Anthropology of Mid-Sized Startups

Guest post by Kevin Simler, who works at Palantir, observes the startup scene, and writes at Melting Asphalt, about… well, go see for yourself.

In their natural habitats, social species organize into characteristic groups. Gazelles form herds, wolves form packs, and ants form colonies. Humans, in the same way, form tribes.

Of course, we’re pretty far removed from our natural habitat these days. But tribes are a large and fundamental part of our evolutionary heritage, and they have a corresponding influence on our mental and social lives. Organizing ourselves into tribes is one of the ways we manufacture normalcy. It helps our paleolithic minds perceive and act, more or less sensibly, in an increasingly complex modern world.

Humans also form kingdoms, nations, states, and civilizations, but those units of organizations aren’t as fundamental to our psychology. Statecraft is an esoteric enterprise; we spend most of our cycles processing social data at the tribal scale. Even Kissinger, for all his mastery of foreign relations, had to play tribe-level politics in the White House and State Department.

Tribal psychology dominates in mid-sized organizations — between 10 and 1000 employees, let’s say. These two numbers flank Dunbar’s number (roughly 150) by an order of magnitude on each side. Dunbar’s number is the “cognitive limit to the number of people with whom one can maintain stable social relationships,” and serves as a good midpoint for the size of organizations we’re going to look at. Below 10 employees, all you have is a simple working group. Above 1000 you’ll start to need “more restrictive rules, laws, and enforced norms to maintain a stable, cohesive group.” Between those two is where the tribe lies.

So let’s see what happens when we treat startups as tribes.

To do that, we’ll need to use the methods of anthropology rather than business analysis. I hope to show that this is a productive lens for understanding startups.[1] It’s not the only lens, of course. It can’t tell us much about the very early days of a company, or about its strategy for going to market. But it can tell us a lot about internal dynamics during the middle stages of growth — after a startup has its foot in the market but before an IPO or acquisition.

Startups as tribes is a useful shift in perspective, I think, because we typically think about startups with a more technical mindset. Startups, especially in tech/software, are notoriously full of left-brained engineering types, people who (myself included) specialize in decontextualized reasoning about formal systems. But if you spend too much time in that mindset, you’re liable to forget that a startup is an all too human enterprise. We aren’t just networked brains exchanging packets between vats. We are embodied creatures, situated in the world, running on evolved wetware. To fully appreciate what goes on inside a growing startup, it pays to remember that an engineer is also a primate.

I’m writing about the world of startups because that’s where I’ve been doing amateur ethnography for the past 7 years. But these concepts should apply just as well to any mid-sized organization or workplace.

Startups in culture space

Most mid-sized startups talk about culture, have a culture page on their websites, and screen for cultural fit. Some have comprehensive “bootcamps” where new employees learn about the culture. And many of the world’s most successful institutions are famously obsessive about it: Apple (largest public company), Bridgewater (largest hedge fund), Netflix, Valve, and Zappos, to name a few. Their cultural guidebooks (bibles?) make for fascinating literature.[2]

Let’s situate ourselves by asking: what region do startups inhabit in the space of all actual (or possible) cultures? One of the more useful tools for doing this, I’ve found, is cultural dimensions theory.[3] This will help locate startups in culture-space, but it won’t help us understand the culture of any particular startup; we’ll get to that in a minute.

Cultural dimensions theory was first explored by Geert Hofstede at IBM in the 1960s and 70s and later in his book Cultures and Organizations: Software of the Mind. Using a survey of IBM’s massive worldwide workforce and some statistics (factor analysis), Hofstede identified the four (and later, five) most significant dimensions along which different cultures vary:

• Power distance: The extent to which the less powerful members of organizations and institutions accept and expect that power is distributed unequally.
• Collectivism (vs. individualism): The degree to which individuals are integrated into groups.
• Uncertainty avoidance: The extent to which the members of a culture feel threatened by ambiguous or unknown situations.
• Masculinity vs. femininity: The distribution of emotional roles between the genders. Masculine cultures value competitiveness, assertiveness, materialism, ambition and power, whereas feminine cultures place more value on relationships and quality of life. (Sometimes this dimension is neutered to Quantity of life vs. Quality of life.)
• Long-term orientation: The degree to which a culture fosters virtues oriented toward future rewards — especially perseverance and thrift. Fun fact: this dimension was initially termed Confucian work dynamism.

