Fighting Factions: How Startups Can Scale Without Mutiny
People & Culture

Fighting Factions: How Startups Can Scale Without Mutiny

Between them, eero's David Loftesness and BCG Digital Ventures' Alexander Grosse have held engineering leadership roles at Twitter, SoundCloud, Amazon/A9, issuu and Nokia. Throughout their careers, they’ve noted a silent force that can be particularly damaging to a fast-growing company: factions.

David Loftesness and Alexander Grosse know a thing or two about hypergrowth. You might say they wrote the book on it.

Between them, they’ve held engineering leadership roles at Twitter, SoundCloud, Amazon/A9, issuu and Nokia — in many cases, during periods when hot products demanded that those companies stretch in new ways, and fast. Now Grosse, Director of Engineering for BCG Digital Ventures, and Loftesness, the Head of Platform at eero, have lived through both the brilliant and bleak moments of scaling teams.

Throughout their careers, they’ve noted one problem that can be particularly damaging to a fast-growing company: factions. Because if you’re going to launch the proverbial rocket ship, you need everyone focused on the same destination—not wrestling over the controls.

In this exclusive interview, Loftesness and Grosse look at five areas where startups can either take action to deactivate destructive factionalism or even prevent them from forming in the first place. They share concrete processes for regaining the efficiency that leaders might not even realize they’re losing to competing mindsets and poor communication. Through these tactics, they show through specific scenarios why empathy is just as vital to a startup’s success as innovation.

The Distinction Between Teams and Factions

Business leaders have thrown out various rules of thumb for when a startup gets too large to function as a single unified team. Some say it’s around 50 people. Jeff Bezos famously prefers to stick to the smaller “two-pizza team.” But factionalism has its roots less in a hard number than in a shift in attitude. “Sometimes you have a ten-person team and you already have five factions,” says Grosse. “They may be divided on interpretations of company goals or different departmental philosophies. One thing’s certain: it can start very early, and, if unattended, it’ll get worse the larger the company gets. You have to zero in on it from the get-go.”

So what should startup leaders look out for? In general, any sign that single-minded collaboration toward a shared goal has been replaced by team- or department-specific priorities. Startups generally begin with a single tight-knit group, breaking into teams as they grow to provide focus around a particular problem domain or business goal. “But those teams are never independent of one another. There are always dependencies,” says Grosse. “If communication between those groups doesn't work, then you create a fertile ground for factions. You get factions when teams cannot seem to collaborate because of differences in vision, differences in working styles. Perhaps there’s even mistrust between the teams.”

Competition between teams isn’t necessarily unhealthy; the energy it breeds has fueled many a startup to success. It is, in many ways, a cousin of collaboration: when people are hustling to achieve some result faster and better, it’s because they’ve agreed on a shared goal. Competition, done intentionally and openly, lets a team pursue multiple approaches to their goal and more quickly identify the best one.

The dead giveaway of a faction is when people aren’t competing by the same rules—they’re playing their own game, whether they realize it or not.

“I define a faction as a subset of the company that starts to put its own priorities before the priorities of the company. It may form for a range of reasons: they've got a particular ax to grind with a colleague, a specific goal they believe is more critical than the current priority, or a distinct way that they want to do their work — and that becomes more important than building a great business. Then you’ve got a faction,” says Loftesness.

If your team is no longer working together toward a shared vision for the company, it’s time—past time, really—to do something about it. Here are five areas where positive action can prevent factions or mitigate them before they fester:

1. HIRING

“Everything begins with hiring. The seeds of factions start there,” says Loftesness. “The number one mistake companies make is not hiring people with values that are consistent with the founders’ vision.”

You hear a lot of talk about “culture fit” in Silicon Valley and beyond, but Loftesness has found that the term implies the wrong kind of homogeneity — and can actually discourage genuine collaboration. What you’re really looking for is a values fit. “I think that's a way that you can allow for broad-based hiring — hiring for differing backgrounds and perspectives — which can be helpful when you're trying to build an innovative product.”

