When you join a small firm, you are joining the firm.
When you join a large firm, you are joining the team.
I asked a US Marine who he was loyal to, and he said his unit. Not his country or the military.
The same happens at schools and universities: you might know your classmates on your course, but you don't know all the people at a university.
Same goes if you're a sports fan. You might be a supporter, but your experience depends a lot on who you're sat with. Ultras, those kinds of hardcore groups are rarely huge mobs.
The key point here is loyalty, or solidarity perhaps, if we're go with the military analogy.
"Google cares deeply about making xyz more ABC." We know this is, at best, flowery language. "Google" can't care about xyz any more than an inanimate object can.
Maybe that's an exaggeration, but I do think personifying an organisation to a think that "wants," "loves," or "believes" is disingenuous. That's not the case, necessarily, for a smaller group.
> When you join a large firm, you are joining the team.
Yes I think to a great degree because if you join a <100 person company you will interact with the founders regularly, face-to-face, and get to know them. If <20 people, you also get to know them outside work.
Beyond about 100-200 people the CEO is just a face on a webpage that takes credit for what you do (and you're happy with that as long as you're getting paid fairly for it).
My first job out of college was at a ~100 person firm that quickly grew to 200, when I left. The culture I was hoping to join was notably different when I got there (at 120 people) and especially when I left (at ~200). That culture change, and the different expectations, really surprised me at the time.
You need reinforced values/mission to care about the bigger things. You may not care about your company as a whole. But if your company's strength helps it, say, reduce climate change, and you care about climate change, now you can care about your company.
Anecdotal but the first two companies I worked at both grew from around 70 when I joined to around 200 when I left them, and in both cases I attested the very thing this article is observing. Influential and capable people became disengaged and frustrated because they could not move things forward by force of will anymore, and morale dropped overall among everyone.
Anecdotally I have seen that from around 50 on people don’t communicate with the CEO anymore but through some middle managers. If these middle managers aren’t really good, the company starts splitting into different camps that don’t communicate anymore.
This has happened at my company that is now 35 people. Too many managers that have fractured communication. Getting things done was easy when there was 10 people.
This can happen even with small numbers of managers. It only takes one territorial or poorly-communicating asshole to fracture a broader organizer. If anything, a small number of managers exaggerates the damage that one bad manager can do.
A startup I’ve watched closely grew to 160 and it’s still unclear they will actually survive this. Early employees are used to the old, startup-y way of doing things, but that’s no longer feasible.
I think there are actually many such points at which teams/companies change significantly. Going from 2 to 3, or 4 to 5 people can feel like a paradigm shift. Team communication often becomes too complex once the team grows beyond 8. I have seen clique formation starting at 5 people (in a band, but that probably doesn’t matter much), but less so in groups of 4.
I've noticed that e.g. over dinner only groups of 5 or smaller can keep one coherent discussion going; if you go to 6 or more people they will inevitably split up into subgroups. That's why I think meetings where people are supposed to discuss an issue and come to a shared understanding of something can only be at most 5 people.
That is why I can't work with others. I normally will eat one pizza all by myself, to the shock of everyone else around me. There is a downside to a fast metabolism (I know a lot of "fat" people who would still gladly trade downsides)
> eat one pizza all by myself, to the shock of everyone else around me
I'm puzzled. Which kind of cultural context does this unfold in? Are these pizzas extraordinarily large? Or is it just a cultural issue that you are supposed to share in any event?
Where I'm at (Northern Europe) ordinary pizzas are about the same size of a LP record (around 30cm, or one foot for the metrically challenged). Unless you specifically order multiple (and/or "family sized") pizzas you are really expected to eat the whole of one pizza by yourself. It's in no way remarkable.
When US people say "one pizza" (in group context), they usually refer to a pie around 40-45cm or so. A size where most people would eat roughly 2 slices.
> I think there are actually many such points at which teams/companies change significantly. Going from 2 to 3, or 4 to 5 people can feel like a paradigm shift. Team communication often becomes too complex once the team grows beyond 8.
We are struggling to get beyond 10 right now. Every time we hire a new person it feels like resetting the entire org chart. The communication thing is 100% on target. At 8 people it felt manageable. Once we got to 10 it was like someone flipped a goddamn switch somewhere. We added a small management layer but this comes at a cost (i.e. we probably need 15+ to justify it fully).
I know a place that went from 200 to almost 900 over just a bit more than 18 months (VC money can be nice sometimes I guess). The new hires included some highly experienced, and thus less start-upy and older, people. Those were highered, across the org because you found them basically everywhere in small numbers, to help get the company ready for production and finalize development (my guess, it can also be just blind luck).
Then the whole thing got delayed, right around the time the new folks ended up, as part of a bunch of projects, exposing the sheer incompetence in some areas. Areas that are vital to get production of the ground. The delay led to a fight between the old and the new culture, and the delay allowed the old to close ranks with founders and clue less career management. As a result, the new culture lost, then the experienced new people left and / or got sidelined. As a result everything got delayed even more, and now VC money is close to run out.
There are also big changes at smaller numbers: ~8, ~30.
Before you hit around 10 people, you all fit at a comfortable table and can have reasonable meeting where there's a chance for everyone to talk. It's also a reasonable number for one person to manage that size of a team and be able to check in on the others on a regular basis. It's hard not to know the rest of the team - assuming you work in the same place (big assumption these days). It's more likely at this scale that many people have to have broad skillsets in order to work on whatever is necessary.