Let’s look at a few of these dimensions that are especially important for delineating ‘startup culture.’

Startups have a very high tolerance for uncertainty (low uncertainty avoidance). It might be more accurate to speak of an appetite for uncertainty, since a startup by definition doesn’t understand its own business model and must explore, both in the market and in design space. There’s a frontier spirit to a startup, evidenced by phrases like pragmatic, risk-taking, pivot, and fail early, fail often. As it grows and ages, a startup will become more cautious. It will need to protect its (increasingly valuable) resources, and the spirit of risk-taking that drove the initial growth will be in danger of fading. The company will also need to become more legible (with titles, policies, rules, and structure) in order to manage its increasing complexity.

Startups have low power-distances. In Silicon Valley, you’ll hear buzzwords like flat, transparent, and meritocratic, or “titles don’t matter here.” But flat refers to more than just the number of layers in the hierarchy — it also means you can criticize your boss’s ideas, or hang out with your boss’s boss after work. Having a flat culture is important, especially in the face of uncertainty, for making sure information flows ‘uphill’ as freely as possible. As it grows and ages, power distance will increase. The company will tend to become more status conscious[4] and less transparent. This doesn’t necessarily spell the demise of meritocracy, but it will require more special attention.

Startups are highly collective enterprises, aka tight-knit teams. Solidarity is absolutely essential to success. If you aren’t all rowing together, your ship is liable to go nowhere. Luckily, there are a lot of factors keeping the collective spirit high, including the compensation structure (stock options), external threats, and having a homogeneous culture. As it grows and ages, a startup tends to become more individualistic. When everyone’s working as a tight-knit team, it’s easy to spot who isn’t pulling their weight, or who’s looking to advance their own agenda at the expense of the company’s. But once the atmosphere begins to rarefy, free-riders and politicians have room to make a home. Turf wars will become more common. And the dilution of shared ownership doesn’t help either. As we’ll see later, solidarity rituals are one of the important ways a startup can maintain a collective orientation in the face of growth.

Analyzing a single culture

If dimensions are useful for comparing cultures, folkways can be used to analyze a single culture in depth, largely on its own terms.

The concept of a folkway comes to us by way of Albion’s Seed by David Hackett Fischer, which analyzes the four (originally British) folkways that dominate American society. What’s interesting here isn’t the precise definition of folkway — we can just treat it as a synonym for culture — but rather than method Fischer uses to analyze it.

In Albion’s Seed, he breaks each of his four cultures into a couple dozen component ‘ways’. Examples of these components include speech ways (how people use language), dress ways (how they dress), time ways (how they think about time), and rank ways (how people are assorted into ranks). The idea is that once you’ve described a society through each of its components, you will have essentially sequenced its genome.

Here’s a suggestive list of what might constitute a startup folkway, along with a couple questions to illustrate each item. Keep in mind that what’s important here is the exercise (of breaking a culture down), not the particular components per se.

• Rank ways. How does rank affect the amount of influence a person has? How are promotion decisions made?
• Time ways. How are product cycles broken down? What activities take place on weekly, monthly, quarterly, and yearly cadences?
• Office ways. How are desks and offices arranged/assigned? [5] What are the core working hours, noise levels, interruption norms, etc.?
• Meeting ways. How, when, and why are meetings initiated? Who talks, who takes notes, and who’s distracted by their phone or laptop?
• Email ways. What kinds of decisions are made over email vs. in meetings? How are mailing lists used?
• Decision-making ways. How are important decisions made? Democratically? Autocratically?
• Compensation ways. How is compensation divided between bonuses, salary, and equity? How much does it vary by rank, seniority, and job function, versus by achievement?
• Work/life balance ways. I’ll just point here.
• Hiring ways and firing ways. How are hiring decisions made, and by whom? How are bad hires identified and weeded out?
• Project management ways. How are the goals for a project determined? How is success measured?
• Coding ways. How is code quality enforced? How territorial are people about code ownership?
• … and many others (celebration ways, onboarding ways, info dissemination ways, planning ways, customer-relation ways, media ways, etc.)