For his favorite example of this principle in action, Loftesness leaves tech behind and considers that most famous team of rivals: President Abraham Lincoln’s cabinet. “In her book about this, Doris Kearns Goodwin describes how Lincoln convinced the rivals that he beat for the nomination to join his cabinet, knowing that they bitterly disagreed on key issues. Lincoln’s ability to use divergent perspectives to inform his decisions may have saved the country. Despite their differences of opinion, they all believed in democracy. They all believed in a government of the people, by the people, for the people. Nobody disagreed with that. What they disagreed about was how you manifest those principles in the specific practices of government.”

In all the hustle to find product/market fit or resonant branding, don’t forget to establish your company’s immutable values. And be precise; explicitly defining each one is part of predicting and preventing conflict. At eero, for example, ambition is a core value. “Ambition can express itself as getting to the next rung of a career ladder. Or it can be defined as bringing on people who are driven by new challenges in an entirely different space than where they were classically trained, which is part of how I ended up working on a hardware product with a software background,” says Loftesness. “Our founders were bold enough to jump into an entirely new area and successfully ship an innovative product, and that’s a quality we want to see in our new hires as well.”

The interview process is your first and best opportunity to prevent factions from forming. Learning how candidates — your future colleagues —  interact with people and how they handle conflict is key. Of course, few people will lead with “It’s my way or the highway!” in an interview. Here’s how Grosse and Loftesness recommend subtly sussing this faction-building mindset out:

Ask candidates about conflicts — note if they focus more on plot or perspectives. Even if you haven’t had time to fully flesh out your company values, hiring for empathy—the ability to see a colleague’s perspective, and understand needs that compete with one’s own — will serve you well.

There are a number of approaches to try. When interviewing engineers, Grosse likes to ask about a conflict they’ve had with the product team. “When it comes to tech debt, there's a common conflict between making things right and getting something out the door quickly to get feedback,” he says. “I just want to hear that engineering candidates understand the views and the needs of the people on the other side of that debate. I don't want to hear that they agreed with everything, but I want to know that they understand that it’s that balance that helps generate the best business outcomes.”

Folks that don't have empathy will talk about the surface-level detail. “They recount: here's what we’re fighting about, here's why I was right and they were wrong, here's how we resolved it. In those cases you'll often hear about escalations to management or some kind of power play,” says Loftesness. An empathetic thinker, on the other hand, will seek to understand and articulate the root of the conflict. “They'll be able to say, ‘Here's what I was feeling. And here's why the other person was feeling differently.’ Reaching this understanding requires empathy and is often the surest path to resolution."

Test an ability to empathize with one of Loftesness’ favorite interview questions: How do you know what you don't know, and what do you do when you realize that? “The intent is to determine whether candidates, particularly for leadership roles, know how to seek and incorporate outside perspectives,” says Loftesness. “Dig into specific examples. You want to hear a demonstrable willingness to hear fresh ideas — even if they don’t ultimately adopt them.”

Inquire about a specific type of collaboration: cross-disciplinary teamwork. “I'm passionate about cross-disciplinary teams, so I usually ask people how their teams worked at former companies. What cross-team interactions are they used to?” says Grosse. “This line of questioning is particularly critical for more senior hires, or anyone who’s spent their career in a corporate environment, where siloes can form. That’s not to say those candidates can’t thrive in a startup—but you need to ensure that they can and want to work with far less structure and hierarchy than they’re used to.”

To get to the heart of this question, ask: What would you improve, if you had complete freedom to change how your team was structured and worked?’ What you’re looking to see is if they naturally incorporate other functional areas and perspectives into their team structure or process. “Give extra credence to candidates who have an instinct for blending perspectives and teams to enhance their overall strength and scale back their weaknesses,” says Grosse. “For example, I remember one candidate who was a front-end engineer, who had taken UX courses to better communicate with the user experience team. While those disciplines share some common language, he wanted to be fluent in their ‘dialect.’ I was impressed by his drive to more precisely and effectively connect with another discipline.”

While these questions are important for any role, when it comes to preventing—or inadvertently enabling—factions, your first leadership hires are particularly influential. New managers often quickly take over the hiring process for their area, giving them an outsized role in setting your company’s tone. “There's a multiplicative effect. The people that they hire are likely to share whatever is most important to them,” says Loftesness. “If your company's hiring process is focused on experience or technical savvy or sales performance but ignores the candidate's core values, you're vulnerable. Your managers are free to hire people who check those checkboxes but have different values than the rest of the company. A penchant for cross-disciplinary collaboration is not a complete failsafe, but is a strong proxy for looking beyond one’s immediate scope of work. That’s crucial for growing early-stage teams into cohesive organizations.”

