Community at Work
Isolation of excellent people makes a company less than the sum of its parts.
It’s the age old tale of a successful startup.
First, the fail-fast board mandates the company to hire. The company goes on a hiring spree. Some of the people that come in are truly fantastic: thanks to the new hiring power, the company can attract some of the best. But the company also does the reverse — it hires a number of people who are mediocre at the job. As managers look to deliver results, they’ll immediately realize they need some of the best people to deliver progress. They’ll create little islands of excellence, where the best people are surrounded by mediocre performers as a way of ensuring some progress on a team. Leaders will scratch their heads and ask “why are we doing so much less with so many people? Why does the organization feel so bloated?” And the great people, burned out from carrying a team and feeling lonely, will leave. My colleague Michelle wrote about this exact phenomenon within Stripe in 2019. After a number of her colleagues left her team, she found herself isolated. Ultimately, she found a great collaborator in the nick of time. But this saving grace felt too much like happenstance and not enough like the company was aware of this issue. Why wasn’t building community for excellent employees higher on management’s priority list?
Isolation of excellent people makes a company less than the sum of its parts. A lone aspiring vocateur will (usually) burn out or produce lower quality work. A network of great people around them is key to improving their happiness, idea quality, longevity, and productivity. Vocateurs need to be emotionally committed to their work – so committed that even their subconscious is chewing on work all the time. Without a community, it’s a lonely path. Excellent workers need community to survive.
Its upside is huge. Community is another form of runway. It makes companies last longer than they should have otherwise and buys time for excellent ideas to emerge. DabbleDB demonstrates the upside: it was created by Avi Bryant and ultimately acquired by Twitter in 2010. The team was so good that their intern was Patrick Collison. Ex-Dabble people identify as a “Dabble person” a decade later, despite their storied careers. Dabble was ahead of its time in many ways, but Community was a big part of why it was great.
Why do companies try to scale headcount so quickly? Why not patiently build community of excellent people instead? In my opinion, this type of “scale fast” approach to hiring is one of Google’s worst exports to the rest of Silicon Valley. It comes from the world where VCs want every business to assume they’re a Google-like outlier until they fail. In this world, ex-Google execs assume a meteoric level of business growth, require a shockingly high headcount growth, and founders fall prey to the same exceptionalism. But it doesn’t have to be that way.
Instead of scaling fast at all costs, businesses should invest in creating community. Here’s what creating great work community requires:
Enable complementarity
Maintain density
Create norms, not rules
Complementarity
Here’s a controversial statement: good company culture has some homogeneity. The dimensions of homogeneity will make or break the culture. People within the community must have something important in common which distinguishes them sharply from the rest of the world.
After you’ve determined the relevant points of homogeneity, almost everything else can be heterogeneous. Those differences are equally relevant. You want the teams’ talents to complement each other, you want their viewpoints to bring diversity of thought, and you want more user empathy in the room. When you build a product for a new market, you want a representative of that market building product.
The balance between homogeneity and heterogeneity is what I’m calling “complementarity.” Finding those key points of homogeneity and enforcing them over time is the hardest challenge and so subject to taste. It should have nothing to do with everyone liking Squid Game, or being the same age. Instead, the dimensions are fairly non-intuitive. You have to carefully consider their work habits, talents, and values. For example, the differences in working style – a really direct debater with quiet and slower thinkers – might never work, no matter how hard you try. Ultimately, the goal of complementarity is to have a group of people who will push each others’ ideas and attack a challenge from all angles, but collaborate with one another, push towards the same mission, and resolve conflict productively.
At Stripe, my colleagues Michelle and Sebastian had a set of very specific criteria for “product shapers,” the people with a set of shared values about product building. These values included the importance of skills like being relentless “engines of progress” and “anthropologists,” or even the sense for what would make a beautiful and surprisingly great product experience.
At Pixar, this went even further. Ed Catmull describes the following with respect to team composition:
Getting the team right is the necessary precursor to getting the ideas right. It is easy to say you want talented people, and you do, but the way those people interact with one another is the real key. Even the smartest people can form an ineffective team if they are mismatched. That means it is better to focus on how a team is performing, not on the talents of the individuals within it. A good team is made up of people who complement each other. There is an important principle here that may seem obvious, yet — in my experience — is not obvious at all. Getting the right people and the right chemistry is more important than getting the right idea.