Basically, this is the set of questions that new employees should be obsessing over during their first few weeks/months on the job — or better yet, while they’re interviewing and negotiating an offer.

Acquiring this cultural knowledge is hugely important, but also difficult, largely because of how implicit and undocumented all these ‘ways’ are. Candidates and new employees need to triangulate, use subtle social cues, and more generally engage in participant observation in order to become fluent in the culture of a mid-sized startup.

One thing that’s great about folkways analysis is that, as you’re breaking a culture into its component parts, you begin to see just how interlocking those parts are. Deeper patterns start to emerge, and you pretty quickly converge on values. If a startup is truly a meritocracy, this will be written all over its office ways, meeting ways, rank ways, compensation ways, and hiring and firing ways. Whether a startup values work/life balance will show up in its time ways, office ways, and compensation ways.

Rituals

When thinking about rituals, it’s easy to imagine something like a raindance — a purely symbolic performance with no causal relationship to its intended effect (rain). But this is dangerously misleading. Most rituals are effective. They may use symbols, but only as a way of mediating some real effect, usually in the social world. For a canonical example, don’t think raindance, think giving a gift.

As in all human cultures, rituals are pervasive in startups, and they serve a number of purposes. They help reinforce the hierarchy (1:1s, staff meetings), define group membership (badges, interview gauntlet as initiation ritual), reinforce shared values and goals (public shaming/celebrations, hackathons), and build trust.

Building trust and enhancing solidarity (i.e. collectivism) is the purpose behind an especially rich set of rituals. Most companies do the occasional happy hour, holiday party, or company retreat. Startups go the extra mile. Here are some of the many rituals I’ve seen startups use to reinforce trust, solidarity, and company (or team) pride:

• Branded apparel. T-shirts, jackets, or hoodies printed with the company logo or an inside joke. One company I know sports bright neon v-neck workout shirts. It functions as a great shibboleth.
• Team sports. Most companies play either pick-up games or in intramural leagues. Spend a couple hours on the court or field (or yoga studio) with your teammates, and you’ll feel a lot closer back in the office as well.
• Themed dress-up days. These are days when employees decide to wear something out of the ordinary. Suits (worn ironically of course) are especially popular. I’ve also seen pajamas, wacky-hats, and mismatched clothing. Wearing something distinctive and synchronizing with your teammates is an especially powerful way of signalling, “We are all in the same tribe.”
• Shared meals. Free food is pretty common among startups, and it goes beyond just saving employees a trip to a local restaurant. Communal meals — from Spartan syssitia to potlucks to college dining halls — are a time-honored bonding ritual.
• Wall art. Going ‘tribal’ on the walls — by hanging posters, painting a mural, or just writing on it — is another classic ritual used to signal a collective spirit in the office.
• Parties and other extracurricular activities. Examples include movie nights, concerts, sporting events, progressive dinner parties, and book discussion groups. As one of my teammates put it, “You don’t discuss books with your coworkers; you discuss them with your friends.”

It’s tempting to trivialize these activities as merely for “morale,” but they accomplish far more than just keeping people happy. They help people identify with the team. They create emotional bonds and social relationships. And they can actually change the communication topology of the company — which, by virtue of Conway’s law, may also have an effect on the product a startup is building.

Tribal bonding rituals can even help mitigate politics. One of the more pernicious disease states a growing organization will face is when employees start treating other teams as a homogeneous out-group they. “IT doesn’t do anything” or “Engineering doesn’t care about the customer.” But this tendency can be softened just by knowing the name and face of someone in that ‘outside’ group, which can happen during dinner, while playing sports, or at a party.

The fact that most rituals have a strong physical component is one of the reasons remote workers are difficult — even if the physical component is just being in the same room together and watching each other’s facial expressions, eye contact, and body language.

Startups as religions

In case startups as tribes isn’t enough of a refactor for this audience, let me propose a more radical one: startups as religions.

The familiar form of this metaphor says that organizations can be cults. Witness Apple’s cult of Jobs, the cult of Bridgewater, “drinking the kool-aid,” etc. But while it’s fun to use the derogatory term (cult), it’s a missed opportunity to reference a much richer concept (religion).

