In a different vein, there's a lot of academic stuff written on Scandinavian workplace management, which is not precisely flat, but has flattish hierarchies based more on rotating and interlocking committees, rather than boss-and-subordinates tree structures. The goal there seems to be to make the sub-groups that are making decisions overt, so informal cliques don't make the decisions. So there are a lot of committees, with rotating membership to make sure they don't develop into hierarchies. E.g. 5 people are chosen to be responsible for something like working conditions, or strategic direction, or integration of foreign workers, for a term of a year or two, and then a different 5 people will hold that power next year. A big controversy there is how much Scandinavian workplace management is tied to Scandinavian culture, i.e. would it work anywhere else?
This is a more problem than networks vs hierarchies since both face this problem. All companies that succeed by definition start with a successful social structure, but as a companies increases the number of employees the social structure has trouble scaling. Clausewitz discusses a simple example of this with an army, as the army grows the number of links in the chain from a commander to a unit increases. He proposes a number of solutions (increase the unit size, increase the branching degree, split the army, etc) but in typical Clausewitzian fashion admits there is an inherent trade off between size and effectiveness.
Valve maybe choosing to stay small to avoid scaling beyond their effective structure. This is a counter-intuitive strategy in the world of bigger is better, but I'm interested in seeing how it plays out.
Valve shouldn't be a surprise. Despite all the ink spilled over management literature, even normal companies are typically full of dysfunction. There is no royal road to getting rid of defacto bosses.
I've been in companies which claim to be "flat". But that's just code for "management is less aggressive than usual". Our institutions make it hard to be flat. Analysis of serious attempts: (http://www.zcommunications.org/the-newstandard-a-parecon-wor...) and (http://www.zcommunications.org/work-after-capitalism-by-paul...)
I wouldn't call 300 people a small number for the size of a company.
In fact, Jeri Ellsworth addresses this point (I'm quoting the transcript by develop-online [0]):
"Their structure probably works really well with about 20 people, but breaks down terribly when you get to a company of 300 people. Communication was a problem. That's where management… Well if I had anything to do differently, would be to make sure a layer of management could do communication correctly."
So it looks like Valve has already scaled beyond their effective structure.
[0] https://www.develop-online.net/news/44746/Valves-perfect-hir...
Such scaling problems are not binary. They are probably less effective than they were at 20 people but they would be even less effective at 30,000. One way we might look at this is to assume (incorrectly) that a flat hierarchy is a complete graph which is the way most small informal groups are run.
A complete graph of 30 people has 435 edges. A complete graph of 300 people has 44,850 edges. A complete graph of 3000 people has 4,498,500 edges.
"Witches are not by nature gregarious, at least with other witches, and they certainly don't have leaders. Granny Weatherwax was the most highly-regarded of the leaders they didn't have."
Freeman was a member of the 70's feminist movement, and wrote this essay to critique how 'flatness' played out - and suggest alternatives.
The big, tl;dr idea could be boiled down to this: Flat organizations substitute unelected, informal leaders for visible, accountable ones.
Essay: http://www.jofreeman.com/joreen/tyranny.htm Wikipedia: http://en.wikipedia.org/wiki/The_Tyranny_of_Structurelessnes...
The origins of the hierarchy/nonhierarchy dispute on the left are a bit clearer, if that's what you're looking for. The polemics between Marx and Bakunin would be one place to start to see that side of it. To over-simplify, Bakunin thought that Marx's views had a dangerous authoritarian tinge that risked just replacing one kind of authoritarian government with a different one; while Marx thought Bakunin was an idealistic hippie out of touch with real-world situations, whose views would at best go nowhere, and at worst result in the workers' movement getting crushed as they spent all day holding disorganized meetings. For a later historical episode, something about the May '68 protests in France would probably have something informative. They were led by young leftists with non-hierarchical views, and opposed by the more orthodox French Communist Party. But I don't know as much as I'd like about that period, either, apart from the Situationists producing a lot of clever slogans.