Density
So, let’s say you’ve figured out the complementarity problem at the early stage. You’ve started a company and hired a small group of early people to meet the challenge. You’ve been extremely lucky – all of these people are vocateurs who “get it”, work hard, enjoy working together, and create positive externalities. Doing a startup requires constantly doing something new, and the teamwork allows them to tackle these challenges together. Their strong communication enables them to share new information and identify and close gaps quickly. They talk to one another and stay in sync.
At some point, the company needs a few more people to grow – you want to build some new products. You hire a few more people and segment them into teams. You are the hiring manager yourself, and everyone’s still excellent.
But nobody’s perfect. At some point, you’ve hired someone who is “just ok.” And that person happens to be a manager, or influential on a hiring panel – they hire a few more ok people as well. You’re a startup, and the company is now filled with mostly good people with a few ok people, here and there. They’re not bad enough to fire – they’re just mediocre.
Over time, the number of mediocre people grows. It quickly grows out of control. It’s harder and harder to hire people who satisfy your complementarity requirements – maybe you just haven’t emphasized their importance to every single hiring manager. Soon enough, fewer and fewer people share the qualities you want as a proportion of the total company.
When that happens, organizations usually isolate the stars so that teams can remain productive. It’s like a bad version of the NBA draft: they send a great player to every team, starting from the worst teams, and the great players end up isolated and alone. The one great player is rarely enough to carry the team successfully, and may burn out or lose their ambition before another great player can be added. It makes the work less fun.
Ultimately, this is terrible for the great employees. To extend my NBA metaphor, you need a Scottie AND a Michael and it’s even better when you have a Dennis. By virtue of being on a team with those guys, you’ll elevate a relatively unremarkable Steve Kerr to greatness.
Without the team – you end up with MJ on the Wizards (less successful!)
The drawback of concentrating A players is that it does leave the rest of your organization a bit bereft. But to be honest, you’re rate limited by your number of A players anyways. You might as well concentrate them, and get work done more quickly!
But…I can avoid this, you might say. I’ll only hire A players and I’ll fire anyone who isn’t an A if I’ve made a mistake! And I don’t need to scale my company dramatically – I can have 230 people run the whole thing, I’ve seen it happen.
Ah, sweet Summer Child. There are solutions to this problem. But they all are subject to tradeoffs, thanks to the Universal Laws of Hiring:
No hiring or perf process is perfect. They can be good, but not perfect.
Price and timeline are meaningful constraints on hiring A players.
Mediocrity spreads like cancer.
No firing process is perfect.
You’ll never be able to perfectly hire, or perfectly weed out the mediocre performers, or even get an A player every time (they may be way too expensive!) So, given these laws, what are the solutions?
Facebook: Facebook allows people to leave and join different teams on their own volition. It means the best people ideally end up clustered on the best teams with the most popular managers. It doesn’t necessarily mean all the mediocre people are fired – Facebook needs low cost engineers working on problems. Its policy certainly promotes density, but the tradeoff is that there’s a strong bias towards sexy, performance-system-rewarding problems, which isn’t necessarily the same as the most important problems.
Apple: Apple has a strict tiering between (paraphrasing) “gods” and “just folks” within the company. Gods, namely executives and product designers, can do whatever they want. Folks must do their bidding, and aren’t allowed to really make decisions. It’s command-and-control to ensure excellence, and mediocrity is tolerated in specific places. The tradeoff is the classic tradeoff of command and control organizations: you allow for very little bottoms up innovation. If you’re “just folks” and have a good idea, will it be heard? Does it…matter?
Google: Ah, Google. When I was looking for my first job, it was unequivocally considered the best place for smart people. It was strongly biased towards a very rigorous interview process and a strong reputation in the Valley. It meant that the first chunk of Google’s history – say, till 2015 or so – was extraordinarily great because Google had excellent people coming through the pipeline and passing interviews. However, Google’s interviews (much like SEO!) became gameable, and the reputation faded. I believe Google’s in a bit of a quagmire ever since. The current reputation is that “Google is where mid people go to retire.”