Hire for the company, not for the team. Tactically, there are a number of ways to make that happen, but they all share one technique: incorporate a variety of perspectives into your interviewing process. Any interview panel should include people from different teams, seniority levels, and disciplines. As your company grows and ages, interview panels should also include voices from both the old guard and the new.

Importantly, those outside voices shouldn’t just be there for show. Give them real sway. At SoundCloud, Grosse recalls, every interviewer had veto rights, with few exceptions. Many other companies have adopted the concept of an interview “bar raiser,” pioneered by Amazon.

“Amazon went through a period of recognizing that its hiring practices were not as thorough as they should be, which led to a lot of performance problems. As a response, leadership identified individuals who had been the most consistent in hiring the best people—and saying no to the wrong people—and were willing to stand up and be vocal,” Loftesness says. “Hiring managers are under pressure; by the time they have an open req, odds are good that they needed it filled yesterday. Infusing the hiring process with trusted interviewers who aren’t directly affected is the best way to ensure that long-term commitments aren’t sacrificed to short-term needs.”

2. ONBOARDING

Hiring for shared values is step one. Establishing an environment where every employee feels included and part of a larger whole is step two. “Onboarding is way more than a tour through the office, showing hires their new desk and the coffee machine.’” says Grosse. “It’s when you get the information and access to the people necessary to do a good job.”

Practically speaking, onboarding is also when new employees have the time to learn about the company, and meet colleagues they might not otherwise encounter. “This is the time to introduce them to each function — the data team, the tools team, the sales team, the product team, you name it — before they get buried in their day-to-day work,” says Loftesness. “Fostering an understanding of what each team does, what their challenges are — the basic act of putting names to faces after a meaningful interaction — is a great way to sidestep factions down the road.”

The specifics and demands of your business will dictate how much time you can spend on this interdepartmental “get-to-know-you.” But even a few half-hour meetings with colleagues in various departments can make a big difference. On the other end of the spectrum, formalized, week-long (or longer) rotations on different teams can establish even stronger bonds. Here’s how a few companies have approached onboarding:

  • “eero just launched a ‘shadowing program,’ where new colleagues from other departments get a day-in-the-life view of your job,” says Loftesness. “The sales team, for example, actually takes engineers, CX reps and managers out to a Best Buy, walking them through their retail training program. They see first-hand the end caps and how the product is positioned. They get to really experience the life of the sales team.”
  • At Amazon, where customer satisfaction is paramount, Loftesness, like all managers of a certain seniority, was required to spend several days training as a customer service rep. “The ‘Customer Connection’ training was an eye-opener, I’ve got to say. It was incredibly motivating, too. The problems you thought your team needed to focus on would shift when you actually saw what real customers’ frustrations were, and where there were opportunities to make them happier,” he says.
  • Issuu goes one step further than shadowing, and actually delays team assignments for engineers until they’ve completed a robust onboarding process. First, new hires spend a week with every team. “After those five weeks, I usually sit together with the new engineer and maybe two team leads, and we decide on a team for their placement. It’s always a compromise between business needs and the personal preference of the engineer,” says Grosse. “In some ways, this onboarding process is an extension of our hiring process: it only suits engineers who we want to attract — those who are open to a broad range of challenges. But regardless, it encourages deeper engagement in the onboarding process. When you assign the team before onboarding, new hires have less patience. Understandably, they want to contribute to their team and be successful. If they don’t know who their team is, they usually are more curious about the company, asking: ‘Hey, what’s here, what’s there, what’s interesting? Where can I contribute?’”

Once you grow beyond five or six teams, or start hiring employees at a faster clip, a rotation-based onboarding process may be impossible to scale due to the sheer number of teams. But while growth complicates certain onboarding practices, it facilitates others. “At a certain scale, you're growing so fast that you might have five to 10 or more employees starting in a given week. Take advantage of these new hire cohorts and help them to bond like a freshman class at college. Even without team rotations, they can use these personal connections to learn more about other teams. It’s a particularly effective way to formalize onboarding without sacrificing rigor,” says Loftesness. “Plus, those cohorts will likely end up in different departments, and they will form and facilitate connections and empathy. When people have friends in other parts of the company, that can be a really useful way of knitting those teams together over time.”

