3 ms·
I have felt this too and it turns out I leave my startups around 200 people -- when I can no longer remember everybody's name instantly. (but where's the line?
by gumby 8y ago
I have felt this too and it turns out I leave my startups around 200 people -- when I can no longer remember everybody's name instantly.
(but where's the line? I worked with a guy who wouldn't work for a company of more than a dozen people. I've looked at his linkedin over the years and indeed he worked for some phenomenal startups...but always left when the started to grow).
- cimmanom 8y agoThe Dunbar Number (the number of relationships any individual can maintain) is typically considered to be around 150-200. So there's a pretty good reason for that 200 number. The number of individual channels of communication in a company hits 100 just past a dozen, which often makes it an inflection point for company culture and might affect why your acquaintance tends to leave at that point.
- n42 8y agoI have a working theory that these inflection points follow the Fibonacci sequence. At least based on my experience, communication and culture changed dramatically at each stage, most notably around 34 and 55
- groby_b 8y agoNot sure if it's the fibonacci sequence per se, but yes, there are inflection points around those numbers. Up to ~15 people with experience and a common purpose, you can just all work together, little structure and roles, and stuff works OK. Teams form ad-hoc, based on problems. From there, you need to define more well-defined roles, so people don't step all over each other. Communication starts being important, because everybody talking to everybody has broken down at that point. Hey, look, a wild manager appears. You're a startup, you try to save on managers, so teams can easily balloon to ~20-30 under a single manager, but then it breaks down again. You realize you need a second manager. As you approach 50, you want to add a third manager. At that point, you have, want it or not, a "leadership team". There needs to be very regular and deliberate communication. Policies start being set, because teams start drifiting apart. You'll hit the next inflection point at 100-200 people. Because at that point, your leadership team is starting to hit communication limits at the leader-to-leader level. (Your leadership reaches the same initial inflection point your original team reached). As that cascades up, you add layers. The problems keep repeating. The next hunk is at ~500 people. The "old hands" feel like the amount of people they don't know really outnumbers the people they know. Those are the sizes I've personally seen happen, several times the same pattern. And based on that progression, I'd expect the next transition point at around ~1,500 people. So it's not quite fibonacci - it seems to skip steps - but it's a good enough guide for the first few 100 people.
- Bartweiss 8y ago> I'd expect the next transition point at around ~1,500 people. I wonder if this is present, at least for the same human-relationships reasons as the others. It seems like there ought to be an upper bound past which people are basically disconnected and lack personal relationships to most of the organization, at which point those relationships become less relevant to structural issues. I guess the ~200 and ~500 problems will keep recurring within subgroups like divisions or offices, but excluding that it seems like non-social constraints might dominate. At 1,000+ person companies, it seems like the problems I hear about are often structural (e.g. the contractors are treated as second-class citizens) or logistical (e.g. we can't get the bulk of the company to an all-hands, even once a year). At 10,000+, purely-statistical treatment of employees seems to take over, like IBM's systematic dismissal of older workers. (And resulting individual-level idiocy, like bringing back some of those essential workers as contractors at higher rates.)
- groby_b 8y agoI'd wager a medium-sized amount on the 1,500 inflection point. It's based on a vague feeling of being in a 1,200 person office, and feeling like things were likely to change soonishly. Also, for large companies, you have separate social dynamics for offices. I've seen the 200/500 point happen in offices of much larger companies. You then layer the "too large to comprehend" dynamics on top of that, for added fun :) (Also, you can get 1,000 people offsite. It's just a lot of work. That might be another of the 1,500 size problems) Either way, social dynamics at large institutions is a fascinating problem. I wish I could find decent literature around the topic, because it's a fun subject :) (Corollary: All the things I said above are anecdata, I don't know if there's a theoretical foundation)
- dredmorbius 8y agoUnder military command hierarchies, spans of control are typically 2-6 of the next-smallest unit. Whether military administration is much lower than business is another question, but it's an argument for smaller SoC. https://en.wikipedia.org/wiki/Military_organization#Commands,_formations_and_units https://en.wikipedia.org/wiki/Military_organization#Commands...
- Bartweiss 8y agoInterestingly, 30-50 people is the standard range given for 'band societies', the most basic pre-agricultural group. That's been cited as a problem with Dunbar's number, but most people seem to think it's just another major "natural group size" among many. Dunbar has also cited 5 and 15 people as rough 'rings' to which we devote disproportionate amounts of our socialization time. I have a bit of trouble judging how Fibonacci compares, since the distinctions I've seen are gradual rather than nonlinear. But it does look like moving from ~15 to ~35 might match a switch from close-association to band structure, and exceeding ~55 would match another restructuring away from that.
- cimmanom 8y agoYou might have 30 people who live together; but even just to avoid inbreeding, their social milieu would likely include a handful of other bands with nearby territory. Thus any given persons social ties could very reasonably number in the low 3 figures.
- Bartweiss 8y agoYep, exactly. Lieberman ought to more than I do, but his argument that "150 connections" can't be selected for because pre-agricultural bands were smaller seems ridiculous. Most fundamentally, because 150 doesn't have to be directly selected for; it's a consequence of working memory that might be valued for unrelated reasons. But more directly, because "bands of 30-50" is not "only knowing 30-50 people". Bands were usually extended family groups, so the value of maintaining personal relationships outside of those groups is obvious. Honestly, the best I can work out is that Lieberman is a few decades out of date on his understanding of how pre-agricultural groups interacted.
- bitwize 8y ago200 is on the order of Dunbar's number, the theoretical maximum number of individuals a human can keep track of. The most commonly cited value for Dunbar's number is about 150, but there's a range between 100 and 250. Dunbar's number is thus a theoretical upper bound for the size of a startup a single person can comfortably relate to every employee of, but some people just aren't comfortable even with a company that size; presumably they want to leave plenty of headroom in their social network for family, friends, church, bowling league, etc.
- Bartweiss 8y agoI've always been a bit sad that there hasn't been more empirical work on Dunbar's Number. It looks like attempts to validate the modern number range from around 150 to around 250 (Bernard–Killworth, who seem to be leaders here), but of course that variance has a lot to do with exactly what counts as a 'connection'. You're very right about the mistake of looking only at company size; Dunbar suggested that maintaining the full connection set would take a very large amount of socialization time, and most people have at least some connections outside the office. Even anecdotally, there are lots of people who describe feeling disconnected at work because they spend more time than coworkers on non-work connections. This seems to have been ignored by a lot of people using the number. Wikipedia claims the Swedish tax service limits offices to 150 people based on Dunbar, but that could easily be 50 people too large for a workplace community. Malcolm Gladwell, for whatever his data is worth, claims that Gore-Tex hit cultural problems as offices grew beyond 150. If those anecdotes are accurate, it implies offices function well with high rather than total connectivity. Intuitively, that seems quite reasonable. Cross-organization connections are valuable, but not every pair of jobs has equal value-from-connectivity. And small teams are infamously hard to build because everyone involved has to get along and agree on how to function. If you aim for 90% connectivity (or 50%, who knows) instead, it's possible to maintain a high-empathy organization while still allowing for personality conflicts.
- gowld 8y agoWhy does it matter? If your startup has 12 or 200 people, it has lots of clients and vendors who you aren't friends with, just like people in a different org in a large company.
- gumby 8y agoThat's a reasonable question. The simple answer is that I just become uncomfortable when I don't know everybody's name (yet they all know mine). The other part is that it's a company (a group of people) not just a corporation; while I am friendly with some clients and vendors (ones I worked on, either old ones or major ones) when we happen to hit it off, it's not the same as the company of people.