25-30 is a size where you start to loose track of the names of new people and don't really know some of the team. It's not possible for one person to do a good job managing all the people, so you've probably already started to break things into subgroups and have more mid-level leaders/managers. There are more likely to be specialists at this scale, but it's hard to manage their workload - either too busy or not enough.
As you grow above that size the subdivision and communication issues increase. If you are in an office, you're probably not all in the same room/floor/building. It's likely that there are some people that you almost never see. When someone new is added, you may not meet them for a while. The specialization issues are different. There's more push and pull around what specialist areas to add or drop, but some of the sub teams are more self-sustaining and managing. However when things go poorly, expect re-orgs and layoffs. Loyalties are more to the local team level, not the overall organization. There's more likely some us vs them within the company.
There is some criticism of Dunbar which I can't evaluate. Even if the specifics are inaccurate within it, there seems to be some truth to how sets of humans scale. I like the Alex Komoroske presentation below about "coordination headwind" as an explanation that doesn't need evolutionary psychology to work. It is easy to understand and deep enough to be well worth paging through periodically.
The ideas started when I was studying the emergent power dynamics of Wikipedia's user community in undergrad, and were also informed by participating in web standards. I believe they are not specific to any one context, but rather emerge inherently any time there are individuals with autonomy who care deeply. Instead of ignoring these dynamics, I think it's important to acknowledge and embrace them with a sense of compassion and openness.
Wouldn't be surprised if this is true also for education. It's bizarre to me that we send children off to these huge warehouse size schools with over 1000 other kids, looked over by a small number of unrelated adults, and expect odd social problems to not arise.
I've seen all kinds of different things. I know of one school near me where it is 1000 per grade. I know of a different district where it is 1000 for the whole high school (4 grades), and another where is is 1000 for the entire district (12 grades).
I think on balance 150-300 per grade is ideal.
Schools that have less than 150 students in a grade just cannot offer students anything other than the same basic actives everyone has. You won't take calculus because there are not enough students to start a calculus class, and this holds back the smart kids futures. You will play football as we cannot field a team without every student there, it doesn't matter how much you would prefer to play any other sport instead we can (not they won't, they cannot) never start a team.
Schools with more than 300 start to get to the point where they can't let students into popular courses/sports and this breeds have/have-not culture wars.
My numbers above are close, but subject to some debate.
Wow. I went to a school with perhaps 500 pupils total covering all 10 grades, and in retrospect I think it was too big. My own children attend a much smaller school.
I should add that there were some specialized classes available, like programming, in the evening where people from multiple schools were gathered.
My high school (this was the 1980s) had 4,000 students. In retrospect I'm surprised there weren't more social problems. A lot of the students did come from a similar socio-economic background (upper-middle class to very very wealthy families), perhaps that explains it.
Likely because humans have evolved to be able to know, remember and relate with ~150 people. Anything larger than that, requires organization and representation.
Gore's approach of splitting factories resonates with me.
Seems like a good point to split a business unit based on function or something. Do some Conway's Law hacking. If that's not possible, then simply horizontally partition.
This reminds me of a part of In Dubious Battle by Steinbeck where Doc Burton talks about large groups, the "group-man".
> ...they seem to me to be a new individual, not at all like single men. A man in a group isn't himself at all; he's a cell in an organism that isn't like him any more than the cells in your body are like you. People have said, 'mobs are crazy, you can't tell what they'll do.' Why don't people look at mobs not as men, but as mobs? A mob nearly always seems to act reasonably, for a mob.
... when group-man wants to move, he makes a standard. 'God wills that we recapture the Holy-Land'; or he says, 'We fight to make the world safe for democracy' ... But the group doesn't care about the Holy Land, or Democracy ... Maybe the group simply wants to move, to fight, and uses these words simply to reassure the brains of individual men. ... You might be an effect as well as a cause, Mac.
Definitely noticed this at Linden Lab. Fewer than 150 people, it felt like a big boisterous sports team with everyone focused on the same goal. Over 150, started noticing infighting, office politics, etc. become more and more of a problem.
The irony is, Philip Rosedale often talked about the Dunbar number in relationship to how virtual communities were best formed/organized!
I usually call this a rogue wave in startups, because it isn't really any one factor that does it but in conjunction many factors like the dunbar number, customer growth, overhead, renewals (especially in B2B SaaS), burnout from multiple years of growth all tend to hit right around the same time.
It's a very real phenomenon in my experience and I've also used the moment to really gain separation from competitors as well, since the stumbling is predictable at right around this point.
Logarithmic scaling factors for your management layers.
Below 50, you basically have a flat org. Even with nominal mangers, everyone is effectively on the same team. 50-150 you need an actual management layer, but these managers are all on the same team, and are also close to the ICs on their team so they have the full context of the entire business between the management group. 150+ you need to introduce a manager-of-managers layer which means you have increased overhead across the board in addition the other communication and coordination challenges.
> Keeping things below 150 means you can manage the system by peer pressure, whereas above 150 you need some kind of top down, discipline-based management system.
> That’s because founder Bill Gore felt that when a unit of workers got big enough, “we decided” became “they decided.”
I'd buy it. It seems like many of the problems we are having in society right now, and the challenges we are facing are results of biggness more than anything.
When you join a large firm, you are joining the team.
I asked a US Marine who he was loyal to, and he said his unit. Not his country or the military.
The same happens at schools and universities: you might know your classmates on your course, but you don't know all the people at a university.
Same goes if you're a sports fan. You might be a supporter, but your experience depends a lot on who you're sat with. Ultras, those kinds of hardcore groups are rarely huge mobs.