David Loftesness

Make room in your onboarding process for informational check-ins with key cultural gatekeepers, too. “Get new hires in a room with your veteran employees, for example, to maintain a thread to your earliest days. Encourage them to share stories, both difficult failures and energizing successes. This can give the new folks some perspective of what the old-timers went through to get the company to where it is today,” says Loftesness. “At Twitter, it was really common for a newbie to come in, look at some cruddy Ruby code in the Monorail and think, ‘Oh my god, what were these people doing?’ But once you’ve heard the origin story for that bit of code, how it saved the day at two in the morning, just before the Grammys, when the site was about to fall over, you’ll have a lot more empathy for what the old guard went through.”

3. TEAM TRANSFERS

Fostering an exchange of ideas and perspectives at the outset is just a start. For long-term sustainability, startups should champion a culture of mobility throughout an employee’s tenure. Grosse encourages leaders to not only put flexible transfer policies in place early, but also to actively encourage people to pursue them.

“In your one-on-ones, of course you want to encourage people to think about what they want to be doing. But for many, the conversation is restricted to what one can do on her existing team or later in her career. There are clearly other opportunities to discuss beyond those two focal points. What about roles on other teams at the same company?” says Grosse. “Try kickstarting this conversation by asking: ‘How excited are you with your work on our team? Are there projects on other teams that intrigue you?’ That’s key for any team rotation policy to actually work. If you’ve stacked your team with secure, values-motivated managers, everyone should see happy staff as a net positive.”

Introduce internal-transfer policies early — and don’t hire managers who focus on headcount.

Other times, internal transfers can actually cure what ails you. Loftesness recalls working with a team that built technology used by a number of internal clients. Many of those clients were frustrated with the quality of the work. “When it didn’t improve, we needed another way forward. It’s not always about a team’s lack of effort, but their lack of exposure to the challenge at hand. Sometimes describing the problem at regular intervals isn’t enough — you have to embed someone who’s living through it,” says Loftesness. “So, we actively encouraged a tech lead from one of those client teams to join and lead the technology team. Not all problems were resolved immediately, but overnight, the nature of the relationship changed.”

There was an infusion of empathy, as a former internal client saw firsthand the challenges that technology team faced — and communicated it to his former team on the other side. The struggling team, too, got direct insight into what the client needed. “What they needed was not just a new manager, but a translator — someone who could be fluent in both team’s realities. Leadership can use this technique to defuse tensions when they notice them occurring,” says Loftesness. “Actively rotating people through a team can really bring fresh perspective and a deeper understanding of the other side.”

Internal transfers may not be as simple as swapping personnel. In many cases, they also require additional training or education. But still, the investment is well worth it. Startups need to retain their star employees, and doing so might mean empowering them to pursue new interests. And that’s better within company walls than out the front door and someplace else.

Grosse recalls an infrastructure engineer at SoundCloud who grew particularly interested in machine learning and information retrieval. “So we gave him time every month to do Coursera courses. Then he switched to the discovery team, and brought all his infrastructure knowledge with him,” he says. “Last time I checked, he’d been with the team for four years and is very content.”

Alexander Grosse

Give employees the freedom to move and the time to learn, and you’ll reap many benefits. It’s much harder for factions to form when people are happy with their work environment — and know that they’ll be supported in making changes when they’re not. Don’t be afraid of those transfers — it’s easier for your people to trust and collaborate with another team if you’ve previously been a member of it. It’s just another manifestation of that key antidote to factionalism: put company over department.

If you’re nervous about disruption from team transfers, or if it’s too early in your company’s life to make that option available, there’s a lighter-weight place to start: cameos. Grosse pioneered the concept for the engineering team at SoundCloud, and it quickly became a cornerstone of the transparent internal communications that keep the multinational company humming.

“I got the idea from a talk by one of Stripe’s founders about their total transparency. Basically employees can go to management meetings — even board meetings,” says Grosse. “So I started opening the engineering leadership meeting to any engineer. Their only duty was to send a written summary of the meeting to the whole engineering team.”

