menlo den, might be forgiven for feeling a little suspicious, intrigued, cynical, and jealous all at once.

Consider, for instance, the fact that hiring at Menlo is handled by committee, with each applicant spending a little bit of time with a group of employees, until a consensus can be reached. That same collective decision-making happens during promotions, layoffs, and flat-out firings.

Consider next the charts in the corner of the office, which display the names and titles of the Menlo employees and also their corresponding pay grades. When I first saw them, I was standing in the midst of a scrum of Menlonians, and I suggested—thus belying my own, frankly square work experience—that it might be a little unnerving to have your salary exposed to your colleagues. And the guy standing to my right actually scoffed. “No,” he said. “It’s the opposite. It’s liberating.”

It’s a relatively safe assumption that most of us have, at one point in our lives, worked for a boss. There is comfort in the arrangement: Someone tells us what to do, and we do it. If we do it well—and “it” here could be anything from writing software to assembling a car—we may get a more spacious cube or more money, and if we do it poorly, we can expect to be let go. Above us, in an ever-narrowing spire, are the shift supervisors and floor managers and vice-presidents, each of whom is subject to his or her own unique hierarchical pressures, and above them is the CEO or president or otherwise-titled grand Pooh-Bah who dictates the rules that the rest of us must follow.

Pay grades are posted for everyone to see.Photo: Illustration by Marc Boutavant

According to Nikil Saval, the author of Cubed, a forthcoming history of the modern office, the top-down management structure first proliferated in the U.S. in the rail era, as corporate barons struggled to maintain control of their sprawling new concerns. The easiest way to govern hundreds of thousands of miles of railroad, they discovered, was to erect a chain of command, which extended from the central office in New York or Chicago to the field offices on the frontier.

It was a strategy that also proved remarkably effective for the heads of large banks and telephone companies and eventually—I’m skipping a few decades here—PC manufacturers and soda-pop-makers and multinational data-­processing firms. There may even be some evidence that the tiered framework is hardwired into our brains. “Hierarchy is prominent across all species and all cultures in the world,” Adam Galinsky, a professor at Columbia Business School, told me recently. “It reduces conflict, helps with role differentiation, and vastly increases coordination.” In other words, employees may need managers because managers define, either implicitly or explicitly, who people are as workers.

In a research paper published last year titled “The Path to Glory Is Paved With Hierarchy,” Galinsky and several colleagues measured the productivity of “mixed power” work teams and found that tiered groups outperformed flat ones. The pecking order, they concluded, is the “universal default for human social organization”—a default that requires minimal social interactions to emerge. The paper cites work by poultry scientists, who have discovered that putting too many high egg producers in the same small space actually decreases overall egg production: “It turns out the best egg producers are also the most competitive birds, and in a group setting, they quickly begin fighting over food, space, and territory; these intragroup conflicts then drive egg production down and bird mortality up. Chicken farmers take note: If you want to maximize group-level productivity, you need harmony, and it seems that hierarchy provides the key.”

The theory that too many bosses may be an obstacle and not a boon did not achieve widespread prominence until the early eighties. The reasons are multifarious, but business historians believe it had something to do with the economic recession, which gutted the ranks of the middle managers and in the process helped companies realize that all those bosses had actually been slowing things down. There was also an increasing sense that ­creativity—an invaluable commodity at the tech firms and software companies of the new “knowledge economy”—might be muffled by hierarchy. A tiered framework had worked fine for railway bosses, but it could have a frankly inhibiting effect on a team whose sole task was to build something new, often out of thin air. For that, you needed space, you needed support, and above all, you needed freedom.

Studies, in fact, show that although some structure is conducive to creativity, the “time to experiment and potentially fail”—to quote Pierre Azoulay, an associate professor at the MIT Sloan School of Management—is vital to the consistent production of innovative ideas.

Among the earliest agitators for a decentralized workplace was Ricardo Semler, the chairman of Brazilian industrial-parts manufacturer Semco. Semler took over Semco from his father, in the early eighties, when he was 23; his first move was to eliminate the majority of the executives and managers. He opened up the company books, organized employees into self-sufficient work teams, and instituted profit-sharing. Semler believed that Semco would perform better as a company if the workers didn’t have a platoon of “bean counters”—one of his favorite insults—constantly looking over their shoulders.

