The Silo Effect

by Gillian Tett

  These Hackathons were not unique to Facebook. On the contrary, they are a well-worn ritual across the technology world. At the City Hall in Chicago, Brett Goldstein arranged hackathons with his nerd herd, to develop programs the city could use. But at Facebook the hackathons had an unusual twist. These events had initially simply occurred whenever some engineers clustered together with their friends and teammates around a laptop. Thus in the early days of Facebook, Zuckerberg had spent all night brainstorming ideas with the other founding members of the company in the house he shared with the engineers. But as time passed the Facebook managers insisted that the groups who coalesced in a hackathon night had to cluster together with people from different teams from their normal projects and work on something outside their day job. Sometimes engineers congregated together into short-term mini-teams because they had made contact in the days leading up to an event, or because they were all interested in a specific problem. On other occasions engineers decided to work with each other almost by chance. Either way, hackathons—like Bootcamp and Hackamonth—were intended to break down the normal departmental boundaries. They were another tool to ensure that while Facebook had dedicated project teams, these could not ossify into inward-looking silos.

  Soon after joining Facebook, Goldfein used a hackathon to write some computing code that could add the category of in-laws and step relations to the Facebook platform. This project was not connected to her News Feed job at all. But it was a project dear to her heart. “I am really close to my husband’s mother, but didn’t want to call her just an in-law, but when I started looking at the Facebook site after I joined I saw that although we had ‘family’ as a group on the site, we couldn’t do in-laws. But that seemed wrong, so I used a hackathon to think about it.” Or as Pedram Keyani, an Indian-born computer engineer who organized many of the hackathons, explained: “The whole point of hackathons is that they shake up the hierarchy. We break down the silos, at least for a period, and get a communal spirit.”

  ANOTHER TOOL THAT THE Facebook managers used to fight silos was the Facebook platform itself. Right from the earliest days of the company, the Facebook managers realized that one of the advantages of Facebook was that it enabled communication to occur in a horizontal way, instead of via a rigid hierarchy. When somebody made a post, everyone could access that piece of information. This was a contrast to the usual communication pattern of many big companies via email, where information tended to get passed up, or down, hierarchies, creating potential bottlenecks and logjams.

  As time passed, the managers spotted that there was a second advantage of the Facebook platform: it provided a way for the staff to build deeper connections with each other, on multiple different levels. The Facebook managers believed that this was another important weapon in the war against silos. “One technique we teach managers here is that we insist that people use names—real names—to talk about each other,” Schrep said. “If we ever catch anyone using a depersonalized moniker then you interrupt them and stop them. We never let people refer to anyone else as, say, ‘those idiots in team six’ or “ ‘those stupid marketing guys,’ since it is one sign of dehumanizing a group. When you don’t know who people are and depersonalize groups, then you get into problems.”

  The Facebook platform could help that goal of humanizing staff, the managers hoped. So the top managers were urged to create their own Facebook posts and share ideas, messages, and details about their lives with their colleagues. Some people hated that. “I am a pretty shy, introverted person naturally,” Goldfein admitted. “Before I joined Facebook I didn’t post very much at all. I was pretty inhibited.” But soon after she arrived, she bowed to the pressure and started posting to her colleagues. Her opening post was reserved: “What’s the worse bug you’ve seen caught in a code review? Can you recommend a better OSX-based IRC client than adium?”31 Then Goldfein started to open up. “Hi there, subscribers!” she declared on her fifth post. “I work in Facebook engineering on core features of the site, like news feed, photos, and search. In my previous life, I ran engineering for VMware’s desktop products. I’m passionate about getting more women into engineering careers and the challenges of scaling software engineering organizations. I have kids instead of hobbies. Science fiction and Glee are my guilty pleasures.”32 As the weeks passed, she started revealing more details about her private life. “I’m kind of fascinated by the variances in my commute at different weird times of day Anyone know of an app for tracking this? Preferably one w/ facebook integration?”33 She posted her favorite recipe for madeleines on her Facebook page. She talked about a cause dear to her heart: the gender imbalance in the computing world. “In the 1970s, boys outnumbered girls 13:1 for perfect scores on the math SAT. Today, the ratio is 3:1 and falling. (And average score is identical.) Is there anyone left who seriously thinks differences in math performance are biological? If so, I have a science lesson for you on how fast evolution works.”34 She championed media outlets trying to buck gender stereotypes. “So Brave is the first Disney princess movie I’ve ever seen in which mom is neither gone nor evil. And, it appears to be pulling down the big bucks. Crossing my fingers in case the world might be changing for the better.”35 When she traveled to Washington on a lobbying campaign, she revealed details of that too. “It was particularly cool to meet Cady Coleman, a colonel and NASA astronaut. My ambition to be an astronaut when I grew up did not survive high school, but I definitely think NASA, Sally Ride, and Christa McAuliffe helped fuel my interest in math and science from early on.”36

