Return to insights

Insights from "Social Chemistry" and "Friends"

Seeking respite from the markets, I’ve just read Marissa King’s ‘Social Chemistry’ and Robin Dunbar’s new book ‘Friends’, which was published last week.

Both books discuss something incredibly important and relevant to every single one of us: friendship and social dynamics. Dunbar’s book is focused on explaining the evolution of friendship and its benefits. King’s book is more about the structure of social networks and how to improve the quality of your friendships. As a result I found her book more interesting and practical.

As is blindingly obvious, Dunbar found that robust social groups correlate remarkably powerfully with everything that’s important (health, happiness and lifespan for a start). And yet they dwindle with age, especially after parenthood. We often don’t give them the right kind of, ahem, attention.

Why this is interesting: full disclosure, I am obsessed with Dunbar’s number. In a world where people seem positively gleeful about identifying society’s many problems, I think Dunbar’s number is a foundational part of any potential solution.

If there’s a single feature that’s common to successful and resilient modern socioeconomic and business structures it’s community. Dunbar’s number is an extremely famous, but quite robust, finding that humans find it cognitively demanding to maintain more than 150 social relationships. More specifically, our tightest circle has just five people. This is complemented by the “five chimps theory” that you’re most influenced by the 5 people you spend the most time with. That most intimate circle is followed by successive layers in multiples of 3x: 15 (good friends), 50 (friends), 150 (meaningful contacts), 500 (acquaintances) and 1,500 (people you can recognize). This relationship has held constant, even now many of our friendships have been digitized.

Stephen Fry quipped that the right way to think about Dunbar’s number is: “the number of people you would not hesitate to go and sit with if you happened to see them at 3am in the departure lounge at Hong Kong airport”.

Globalization has made capital vastly more mobile than labour, fracturing many rural communities. At the same time, the staggering increase in the quality of digital leisure time has coincided with a more globally competitive jobs market. As a result, communities are increasingly becoming virtualized, especially for young people and at the high-school education level.

This unmet yearning for tribalism is finding expression in some dangerous ways. At the same time, societies that are built around smaller units, like Switzerland and Utah, are seeing wildly beneficial results. In fact, Utah is such a staggering outlier on almost every metric in America that I think it holds the solutions to many of our current social problems. I hope to cover it in more detail in future.

Finally, there’s a practical business lateral here: Safi Bahcall’s book ‘Loonshots’ also found that once an organizational unit tipped over 150 people, incentives began to favour politics over innovation.

Insights from ‘Social Chemistry’

[13min Read]

King links to a survey at It takes about 10min but reveals the structure of your personal network as well as illustrating which people are your strongest social connections. It will classify you as a Convener, Broker or Expansionist. I’m generally pretty skeptical of this kind of broad categorization of inherently complex people, but still found it surprisingly revealing.

Source: Social Chemistry.

Essentially the difference is being a part of a robust group, being a linkage between different groups or being at the epicenter of a group. It’s hardly perfect: a third of people have no clearly defined style and 20-25% are mixed.