“Bureaucracies are built by and for people who busy themselves proving they are necessary, especially when they suspect they aren’t,” he wrote in his 1988 book Turning the Tables (published in the United States under the title Maverick). “All these bosses have to keep themselves occupied, and so they constantly complicate everything.” Rules are good for prisons and armies, Semler wrote, but “for a business that wants people to think, innovate, and act as human beings whenever possible,” they serve only four purposes:

1. Divert attention from the company’s objectives.

2. Provide a false sense of security for the executives.

3. Create work for bean counters.

4. Teach men to stone dinosaurs and start fire with sticks.

Maverick was an international best seller, and it remains a foundational text for advocates of the flattened workplace today. Tom Preston-Werner, a co-founder of the tech firm GitHub, which allows all employees to set their own schedules and choose their own projects, told me that the book “changed my understanding of how a business could be run.”

In 1980, less than 20 percent of the companies on the Fortune 1000 list boasted at least some sort of team management structure. By 1990, it was 50 percent. By 2000, it was 80 percent. “Companies were trying to figure out the best way to foster creativity, to effect rapid change, to deal with growing global competitiveness,” says Stephen Courtright, an assistant professor at Texas A&M, who specializes in the study of self-governing workplaces. “In many cases, that involved flat, horizontal management.”

Work is just as transparent, with lots of team projects—and scant privacy.Photo: Illustration by Marc Boutavant

But only in recent years have we really seen the ideal of the democratized workplace brought to its logical conclusion: companies that don’t just have fewer managers and bosses but have hardly any bosses at all.

Last month, Fox debuted a reality show called Does Someone Have to Go? In each episode, a squadron of Fox producers descends on an office, puts the employees in charge of day-to-day operations, and turns on the cameras. “You will all be the boss,” one of the teams is told in an early promo. Cut to a close-up of the furious, disbelieving mug of an employee who looks like he’s just been asked to guillotine his golden retriever.

Much of the show works like this: It turns the concept of self-management into the stuff of nightmare. As the New York Times noted, the program assumes “that what’s wrong with the American workplace is the workers. The problems discussed aren’t about the structure of the company, or the state of its chosen industry and market, or the economy as a whole. Employees are the enemy.” And bosses, by default, the heroes.

All of which would deeply surprise Simon Anderson, the CEO of the web-hosting company DreamHost. “Management is a term to me that feels very twentieth century,” he says. “That 100-year chunk of time when the world was very industrialized, and a company would make something that could be stamped out 10 million times and figured out a way to ship it easily, you needed the hierarchy for that. I think this century is more about building intelligent teams.”

Anderson was one of three finalists for the CEO job at DreamHost in 2011. In choosing their new chief executive, the founders of the company decided to open the floor to an online vote. To make his case, Anderson was trotted in front of a majority of its 100 employees, asked to give a little speech, and answered questions from the crowd. He won narrowly, by a margin of 53 to 47 percent. Anderson told me that he has since spoken to some of the employees who voted against him. “There were no hard feelings,” he said.

DreamHost has a management structure, but workers have the freedom to select their own projects, and the offices, in downtown Los Angeles, are open 24 hours a day.

And there is an increasing number of companies around the country that function similarly. Development at Valve Corporation, a video-game company based near Seattle, is conducted by a network of self-governing teams. Employees choose which team they join and also choose when they’d like to join a new one. Jason Holtman, who until recently worked in business development at Valve, has, in the past, called the setup “an organic gravity well”—the number of people on a project helps determine which games are shipped and which are held for more work.

At W. L. Gore & Associates, the makers of Gore-Tex, there are few titles (almost everyone is referred to as an “associate”), and once a year employees gather to rank their colleagues based on their contributions to the overall success of the company. Those rankings are used by a separate committee of associates to determine pay raises or cuts.

And at IDEO—a sprawling design firm with offices in New York, Mumbai, and London, among other cities—work teams are multidisciplinary and mostly self-governing. Every team opens a project by setting a series of personal and collective goals, reviews its progress midway, and closes the project with a self-analysis session.

This structure—largely flat and very flexible—is especially appealing to those new to the workforce, twentysomethings who tend to approach work differently from their parents. “The way workers are motivated is changing,” says Anderson of DreamHost. “Twenty years ago, it was about higher pay. Now it’s more about finding your work meaningful and interesting.” As more and more millennials enter positions of power in the business world, Anderson believes we will soon reach a point where hierarchy itself is “passé.”