  Sometimes Goldfein felt nervous about revealing so much. But as time passed, she realized that it was the personal that helped her connect. “Every time I share a little more outside my comfort zone it feels so rewarding. I end up making new connections in the company, and it’s become like a positive feedback loop.” She was not the only one. During 2012, the reticent Zuckerberg started to post details of his wedding, his backyard, and his daily life, albeit with a corporate spin. “The Facebook foxes that live on our campus are pretty amazing,” he mused to staff in one post.37 Or: “I updated my grilling app, iGrill, today and it now has Facebook integration that lets you see what other people are grilling right now around the world. Awesome! I’m making a Fred’s steak!”38 Sandberg posted information about the company and used the site to promote her book, called Lean In, that urged women to take more responsibility for their careers. Schrep used the platform to share tips with other engineers on how to work effectively. “Focus. Focus, Focus. Facebook engineers are used to hearing me say the word focus A LOT. De swiss cheese your schedule! Abide by no-meeting Wednesday. Mind your health as this is a marathon not a sprint: exercise and good sleep make you more productive.”39 Then he revealed that his secret tips for success were “1) Eliminate all visual distractions on my screen and focus on one app at a time. FocusMask works great for this. 2) Set up intervals for focused time, breaks, and interrupt driven time. Using a simple timer during focus intervals helps me not accidentally wander. 3) Large over-ear headphones which block out sound and tell others to not interrupt me now. Music without lyrics—Mogwai or classical depending on mood. What works for you?”

  Junior staff shared thoughts too. In the spring of 2013 an engineer named Ryan Patterson posted a note to tell colleagues that “this week marks the beginning of my 4th year at Facebook.” By the standards of the company, that made him a veteran. So he decided to offer advice to new recruits. “As a hacker, having agency over your world is critical to fully explore the boundaries of problems and find how to best leverage your solutions. . . . I’m referring both to the code that you write and to the interactions you have within the company. . . . Take ownership over the entire company. Find a random person with whom you haven’t worked in months and have a quick chat with them. It can provide a fresh insight.”40

  ON AN EARLY SUMMER evening in May 2013, several hundred young(ish) men and women wearing sneakers, jeans, and T-shirts assembled around the yellow crane in Hacker Square. The s
un was dipping below the horizon, leaving pink streams in the luminescent blue sky. A man on stilts bounded across the square. Another waved an old-fashioned boogie box throbbing with dance music. A third man, bearded, wearing a blue mask and indigo Zorro-style cape, jumped onto the base of a crane. The Facebook engineers had stumbled across this hunk of machinery at another office a few years earlier, entirely by chance, and become attached to it. They used it for office meetings and pranks. So when the company moved to the new campus at Menlo Park, the engineers insisted on moving it too, as a ceremonial prop and symbol of their origins. The company had only been in existence for a decade. However, the Facebook engineers had already forged a multitude of group rituals and founding legends. The pattern would have looked familiar to an anthropologist such as Bronislaw Malinowski, the Polish academic who studied the Trobriand Islanders in the 1930s and pioneered the idea of “participant observation.” The ceremonies at Facebook, like the Kula exchanges in the Trobriand Islands, had an important function; they were helping to glue the wider social group together by forging a common sense of identity.

  “Friends, Romans, Hackers! Lend me your ears!” the man in the cape shouted. The crowd temporarily fell silent. Pedram Keyani, the nominal leader of the Hackathon group, stepped forward. “You know the form!” shouted out Keyani, sporting jeans and an olive green T-shirt. “If you are still here at 5 a.m. there will be breakfast! And Chinese later!” The crowd laughed again. Hackathons always started on the yellow crane, and then moved to another room where the engineers always ate exactly the same Chinese takeout several hours later, in the middle of the night. Zuckerberg had first started buying food from this particular Chinese restaurant when he created Facebook a decade before. By 2013, the restaurant was no longer a remotely convenient place to order from, since it was a long way from Hacker Way. But nobody wanted to mess with that ritual by ordering from somewhere else.

  “Is there anyone here who has not been to a hackathon before?” Keyani shouted to the crowd. A few hands waved. “Well that’s awesome! The only rule is—go enjoy yourself, go mingle, go have fun! And just remember our hackathons are not like anyone else’s hackathons. Enjoy! Let’s write some code!” The gang of hoodie-wearing engineers dispersed in small groups and walked to a large conference room studded with inspirational slogans: Move fast, and break things! Done is better than perfect! What would you do if you were not afraid? “Posters are one of the biggest culture carrier at Facebook,” observed Sandberg. “We will be in a meeting and someone will quote something to somebody else like ‘Make Great Decisions.’ That’s the type of company we are. A lot of the initial cultural ideas came from Mark, but we are very aligned as a group. You cannot impose this from the top. It has to grow from below too.”

  COULD THESE LESSONS BE transplanted into other companies? From time to time, Sandberg and the other executives would ask themselves that. By 2013, other companies across Silicon Valley were using variants of the same silo-busting tactics that Facebook had developed. At Google and Apple the employees staged hackathons, and rotated staff. The idea of conducting common induction and training programs for new recruits was spreading. The concept of using architecture as a tool to promote employee collisions and collaborations was also widespread, both inside and outside the technology world. 3M, the manufacturing group, prided itself on running research laboratories that deliberately mixed different specialists up. Google had imaginatively designed facilities that enabled staff to collide.