Netflix: The classic “A player” culture. It comes the closest to the “fire well” ideal, with great severance packages that go to anyone who can’t cut it. It’s a scary culture with a lot of turnover and perhaps no psych safety. I imagine you can’t have that type of turnover at the leadership level (who keeps state in that situation?) I don’t know what it looks like on the inside, and I don’t know how they handle cost, but the numbers look good – CEO approval is high, the company is growing, and the company’s financials are doing well.
Create norms, not rules
Most companies love making a big formal set of rules for how people should engage. I think about norms in the context of a work community as the better alternative to rules. Unlike brittle rules that are meant to constrain behavior in a relatively static environment, a set of norms are malleable and dynamic. Communities need norms so that they can create cohesion, but aren’t limited by rules that are irrelevant to changing circumstances.
Norms are translations of shared values into more practical manifestations. For example, if there was a shared value of pursuing excellence, norms could include things like writing rigorous documents defending a particular position. If there was a value of merit over status, a norm could include junior people taking the helm of major projects.
Richard Hamming (can I work in a quote from him in every newsletter?) describes the norms of Bell Labs as the following: At Bell Labs everyone expected good work from me - it was a big help. Everybody expects you to do a good job, so you do, if you've got pride. I think it's very valuable to have first-class people around. I sought out the best people. The moment that physics table lost the best people, I left. The moment I saw that the same was true of the chemistry table, I left. I tried to go with people who had great ability so I could learn from them and who would expect great results out of me.
The problem with norms is that they can be a bit imperceptible. Norms require time to learn and understand, while rules can be set in stone. Norms can also (when they become memes!) get weaponized. We’ve all seen examples of cultural norms like an “obligation to dissent” (McKinsey) turn into excuses for the loudest man in the room to dominate the conversation.
But their value is that they are much more resilient than any rule, and create a solid foundation for collaboration amongst the community. Once learned, they can be carefully pruned and adapted in a more elegant way than the very explicit repeal or revision of a rule.
Remote work
The elephant in the room here is remote work: can community exist when remote work is on the rise? Nicholas Bloom, in last week’s Economist, believes that we will see a “Nike swoosh” of remote work’s revival. His take is that the minor depression in remote work post-pandemic is just crusty executives demanding in-person offices because it’s what they’re used to. Many people, sad new grads included, believe that community formation is almost impossible in a remote environment. If community is so essential, how can it survive in a world where remote work is contributing to 1.5% productivity gains over the last year?
If you’ve had close friendships formed over the Internet, you know it’s possible. But remote community’s going to take a whole new set of tactics that seem as unconventional to us as Patio11’s internet forum tactics felt to traditional event organizers.
Look – I hate Zoom happy hours as much as the next person. Yet, it has to be possible to create a community remotely because remote work holds so much hope for the future of labor.
As Claudia Goldin (who won the Nobel Prize in Economics this week!) researched, the key to advancing gender equality is to make work more flexible:
But what must the "last" chapter contain for there to be equality in the labor market?...It must involve changes in the labor market, especially how jobs are structured and remunerated to enhance temporal flexibility. The gender gap in pay would be considerably reduced and might vanish altogether if firms did not have an incentive to disproportionately reward individuals who labored long hours and worked particular hours. Such change has taken off in various sectors, such as technology, science, and health, but is less apparent in the corporate, financial, and legal worlds.
Searching for these tactics feels like the next most important management challenge — what’s it going to be? If someone sends me another remote cocktail kit, I’ll scream.
Early in my career my team got to work on an Apache project on behalf of our employer. Folks from the foundation came in and gave us training on their norms and best practices—to the effect of "the way you work at your company won't be successful here, you need to explicitly change how you work in order to be part of this project". We got copies of https://www.amazon.com/Producing-Open-Source-Software-Successful/dp/0596007590 and were expected to read them.
Working on that project was great: people could be productive in any geographic location, organizational hierarchies were somewhat meaningless (we were working collaboratively with competitive companies), there were very few meetings (the occasional in person onsite) and very little "I don't understand the strategy/roadmap/vision/etc" whinging.
Not that every open source project is successful or has a healthy community, but I've been a little surprised/curious that this model doesn't seem to get adopted for significant internal projects. I wonder if it's that most people come into command-and-control environments early in their careers and don't get the chance to see other patterns of working. Have you come across any references of companies where teams follow "chaordic" (RIP Dee Hock) internal development processes?