Then there are the employees of Morning Star, a California tomato processor that offers an interesting demographical contrast (its employees are not young creatives). Morning Star annually processes thousands of tons of ripe tomatoes, which are ground down for use in ketchup, pizza sauce, or tomato paste. A little over 2,000 people work for Morning Star at the height of the tomato season. In order to be hired, you first sign something known as a clou, or a Colleague Letter of Understanding. The clou outlines your priorities for the year ahead. If you are a tomato sorter, you pledge to sort a predetermined amount of tomatoes a day for the duration of your stay at Morning Star, and if you are the man who is responsible for helping evaporate the water out of the squishy tomato pulp, then you sign an agreement to evaporate a specific number of gallons of water every week.

Co-workers consult on who gets to stay …Photo: Illustration by Marc Boutavant

Around the facilities, everywhere you turn, are whiteboards and chalkboards and big-screen TVs displaying names and cascading numbers—the tally of what is being done and what remains to be done by each worker. (Morning Star frames it as emancipative—your impetus to work is no longer tied to the whims of your boss but to your own motivation—and most employees seem to agree. But to an outsider, the whole thing seems a little oppressive. What if you have an off day?)

Morning Star was founded by Chris Rufer, a graduate of the UCLA business school. In interviews, Rufer, like Semler, has described the structure of his company as deeply humane, something that buoys the spirits of the workers. The relationships between members of a self-managed team, he has said, are more organic and thus more substantive than the traditional relationship between boss and subordinate, which he describes as both “forced” and “artificial.”

To that end, yearly evaluations at Morning Star are handled not by a manager (there are none), but by a group of colleagues who administers a written evaluation of your progress. Paul Green Jr., a member of Morning Star’s Self-Management Institute, a kind of internal R&D arm of the company, says the goal is to meet a series of self-defined (capital S) Steppingstones: X amount of gallons, Y number of tomatoes, and so forth.

Conflicts, meanwhile, are handled via a four-step procedure called “Direct Communication and Gaining Agreement.” The first step is to attempt to “appeal directly to the colleague,” Green told me. “See if you can sort out your differences.” Second step is to bring in a third colleague to mediate the argument. The third is to create a panel of six to ten additional colleagues. If all that fails, Rufer can be called upon to join the panel and help render a decision, which occurs about ten times a year.

It may be worth noting that the egalitarian spirit at Morning Star only goes so far: The company is privately held, and no employee, no matter how hard-performing, is entitled to a share of the profits.

One of the first things that Rich Sheridan, the co-founder and CEO of Menlo Innovations, told me, upon my arrival, was Menlo’s origin story. He tells this story a lot, at conferences and panels, and the whole spiel is practiced, and fluid, and very TED Talk–esque. It starts in the late fall of 1997, when Sheridan, at the time a V.P. at a document-translation firm called Interface Systems, invited his then-8-year-old daughter, Sarah, to join him for a day at work. Around five in the afternoon, Sheridan turned to her and asked what she had learned.

“Well, Daddy,” she replied—and here, Sheridan screws up his face and crunches his body low into his chair and, for a 55-year-old man, does an admirable impression of an 8-year-old girl—“I learned that you’re really important.”

“What makes you say that?” Sheridan asked.

“All day long,” she said, “people came in here and asked you to make a decision for them. And you made a decision, and they went on their way.”

Sheridan was mortified. “I realized that the organization couldn’t move any faster than me,” he said. “That I was the bottleneck.”

He shared his epiphany with a young consultant, James Goebel, who harbored some admittedly radical ideas about modern management. Sheridan had originally brought Goebel onboard to help teach his team some new programming techniques, but they decided instead that Goebel should help him with a redesign of Interface.

Over the next six months, cubicles were dismantled, managers were moved out of their offices and into the middle of the floor, and pair programming was introduced. No longer could any code belong to any one person—everyone would share everything. “I had one guy raise his hand,” Sheridan remembers, “and say, ‘Rich—blood, mayhem, murder. For God’s sake, don’t put us in an open room, don’t make me share my computer with another human being, and please, please don’t make me share my code.’ ”

Sheridan believed that by going from a “siloed” environment to a collectivized one, Interface would be more productive. And happier. And by all accounts it was, until 2001, when the dot-com bubble burst, and Sheridan and his entire team were shown the door. That same year, Sheridan and Goebel and two partners founded Menlo Innovations. There were the usual start-up obstacles (wary clients, financing challenges, anemic bank accounts), but over the past decade, Menlo vastly increased its revenue, outgrowing one office space after another.