  Numerous other entities were using social media sites to promote better employee communication. Over in Europe, the consumer company Unilever installed an internal social media platform that had been developed by Salesforce, the cloud computing group, to promote cross-company horizontal communication. This system, called Chatter, was initially created to enable the top managers to disseminate messages to everyone else. But then the Unilever executives realized it also could be used to break down silos inside the company. “Chatter has helped us to connect all our global groups and teams, share ideas, news items and latest developments and minimise the chances for unnecessary repetition in markets,” chief scientist Kim Crilly says.

  But what made the managers at Facebook unusual was the degree to which they kept turning the lens back on themselves and trying new iterative experiments in social engineering. Having built a successful business by using computing ideas to analyze human friendships, they remained endlessly fascinated by how they all interacted with each other. “I never used to think about this social stuff. It didn’t seem that important,” confessed Schrep. “But then, when I came to Facebook, I realized just how much it matters. That’s a real change! And now I cannot stop thinking about it.”

  The Facebook managers had another reason for introspection: they knew that they were surrounded by competitive threats. There were dozens of smaller, more nimble start-ups who were eager to challenge Facebook, and technology was changing at a furious pace. When Zuckerberg had first created Facebook in the early years of the twenty-first century, he built the business for desktop computers and personal computing screens. Facebook was so successful in exploiting that niche, that it was slow to adapt to mobile technology. The Facebook engineers made the mistake of assuming that they could simply take the web product and put it onto mobile. But web products did not translate well onto a tiny mobile phone. “It’s pretty well documented that we were not very good at mobile [initially]—terrible, in fact,” Goldfein admitted in the spring of 2014. “I joined Facebook in the summer of 2010, and at that point Facebook’s mobile team was like four guys in the basement. [For the web] there was a News Feed team; there was a Messenger team; there was a Photos team. . . . Collectively there were 100 or 200 engineers building all of these features [for the web], and then you had this one tiny team that was trying to keep feature parity with all of that in the iOS app. And I won’t even talk about how we were approaching Android because it was just embarrassing.”41

  When Zuckerberg and the others eventually realized their terrible mistake, they scrambled to respond. Facebook bought Instagram, the photo-sharing site, and other mobile companies.42 Goldfein and the other engineers were shuffled around, to focus on mobile technologies instead of the web. “What we learned to do is start with the [mobile] platform and make the best possible application for it and to the extent that the web has things that are useful to bring them over—but don’t start from the web and try to bring them over. Start from mobile, and take what’s good from the web,” she observed. By 2014, this switch in strategy seemed to be working. The company’s mobile app was gaining traction, and the company was harnessing advertising revenue from this mobile platform, alongside its more traditional PC-based product. Indeed, when Facebook reported its results for the 2014 calendar year, these showed that revenues had jumped $12.47 billion, 57 percent higher than a year earlier, largely due to a sudden surge in mobile advertising.43

  But nobody inside the company felt complacent. The bigger the company became—and more successful—the more nervous they felt about the threat from nimble, small rivals. When Facebook conducted a successful IPO in late 2013, which valued the company at an eye-popping $100 billion, this expansion seemed to increase, not decrease, the unease. “I don’t think we are in any danger [of getting silos] in the next two years, but for all companies it’s a question of scale,” Goldfein observed. “But social problems are like computer problems: you have to keep asking if the same approach that handles 1,000 users be applied to 100,000.”

  There was another, more subtle challenge. One notable feature of Facebook was that its employees were homogenous. The engineers were mostly trained in the same set of computer skills and were in their twenties or early thirties. Most wore a common “uniform,” sneakers and jeans, and had a similar outlook on life. That made it easier to foster a common group identity and break down silos inside the company. Moving the engineers between teams was not so hard, because they were so similar. But as the employees inside Facebook became defined as a
social group, with a sense of their own identity, this created a new risk: in the future the company could end up acting like a gigantic social silo of its own.

  That danger was not unique to Facebook. The boom in Silicon Valley was giving birth to a wildly successfully technology elite. The wealthier these techies became, the greater their sense of confidence, if not arrogance. Like the bankers a decade before, the techies risked becoming detached from less successful mortals around them, not just in terms of wealth, but education and outlook. To anyone outside the tech world, it was almost impossible to understand exactly what computer engineers at places such as Facebook did. The algorithms were as impenetrable as financial jargon. And the techies sometimes lost perspective of what outsiders thought of them. There was a risk of becoming a techie ghetto.

  But the engineers who congregated in the orange conference room at Facebook during the hackathon on that warm May night were not too worried about those future dangers. The present seemed too exciting. They just wanted to write code, and push the frontiers of computing science, as the managers kept producing new ideas about social engineering. “We think we have a system that works, which breaks down silos,” said Keyani as he looked across the groups of engineers as they did their hackathon, huddled together in clusters around computers. “But we need to keep experimenting all the time.” It was the spirit of the “hacker way”—and perhaps the most crucial weapon in Facebook’s fight to avoid the fate of Sony.