Immediately, Grosse saw two benefits. The first and most obvious was a greater understanding of goals and process across the engineering org. The second was equally powerful: it forced leadership to think more carefully about how that meeting was run and ensure that it was something they wanted their team to see and hear. “To be honest, those meetings were kind of a shit show before that! Then suddenly we had a guest coming, and it was a completely different meeting.”

Don’t be afraid to think — and act — outside the box. If you have a hunch that something might bring teams together, try it. Loftesness recalls hearing a story from Marc Hedlund about his time as VP of Engineering at Sana Security. The leadership team was grappling with the age-old struggle of engineering versus sales, so they decided to try something new.

“The VP of sales delivered a typical sales pitch to a simulated customer in front of the full engineering team, with a Q&A session afterwards. The benefits were huge: It was obviously very rewarding for the engineers to speak up about the most outlandishly sales-y claims, and make sure the VP of sales wasn't selling vaporware,” says Loftesness. “But some of them also went back to their desks and started working on those crazy features. They saw that the VP of sales wasn’t pitching them to be a jerk — he just knew what was going to motivate current and potential customers.”

4. TECHNOLOGY SELECTION

Compared to big-picture concerns like hiring and career advancement, the bug tracker or workflow tools you use may seem like relatively minor details. But in Loftesness’ experience, those choices reflect bigger-picture dynamics at play.

Divergence of vision or divergence of values will often express itself in the form of already established brand tribes or allegiances. Like arguing about Emacs versus vi, or Trello versus Asana, or Java versus Scala. It may seem like it’s a trivial, surface thing. You might say, ‘Oh it’s fine if people want to use their tool, as long as it gets the job done.’ But I think it’s really important to look deeper and actually ask: does this represent a more fundamental difference?”

Twitter’s engineering organization, for example, spent years trying to bridge the gap between the Java and Scala camps. “One group felt like, ‘We don’t need fancy language features. We just need to be able to hire people that can understand our code and ship things for the business as quickly as possible with reliable tools,” says Loftesness. “The Scala camp would counter: ‘No, we actually care about certain language features and believe they will lead to higher-quality code. We're willing to put up with a little funkiness in our tools.’”

Sure, on one level it’s a classic engineering debate. But there’s a much deeper philosophical issue at play. “It raises the question: What's more important? Is it shipping code quickly, or shipping the right code? Those values-based questions have broad implications and letting technological divides deepen is a recipe for factions,” says Loftesness. “It took us years to come to a resolution. And in the meantime, teams grew and those divisions persisted and hardened throughout the company. It led to other conflicts between those groups — and even departures.”

To nip those ideological-based factions in the bud, seek out the torchbearers in the factions. The leaders may not be the most senior members in the factions, so look for exhibitions of influence. Once you find them, consider the following plan: “When two managers are at loggerheads, ask them to collaborate on an unrelated project. Perhaps you’ve been meaning to revise your hiring or promotion process. Ask them to step up and work together to get the job done,” says Loftesness. “Make it clear to both of them what the goals are — and that you expect them to set aside their differences for the good of the company. If they realize their success — or failure — is tied together, they have extra incentive to find common ground. And, most importantly, will take that new level of understanding back to their ‘factions’ once they’ve finished. That newfound empathy can dissolve factions — I’ve seen it.”

5. TEAM AND OFFICE STRUCTURE

Clearly, Grosse and Loftesness are big proponents of cross-functional teams, which they prefer to call “delivery teams” to emphasize that they’re aligned around a shared objective. “It means that a team should be composed of everyone who is needed to deliver a product, service or experience,” says Grosse.

"I see a lot of startups say, ‘Okay, we need to hire a lot of engineers because that’s how we’ll get stuff done.’ Developers get hired, get hired, get hired. And other roles — in product, marketing and design, for example — get neglected, become short-staffed and overloaded. I’ve seen this happen. When it does, you’ll be able to easily find your product manager. He’ll be the one running around like a headless chicken,” he says. “But building larger teams without the infrastructure to support them is not only counterproductive, it can also breed frustrated factions as engineering teams find themselves vying for the same resources.”

Don’t just add more bees to the hive. You can actually make your output worse by hiring more engineers.

In preparation for one particular release at SoundCloud, leadership asked him to hire more front-end engineers to boost productivity. But first, he went to his teams to ask what most kept slowing them down on a daily basis. “Everybody said, ‘We only have one designer. And every time I have a question about a design, it takes me seven days to get an answer.’ I realized that hiring more engineers would actually make the problem worse. Instead, I needed to hire more designers.”