When I first contacted Sheridan, by e-mail, he wrote that it was “quite bossless at Menlo! Team makes hiring decisions, team gives feedback, team decides promotions.” This is only partially true. Overseeing strategy, the long-term vision of Menlo as a whole, still falls to Goebel (now the chief architect and COO) and Sheridan. If one worker accused another of something ­illegal—sexual harassment or theft—Goebel and Sheridan would be responsible for taking the appropriate measures. They also serve as representatives for Menlo at scads of management and business conferences here and abroad. (Sheridan’s business card reads “Chief Storyteller and Tour Guide.”)

… And who must go.Photo: Illustration by Marc Boutavant

In every other way, however, Goebel and Sheridan are not traditional bosses. Overarching strategy is their domain, but the tactical stuff—the daily squall of code-making and high-tech “anthropologizing” that drives the company—belongs entirely to the employees. On the days I spent at Menlo, I never once saw any worker pay Sheridan or Goebel any special deference. No one talked any differently in their presence. They were there as team shrinks and advisers, and yet they were also not there—the rest of the office thrummed on around them, regardless of what either of them did. Sheridan’s troubling “bottleneck” had been removed; in its place was a largely self-sufficient software-­coding machine.

I asked Sheridan if he ever missed his old office and the clout that came with it. “I liked being the person everyone came to,” he admitted. “There was glory to it. I felt like the smartest guy in the room. But that doesn’t matter to me anymore. It’s not my goal.”

Menlo’s hiring process is called “Extreme Interviewing,” and it bears a striking resemblance to speed-dating. Applicants, sometimes as many as five for each open position, are brought into the offices for a series of rapid-fire interviews with a range of current employees. The emphasis is on “kindergarten skills”: geniality, curiosity, generosity. As the Menlo white paper on Extreme Interviewing puts it, technical proficiency is less important than a candidate’s “ability to make [his or her] partner look good.” (Sample interview question: “What is the most challenging bug that you helped someone else fix?”)

Later the Menlonians gather to compare notes. Lobbying is common; so is argument. Eventually, the candidates are ranked partially on their technical skill and partially on what the white paper describes as the team’s “value set.” Offers are extended to the applicants at the top of the list.

“The collective aspect allows one of us to say, ‘Okay, he did really well at this,’ ” one employee, Greg Haskins, told me. “And another of us to say, ‘But not so well at that.’ And we bounce opinions back and forth before deciding. Basically, we build our team together.”

In an old-fashioned, top-down work environment, employees get rewards from three main sources. One is pay (the better we do, the more we get). The second is title (the better we do, the higher we ascend). The third is a sense of achievement (recognition for having performed a task admirably). In a bossless environment, the calculus is significantly changed. Achievement is tied to the team as a whole—you can write a particularly nice piece of code, but it won’t mean squat until the whole project is up and running, and for that, you’ll need the assistance of your peers. Individual achievement is obviated and replaced by shared achievement. Friendliness and congeniality start to matter more, ladder-climbing a whole lot less.

And herein an important point: Horizontally managed companies work in large part because they tend to attract people who are okay working in a bossless environment and weed out the ones who aren’t. (“I can smell a corporate-ladder climber from a mile away,” one Menlonian explained.)

Which is not to say that everything at Menlo is peace and love and synergistic intellectual communion. It’s not. It couldn’t be—no matter how lofty the theories espoused in the glossy, professionally bound 73-page Menlo guidebook (available for free to visitors), its adherents are human, who by their very nature gossip, back-talk, and bicker. “Are there fewer problems here than elsewhere?” Goebel asked me rhetorically. “I’m inclined to say no. We just deal with our problems differently. And the problems themselves are different.”

Most of those problems originate in the vacuum left by the traditional management structure. Stephen Courtright told me the story of a manufacturing company that had “misunderstood,” in his words, what self-management was all about—few boundaries or guidelines were set for the work teams, and the feedback loop was practically nonexistent. (Courtright, who was consulting for this company, declined to disclose its name.) “They were just kind of set loose to do their thing,” Courtright said, “and they produced too much inventory, and the company ended up with a stockpile of unused stuff.” An executive decision was made to go back to a top-down hierarchy.