  • The importance of “Conveners”: A key insight of King’s book is the strongest kind of network is where your friends are also friends with each other. As Dunbar also found, roughly 60% of human conversation is gossip. Gossip performs a valuable social function of building trust and punishing bad behaviour. Because more people know each other, convener networks are likely to have more gossip. Convener networks also have high-bandwidth and redundancy. People who are Conveners are also more likely to be able to intuit the perspectives of other people.
  • “Brokers” produce creativity and information diversity. This is where the survey placed me, although I obviously wanted to be a Convener. It’s not totally surprising, given I’ve been a professional information broker for the last 16 years. The most robust connection between personality and network type is ‘high self-monitoring’ and being a broker. Essentially people who are keenly aware of how they are perceived and can adapt themselves quickly to new surroundings. Whereas Convener networks are often so similar they are prone to echo chambers, Brokers’ networks span more different groups, so are diverse and creative. Brokers produce better ideas and have higher job performance (I hope my boss is reading this). However, broker positions can often be short-lived and require considerable effort to maintain. Because of superior access to information, a broker is dependent on their reputation. This network structure can lead to psychological cognitive dissonance; if your identity is too chameleonic, it’s personally destabilising.
  • “Expansionists” are central nodes. They know dramatically more people than the other groups, and attain popularity, status and power. The more connections you have, the easier it is to make new connections. Physical beauty is a factor, but expansionists are also great at reading others and identifying who else is popular. Public displays of confidence (not necessarily accuracy) are important- and brought to mind the kind of ‘memelords’ that now dominate social media. Successful expansionists tend to be generous and elicit reciprocity. However, huge networks often make it hard to generate strong linkages; so many have systematised ways of managing their networks. High volume means low quality. Having too many weak-tie friends can actually be correlated with depression.
  • Give: King talks about the totally justified distaste for calculated “networking”. But makes the point that if you go into every interaction focused on what you can give, it makes the process easier. She postulates that’s why powerful people find it easier to network: they know they are giving value to those with less influence. In order not to be creepy, successful expansionists tend to be inherently likeable not actively status-seeking.
  • Get a professional sponsor. More than a mentor, a sponsor will use their reputation to support yours. ‘Sponsorship is one of the strongest predictors of promotions and salaries—roughly equivalent to the number of hours someone works—according to a study that examined the careers of tens of thousands of employees and dozens of predictors of job success. For promotion prospects, sponsorship matters more than someone’s gender, personality, education, and experience. The same was true for career satisfaction. Career success is as much a game of getting a sponsor as it is one of performing well.’
  • Be vulnerable: King talks about vulnerability as a facilitator of both friendship and trust. Escalating self-disclosure leads to personal closeness, and is obviously best when reciprocal. This is consistent with Brené Brown’s fascinating work on shame and I believe is a foundational concept in life.
  • Vary your routines. Somewhat surprisingly, social networks are still largely determined by frequency of in-person contact and geography. Due to routines, most people’s locations at any one time are predictable to at least an 80% margin every day. If you want to meet new people, vary the loops.
  • Reactivate dormant ties. Social connections that were stale for more than 3 years provided more useful and novel knowledge than either current strong or weak ties.
  • Be present, put away the phone. Yes, this is obvious but we still don’t do it. Head up, eye contact, smiles. I like the idea of everyone putting their phone in a big bowl at the start of a meal and the first person to reach for theirs pays the check. When discussing meaningful things, the mere presence of a phone materially diminished the perceived quality of the conversation. Dunbar found the time spent in meaningful conversations is correlated with life satisfaction. You can [allegedly] test your social intelligence here. I am only linking to it because I got 33/36.
  • Listen, then ask follow-up questions. More questions are generally good, but follow-up questions are positively transformative. It shows you’re curious and listening. 96% of people rate themselves as good listeners. You can likely get better at it… I spent a year training as a coach, this was the most important part by a factor of 10. A fun tip from the book is to ask someone “what is it like to be you today” and not interrupt for 2 minutes.
  • Touch. Obviously a total social and professional minefield even before COVID, but a short touch on the forearm can be transformational. Touch is 10x stronger than verbal or emotional contact. Perhaps a legacy of simian grooming.
  • Workplaces need psychological safety. This doesn’t quite mean what it sounds like- it means that employees can admit to errors without fear or blame. It’s contagious within groups and helps the network become a robust ‘Convener’ structure. Rudeness destroys this dynamic.
  • Uncommon commonalities build strong bonds. Greater disclosure increases the likelihood that you’ll discover someone else likes something niche you like, which itself produces tighter relationships. As Dunbar notes: ‘not only are you more likely to share genes with a friend, you are also more likely to think like your friends. That’s not because being friends makes you think alike, but that you gravitate together because you think alike.’
  • The health impact of loneliness is staggering: ‘Loneliness is turning out to be the modern killer disease, rapidly replacing all the more usual candidates as the commonest cause of death…. Perhaps the most surprising finding to emerge from the medical literature over the past two decades has been the evidence that the more friends we have, the less likely we are to fall prey to diseases, and the longer we will live.’
  • More friends, bigger brain. Brain size correlates to social complexity, both within mammals and individual humans. ‘Social Brain Hypothesis is a two-step explanation: ecological problems are solved by living in a group, and living in a group is solved by having a large enough brain to manage the stresses involved.’ Unlike any other animal on earth, by a mile, it takes us 20 years to become socially mature. The ability to delay gratification or inhibit inappropriate responses is linked to better social skills, in particular theory of mind, in children.
  • Most people can simulate a maximum of five other mental states at a time. As an example, here’s a fifth order joke: ‘A young boy enters a barber shop and the barber whispers to his customer, ‘This is the dumbest kid in the world. Watch while I prove it to you.’ The barber puts a dollar bill in one hand and two quarters in the other, then calls the boy over and asks, ‘Which do you want, son?’ The boy takes the quarters and leaves. ‘What did I tell you?’ said the barber. ‘That kid never learns!’ Later, when the customer leaves, he sees the same young boy coming out of the ice cream store. ‘Hey, son! May I ask you a question? Why did you take the quarters instead of the dollar bill?’ The boy licked his cone and replied, ‘Because the day I take the dollar, the game is over!’
  • Intuitively, friendship strength is correlated with time. We have an average of 3.5 social hours per day, so that’s 1min 45 secs per friend. But it’s not linear obviously. ‘Something like 40 percent of this social time is devoted to the five people in our innermost social circle, the support clique, with another 20 per cent devoted to the ten additional people that make up the fifteen members of the support group.
  • Cousins are the limit of a family group. ‘None of the six main kinship-naming systems in the world have terms for anyone who is less closely related than cousins. It is as though this is the natural limit for human communities, and everyone beyond that magic circle of cousins is a stranger of no particular importance.’

Some practical conclusions:

THE KICKER. This absolutely blew my socks off, because every excellent book I’ve read eventually relates back to The Tao, or the principle of balanced order and chaos. You will see it repeated ad-infinitum in my work because it seems to relate back to absolutely everything. Human networks are what Watts and Strogatz famously called ‘small world’ networks- which we understand through the strange (and proven) property of roughly being able to link any 6 people to each other. You can link everyone to everyone else because we have small, orderly groups of friends, but also friends that know a vast number of people a long distance away. King concludes that small world networks are comprised of a mix of her three categories.

‘Small world networks, of which human networks are just one example, lie in the middle of the continuum. They have enough order to make them searchable and enough randomness to provide the shortcuts necessary to create a small world. They strike a perfect balance of chaos and order.’

Sign up for updates