Building cross-functional, product-oriented teams means supplying the resources they need, but also demanding that they take full ownership of their deliverables. “When I joined issuu, one of the first things I did was dissolve the QA department to improve quality. I know that sounds counterintuitive,” says Grosse. “There was a little chaos for a week or two, and then teams embraced it and quality improved. They realized that we had to own our own quality — a key value for our team — especially since we didn't have another team to rely on.’”

Embedding all the resources a product demands in a single team — engineering, marketing, operations, and more — fosters resilience and, in most cases, best supports scaling efforts. “When you keep disciplines separate, then you have battles between sales and marketing, or development and QA. Without collaboration, these groups end up just throwing things over the wall to each other,” says Loftesness. “But if they're actually sitting together and working on a common problem, that’s less likely to happen.”

As companies grow, the very notion of “sitting together” can pose logistical problems. But this isn’t a reason to abandon cross-functional teams. When it comes time to add a second location, Grosse and Loftesness urge leaders to think carefully about who goes where. “At one company, we had a problem for a long time because all of management was in San Francisco, and we had an engineering office in Europe, nine hours away,” says Grosse. “Obviously it was very difficult for each office to understand what the other needed, because they were so distant from each other. That created all kinds of conflicts.”

It’s an increasingly common problem, particularly as startups seek to find untapped, and often lower-cost, pools of engineering talent. But in Grosse’s experience, whatever you might save isn’t worth the loss of efficiency to confusion and strained personal connection. So he made a simple rule: one product, one office. This meant that all the resources needed to deliver that product were based in that office and could be grouped into cross-functional “delivery teams.”

It’s a common factions fail case. Different locations battling. Or the mother ship versus remote offices.

Loftesness lived through the challenges of multiple locations during his time at Amazon, when his Palo Alto office found itself at odds with a team out of the company’s Seattle headquarters. “My team owned all of product search for Amazon. But there was a team in Seattle — they were very creative and innovative — that had identified a way to use personalization technology to actually make search better,” The trouble was, they put innovation above all else, and started running A/B tests on the search pages that we were responsible for. We were getting paged when there were bugs in the middle of the night. So that definitely exacerbated the tension. Both teams got stuck in their way of thinking — innovation versus reliability — and wasted energy looking for the other team’s fail cases to make their points.”

Ultimately, though, those A/B tests were proving that the new technology could have a big impact on Amazon’s user experience — and its bottom line. “We started calming things down by having a weekly call to share notes and projects, which began a little bit of healing,” says Loftesness. “But both sides ultimately agreed that the new technology needed to be transferred to the search team so we could incorporate it into our technology roadmap. This would avoid both technical and human conflicts, and the other team could focus on new innovations.”

It wasn’t a perfect resolution, but it reversed intensifying factionalism. “The tension stemming from location differences and the distinct styles of each group were problematic,” says Loftesness. “Not everybody in the end was totally happy. But the conflict was gone — and with it the distractions and the inefficiencies that came from fighting back and forth. It allowed each team to pursue the goals of the company more efficiently, which everyone agreed was far more important than our petty squabbles.”

Fewer Factions, Less Friction

Early-stage startups are as prone to factions as large companies — it’s less about headcount and more about getting ahead of them. To keep them at bay, internalize the difference between teams and factions. Take steps at gateway experiences — hiring, onboarding, transfers, technology selection and team or office growth — to prize cross-disciplinary collaboration, functional immersion, career mobility and the values underpinning different tools and offices. But, even with all these efforts, there’s an important truth: you may not solve every factional struggle with an outcome that thrills everyone. But if you can implement processes in these five areas and avoid issues before they arise, so much the better. Even in those earliest, breakneck days, it’s well worth your time.

“I liken it to coding. You try to catch bugs as early in the process as you can, because it costs much less then than it does in production, when it's already affected your customers,” says Loftesness. “The same is true of any cultural problem. Maybe you hired the wrong people — or you’ve allowed members of your team to rally around an unhealthy theme or practice. Unwinding that is going to be much harder than taking a small step early on to counteract factions. You owe that to your people.”

Photography courtesy of the subjects. Art by wildpixel/Getty Images Plus/Getty Images.