At Menlo, employees must necessarily be self-starters and willing to work collaboratively; if they aren’t, the rest of the team will show them the door. Several Menlonians recalled for me the story of a man I’ll call John. At first John seemed like an ideal employee—diligent and hardworking. But the horizontal structure at Menlo clearly ate at him, and under-the-breath grumbling soon gave way to very loud bellyaching.

One afternoon at lunch, he stood up and told his colleagues that they were “insane—that they needed to be freed from this weird Nazi regime,” in the recollection of someone who witnessed the explosion. Sometime after, the team fired him.

Firing, developer Kealy Opelt told me, is something the company takes very seriously. “Usually, I wouldn’t think about bringing in the group unless I’ve had to talk to someone several times. If it keeps happening, and you both can’t adjust, that’s when you start talking to everybody else and see if it’s just the two of you.”

Over subsequent weeks, a series of meetings will be held, and as with the Extreme Interviewing process, a consensus will be reached. If it is determined by the team that a person has to go, there’s very little that Sheridan or Goebel can do about it.

To illustrate, Goebel told me about his niece, Erin, who spent a couple of months at Menlo as an admin, before the other employees decided to let her go. Goebel objected, but at Menlo, nepotism has no place. “Actually, my niece lives with me,” he said. “And she was really pissed. Thankfully, she didn’t take it out on the relationship with my wife or my kids. Still, it was a little frosty for a while.”

This kind of team decision-making can be seen as the ultimate motivational tool—always work hard, and your co-workers won’t can you. It can also be the kind of thing that yields a whole lot of ad hoc office alliances.

Opelt and colleague John Martin, for instance, told me that they had both recently recommended each other for promotions. They swore it was coincidental—they both believed the other deserved it—but it’s not hard to imagine the existence of a lot more deliberately constructed pacts.

And then there is that chart that’s plastered prominently on the wall—the one that displays the titles of Menlo employees and their respective pay grades. This is a new innovation, introduced by Goebel over lunch a few months earlier to address curiosity from staffers about where they stood in the overall scheme of the open office. Although almost all of the employees had written their names in Magic Marker on the poster, there were a few abstainers. “I think people are still deciding how to handle that,” one employee told me. “It’s stirred up a little bit of controversy. Like, that it might actually be a step too far.”

Goebel and Sheridan confessed that the idea was still in its infancy—“an experiment,” Goebel said. Paul Green used much the same language to describe an analogous open-book policy at Morning Star. “Full transparency in terms of compensation is something we’re working on,” he said. “It’s a process.”

One afternoon, I was invited to attend the weekly “kickoff” meeting for the fifteen team members assigned to the account of a local biotech firm that manufactures flow cytometers—machines that test blood and sundry other fluids. Menlo is developing the software that organizes and analyzes data collected by the device.

The purpose of the kickoff meeting was to discuss the code that needed to be completed by the end of the week, and the fifteen Menlonians arranged themselves in a loose circle around a set of tables. Notepads and pens were produced. Goebel and Sheridan were absent—neither man typically attends these meetings because, as Jack Coy, a developer on the account, told me, “to be honest, we don’t need them.”

Initially the proceedings moved smoothly. A pair of programmers would walk up to one of those big easel pads and discuss the code they’d been assigned and sometimes sketch on the paper a couple of key components. The floor was then opened to the rest of the team, who might shout out questions or suggestions.

At one point, I watched Jeff Jia and Jack Coy, two programmers in their twenties, describe an issue with an important piece of code. Caveat: Software design is an extremely complex science, and people who work in software tend to talk in jargony, acronym-heavy shorthand. For that reason, I will not attempt to recount the exact parameters of the problem Jia and Coy were facing, but suffice it to say it involved a digital filter that had been built to catch certain subsets of cytometer data and arrange those subsets into colored charts.

But the filter wasn’t functioning properly, and Jia and Coy were stumped. They had the look of stumped people—lips puckered, arms crossed, shoulders raised. After a little while, they were joined at the easel by Greg Haskins. Haskins is 27, with a babyish face and unruly brown hair, and although he has only been at Menlo for little over a year, he clearly maintains an outsize presence among his peers; when he talks, the other coders tend to listen.

