The problem with explicit leadership positions is that people have no downward mobility. If you are bad, then the organization loses a great resource. I like the model of dept head in academia. Professors really want to be doing research, but they also need a dept head who handles administration. No one really wants to do it, so it rotates through the department. There are some perks like a pay bump, reduced teaching, extra grad assistants (or whatever).<p>Many more people get a chance to try out a leadership role to grow, but it is expected to be temporary so people can go back to being technical with no repercussions. Individual contributors get better perspective on management. People that are good at it and have an affinity for it stay in the role longer.<p>Self organizing teams still have a leader, but the team selects the leader.
There's one thing I've learned from white water rafting: On slow parts democracy works fine - everyone discussing about the best path - but when things start happening fast you need a single captain to make the calls, otherwise you end up hitting every single rock in the river while trying to agree if it's better to go left or right around it.
I would like to refer to the following - The Tyranny of Structurelessness, originally meant to address the concerns regarding 'flat' organizations.<p><i>"This means that to strive for a structureless group is as useful, and as deceptive, as to aim at an "objective" news story, "value-free" social science, or a "free" economy. A "laissez faire" group is about as realistic as a "laissez faire" society; the idea becomes a smokescreen for the strong or the lucky to establish unquestioned hegemony over others...<p>...Thus structurelessness becomes a way of masking power, and within the women's movement is usually most strongly advocated by those who are the most powerful (whether they are conscious of their power or not). As long as the structure of the group is informal, the rules of how decisions are made are known only to a few and awareness of power is limited to those who know the rules. Those who do not know the rules and are not chosen for initiation must remain in confusion, or suffer from paranoid delusions that something is happening of which they are not quite aware."</i><p><a href="https://www.jofreeman.com/joreen/tyranny.htm" rel="nofollow">https://www.jofreeman.com/joreen/tyranny.htm</a>
"A chair who asks, “Is everyone OK with choice A?” is going to get objections. But a chair who asks, “Can anyone not live with choice A?” is more likely to only hear from folks who think that choice A is impossible to engineer given some constraints."<p>It's funny - I've adopted a similar approach towards getting my kids to agree on something with each other. Instead of asking, "Should we do A or B or C?", and having one want A and the other dead-set on B, I ask which one do they definitely NOT want to do. Works pretty well for kids under 10...
Assuming whatever you're trying to do is a good idea, my method is to just write a "heads-up"-type message and just go for it.<p>For example, "I'm going to do foo. If anyone has any concerns or anything, please let me know". This covers your ass, provides a discussion/feedback area, and lets you actually get things done without a lengthy bike-shedding process.<p>One important lesson I have learned in my career is that fortune favors the bold.
Interesting that most comments deal with why flat structure doesn't work. Newsflash: of course there is always a boss. Someone needs to sign the paycheck after all. But the term "flat structure" usually refers to lack of <i>middle</i> management. And yes, that <i>can</i> work, and much better than traditional hierarchical organization. But it takes a great leader and capable employees to achieve that. I am lucky to work at such place and it rocks.<p>As for reaching consensus, I think the idea is great. Instead of seeking consensus try to find a few proponents and make sure nobody sees fundamental flaws... I think this is a perfect tradeoff between doing the right thing and moving fast.
To bastardize Greenspun, any sufficiently large flat org contains an ad-hoc, informally-specified, bug-ridden, slow implementation of half of top-down org.<p><a href="https://en.wikipedia.org/wiki/Greenspun%27s_tenth_rule" rel="nofollow">https://en.wikipedia.org/wiki/Greenspun%27s_tenth_rule</a>
While I'm very much in favour of reminding everyone of Tyranny of Structurelessness and shadow hierarchies, the article is referring to a more IETF-like situation of voluntary collaboration. Here there are other sorts of shadow hierachy:<p>- Postel decentralisation: it used to be the case that a lot of important internet functionality was run through one person. Less so these days, but there will still be critical individuals.<p>- Fait accompli. This is really apparent in browsers, where the W3C inevitably lags actual practice and features tend to be introduced by deploying first and standardising later.