There are a lot of different directions we could take this metaphor:

• Founders and other top leaders as cult figures
• Middle managers as priests (tending to the flock, hearing confessions during 1:1s?)
• How the transition from startup to big company mirrors the stages of religious evolution
• Mission statements and corporate values as an articulation of shared (quasi-)religious beliefs (e.g. “Don’t be evil.”)
• How religion can become an instrument of social control by those in power
• Evangelism and recruiting
• Rituals of sacrifice
• World-rejection as a driving force
• Founding story as creation myth

And so on. Startups can learn a lot from the sociology of religion. Unfortunately we only have time to scratch the surface here, so I’d like to look at religion’s role in caring for the collective interest.

Religions have endured for millennia because they are very good at what they do: uniting individuals into a community with shared values, goals, and culture. And if you think that religious beliefs are mostly nonsense, you should be even more impressed by what religions have been able to accomplish.

Émile Durkheim understood religion not as belief in the supernatural (as the popular conception would have it), but as a system of beliefs and practices relating to sacred things. Something is sacred, according to Durkheim, when it represents the interests of the group. Everything else is profane, i.e., unrelated to the group’s interests.

Like a religion, a startup will care for its collective interest by defining certain things as sacred. A classic example is the company’s logo. This symbol is, quite literally, “set apart and forbidden” by brand guidelines, which often specify exactly how the logo must be presented and how far it should sit from the other elements on a page (thus separating the sacred from the profane).

A startup may choose to sacralize other things related to its collective interest, such as team rituals or hiring and firing (determining who gets to be a member of the community). Or it may choose to treat certain values as sacred, such as code quality, product quality, or customer satisfaction. This kind of thinking is a bit dangerous as it tends to ignore the existence of tradeoffs (sacred things are considered inviolable), but it can be valuable when used judiciously.

The religious perspective also sheds light on why solidarity rituals are so important, and so prevalent, at a growing startup:

“Religions often maintain intragroup solidarity by requiring costly behavioral patterns of group members. The performance of these costly behaviors signals commitment and loyalty to the group and the beliefs of its members. Thus, trust is enhanced among group members, which enables them to minimize costly monitoring mechanisms that are otherwise necessary to overcome the free-rider problems that typically plague collective pursuits.” [From Signalling, Solidarity, and the Sacred, the best 10 pages I’ve read all year.]

Religion is, of course, a tricky proposition. If you take inspiration from religion’s playbook, just make sure it’s not the page on orthodoxy. Once your community decides that some beliefs are wrong (and punishable!), all hope for acquiring knowledge is lost. To succeed in the risky, uncertain world a startup inhabits, you need a community of free-thinkers, not true-believers.

Endnotes

[1] I’m not the first to identify “tribes” as a useful metaphor. Seth Godin (Tribes) and Dave Logan (Tribal Leadership), for example, have built whole books around the term. I’m just pointing out that the formal study of tribes, i.e. anthropology, can shed light on what goes on at a startup.

[2] Cultural guidebooks: Bridgewater, Netflix, Valve, and Zappos (see also Delivering Happiness).

[3] There are many other ways to slice up cultural space, for those who enjoy such things. E.g.:

• Steve Yegge on liberal vs. conservative engineering practices
• Edward T Hall’s high context vs. low context cultures
• Robert Cringely’s commandos vs. infantry vs. police

[4] Be careful with status symbols. Status symbols create envy, making it more difficult to maintain a flat, collective culture. Visible status symbols are especially tricky. Examples include titles, desk location, parking spots, and office equipment (fancy desks or chairs, size and number of monitors).

[5] Desk location is critical. A study done on Google’s internal prediction markets “found that [knowledge and opinions] were most correlated among employees sharing an office, that correlations declined with distance for employees on the same floor of a building, and that employees on different floors of the same building were no more correlated than employees in different cities.” If you want people working together as seamlessly as possible, pack them in.

[image credits] Google whiteboard (Scott Beale / Laughing Squid); Mark Zuckerberg and David Choe (Wang Newton); cave paintings (Wikipedia).