1. information is often collected at the leaf nodes, so it must flow up the chain-of-command to be used effectively
2. power creates perverse incentives to distort information as it travels up the hierarchy, even when it doesn't information is distorted as it is compressed for fewer and fewer people (see executive summary)
3. hierarchies have to wait for signals to go up the tree and back down to act (which is why the US military has non-hierarchical decision making in places), this makes "tall" hierarchies slow to react
4. corruption of single node in a hierarchy can disrupt all nodes below or above it in non-obvious ways
5. hierarchies tend to have extremely critical nodes and cascading failure modes
Use of such structures should always be considered within the framework of a trade-off.
I would argue that well designed organisations or networks should have enough connections to route around/detect minor damage/corruption and quickly isolate the problem. For instance small world networks have this property http://en.wikipedia.org/wiki/Small-world_network .
Not all nodes in flat human networks are likely to be equal in information flow, but a plurality of high degree nodes does somewhat mitigate cascading failure due to disconnection but may increase cascading failure due to interconnection such as viruses.
The company was setup like a collection of startups. You would pitch your ideas to executive management who would green light hiring and what not. You'd also borrow or entice engineers over to your group. When I joined, it sounded very interesting, but it was a complete nightmare.
The executive management would green light things, but they'd green light 10x more than could be done by the engineering staff. The hiring queue was so long that you'd be lucky to see engineers hired for your project for 6 months. That meant stealing engineers from other groups which encouraged serious protectionism by doing things like say, loading down your group with busy work to keep your empire intact.
It was a political nightmare because no one was really in charge, so everyone claimed ownership. You had review processes for every single step of development by whatever group claimed ownership over something you touched regardless of how trivial.
The executive management rarely stuck their necks out making decisions. You had to prove something was better with data. The fact that data was easily manipulated didn't bother anyone, as long as they didn't have to take any blame.
The thing is, the company itself is wildly successful and could be held up as a model for flat organizations if you didn't look too closely. But really, it is in spite of their organization rather than because of it.
This bit of misinformation has now apparently calcified into conventional wisdom; however, 'Gatekeeper' explicitly permits code-signing without publishing through the App Store, with no attendant restrictions, and signed software is accepted by the default security settings.
All other fears boil down to merely the possibility that Apple will change their mind and disallow 'sideloaded' software in the future.
For now
What market is this company in, if I may ask?
Notably, the one which was proud to loudly proclaim it's flatness and openness was far and away the worst. It didn't quite fit the high school analogy, but it wasn't far off.
In reality, the "flat", "open" structure was actually a concentric hierarchy radiating from The Boss. The Boss > The Inner Circle > Untouchables > Loyals > The Rest.
As bad as that was, I'm still inclined to something generally flat given an exceptional leader to set the tone.
All this really tells me is, given a lack of formal hierarchy, people tend to organize themselves into the exact same hierarchies that are formalized everywhere else.
Also, I'd expect heroes to be small subset of exceptional people.
I think that open allocation in terms of project mobility is the right way to go. I also think that flat structures are rarely truly that, and that they're certainly not resistant to fast growth (more than 20% per year).
There's a paradox inherent in the fact that if there's no management at all, informal psuedomanagement appears and that's worse (because it's unstable and must compete to hold ground). What you want is minimalist, benevolent management that operates in the joint interest of the company and its people, but that seems to require a relatively rare lack of ego. You need a cultural shift to a maker's culture where people who build things are held in highest esteem, and the long-term builders/individual contributors have as much influence and credibility as management.
It's pretty rare that one can do away with management outright, though. Someone has to protect the good.
Management is basically there to provide accountability. But peer systems can also provide some accountability.
Semco has an interesting model because they do insane things like allow employees to set their own salary. But your salary will be public, and if you can't prove to your peers on a regular basis that you are still needed, you will be asked to pack up your things.
This is just one of the way that they've delegated management functions downwards, which by all accounts works very well. There is still an official hierarchy but it's more about mediating disputes or very high level strategy.
http://www.freibergs.com/resources/articles/leadership/semco...
Do you mean the opposite? Either way - "certainly" is a cheap trick. Reasons please.
The issue with fast growth is that it requires people to continuously adapt to the changing power structure and the influx of new people. Beyond a certain point, people are no longer focused on their daily work, but on making sure the new hires don't get "slid in" above them.
I've never quite believed that things could go so smoothly as they say. I'm not saying this article is necessarily true, but that I wish there were real research into the negatives of these kinds of management structures (or lack-of-management structures).