Now Haskins gestured a few more team members to their feet and began arranging them in front of the easel pad like mannequins. What followed was a piece of anarchic and improvised street theater: Rock, the ponytailed quality-advocate dude, was assigned the role of filter, and Kristi Sitarski and Joe Ptolemy were the data that the filter was meant to be sorting. Sitarski stepped forward (she was in), and Ptolemy stepped back (he was out). Everyone laughed. This was team spirit in action. This was collective action.

Except it wasn’t, not really. Instead it was an action that had been driven in large part by an individual—Haskins. I was reminded of the control group of chickens referenced in Galinsky & Co.’s “The Path to Glory Is Paved With Hierarchy.” Some of them were prolific egg producers and some of them were low-level layers, but together, in a mixed-power group, they formed a nice symphony of egg production. Perhaps Galinsky was right that as animals, we do naturally slot ourselves into some sort of hierarchy. Certainly I was seeing something similar now: However ostensibly egalitarian this Menlo work team was, in practice, a leader had quickly emerged, and everyone else was falling in line.

A few weeks after my visit to Ann Arbor, during a conversation with Tom Preston-Werner of GitHub—itself a mostly horizontal company—I happened to use the word flat. He stopped me. “With people, there are always senses of authority and matters of seniority and so many other social factors at play,” he said. “You’re fooling yourself if you think a group bigger than a single person can ever really be flat.” But the people at companies like Menlo or GitHub or Morning Star don’t seem to care. Instead, it’s possible that the triumph of the flattened office may be the creation of work environments in which leaders organically arise, and all employees feel a sense of ownership, whether real or imagined.

Jason Holtman, formerly of Valve, has said that the big negative of going bossless is the loss of the “quick mandate”—the ability to swiftly and effectively mobilize the organization. A management structure may be democratic, and thus empowering to the workers, but that does not mean it is neat, nor particularly fast.

That kickoff meeting for that biotech firm was long—it started at 11 a.m. and went until almost two in the afternoon, with a break in between—and by the midway mark, the proceedings were moving a little more slowly, with more exasperated sighs, or slight but conspicuous head shakes, and sometimes everyone seemed to be talking simultaneously, in one big, warbly squawk.

Kealy Opelt, leaning forward, attempted to wrestle the proceedings back on track.

“So where were we?” she’d say. Or: “I didn’t hear that, Jeff. Could you repeat it for everyone?” Or: “One at a time, please.”

Later, I asked her if she’d been frustrated.

“Look,” she said, “we don’t want people having their arms all crossed and not saying anything, with this brooding idea they haven’t shared with anybody. That would be worse. But yeah, it can get chaotic.”

Chaotic, noisy, somewhat lumbering—Menlo may be all those things, but it is also a profitable company. Revenue in 2012 was about $4 million, and employee retention rates are notably high. As the organization has grown its client list, it has expanded the number of programmers it keeps on staff, and when the company posts an advertisement for an open gig, a flood of applications typically pours in. Around Ann Arbor, Menlo Innovations is considered an extremely desirable place to work.

Not that it’s an easy place to work. More than one employee told me that it took them some time to get used to the peculiar demands of the office—the lack of doors, the lack of privacy, the constant and sometimes exhausting presence of the rest of the staff.

To help relieve some of that pressure, “walkies” were introduced several years ago. A “walkie” is exactly what it sounds like—a ten-minute group walk around the block. The daily ritual is intended to allow the basement-bound Menlonians a chance to unwind and to see the sun. On one walkie, I found myself between two Menlo employees, Natalie Svaan and Jack Coy. They could not be more unalike. Svaan, a Michigan native with silver-streaked black hair, came to Menlo after a career in IT. Coy is still in college; he works at Menlo part time.

“I remember when I first got hired, I was so wiped out,” Svaan was saying. “I’d just come home and collapse. But you know, it’s strange; it’s gotten to be energizing. The hours do actually fly right by, probably because you’re always working.”

“And there’s the team thing,” Coy added. “It’s part of your task to make the team succeed and to help other people do their jobs. It’s not to promote oneself at the expense of the rest of the team. I like that.”

He shook his head. “I think there are actually people,” he added, in a tone of deep and abiding wonder, “who want to just go to work, be told what to do, do it, and go home.”

The Boss Stops Here