I currently work within a profit-share-based flat-org. We have a rule to never exceed 12 people in company size. Sure there might be several of us that are "core" members and others that come and go throughout the years. So in a sense this seniority has translated to hierarchy to some degree.<p>We're all mostly web developers so the mind-share is important since our technologies move fast. Decision making doesn't cost us a lot of time because there aren't many big decisions to make. Our only producer and founder see client faces the most, but not exclusively. We do good work and rely solely on previous clients and referrals banging our doors down for more work – so we fortunately don't have to worry on that front or "crown someone" with biz dev because it's on autopilot for the most part (definitely over-simplifying the work it takes to get a client in the door but you get the picture).<p>One would assume that our profits would cap out eventually because we aren't scaling the labor-force, but we just keep getting better paying projects and have a few internal projects that provide some passive revenue – so we've been able to nearly double our revenue year-over-year for the last few years and is now stably seven figures.<p>Our big decisions come down to whether or not to hire a specialized dev for an upcoming scope or maybe what conference we attend that year. These decisions don't happen often but they happen openly and almost everyone is satisfied with the result almost all of the time. "Almost" is used unapologetically because we all know there's never been a perfect organization and that keeps everyone involved perfectly happy. We can't always be skipping around in the cotton candy fields and enjoying the sugar rain like the unicorns we truly are on the inside – so when we have to go to work we recognize that our pattern is better than the vast majority of agency-models here in NYC and are still delighted by our cumulative work experience day to day.<p>Flat-org is sometimes tossed around as if it projects itself as a silver bullet to organizational overhead. But anyone that cares to think clearly for a moment can recognize that any org structure has a place and time and that there will never be a one-size-fits-all solution to managing people.
It seems there's an ideology here (in the Gary Bernhardt sense: <a href="https://www.destroyallsoftware.com/talks/ideology" rel="nofollow">https://www.destroyallsoftware.com/talks/ideology</a>), that flat organizations are <i>impossible</i>. Not just bad or inefficient, but literally cannot possibly happen, that it will always and forever immediately decay into a shadow oligarchy (which to be fair can happen).<p>Is it really that hard to take these people at their word, along with other companies without a stated strict and rigid hierarchy? Little wonder such organizations are hard to build, especially in a culture that holds the very notion in contempt.
The comments here have completely derailed. Remove "flat organization" from the post title, and the article is still valid. It's talking about consensus making which applies to any org structure.
Rough consensus is a really naive solution to making decisions. Being in a flat structure doesn't change the fact that different people have different requirements.<p>Just because you have more people "agreeing" to one particular idea doesn't make it the right idea.
There should be someone designated to gather rough consensus and act upon it. Someone who leads the group. Perhaps they could be called a "leader"?
I’ve only had about two years of software experience, but my (junior) opinion is that I’d prepare to spend even more of my brainpower on political cagematch fights than usual, as non-hierarchical means my “equals” are going to start trying to intercept credit, take lead on things, etc. and position themselves as de facto managers without me being able to tell them “fuck off, you’re not my manager.”
It just seems too fashionable to talk about flat organizations nowadays. And, as many have pointed out, it's a bit ironic since organizations that claim to be flat tend not to be in practice. It's obnoxious when people claim that things are one way when, in fact, they are another.<p>It seems people have developed some kind of allergy to leader figures. In the US, this probably has a lot to do with the political situation. The current president is non other than the antithesis of a good leader --- and he's "on top", if not literally, then figuratively. Elsewhere, it's probably just cyclical. We're at the bottom of a leadership trough in history.<p>However, I can't count the number of times I've wished for a good leader lately. Call me mentally lazy or a conformist, but I want someone to give me a sense of direction and purpose. I want someone I feel like I can relate to and trust. Maybe I'm just the type that gets a lot out of mentoring relationships.<p>I guess it's just one of those things where you need to be the change you want to see in the world, eh?
Our hackerspace runs in a "flat manner".<p>In reality, there are people that are, in a Animal Farm sense, "more equal" than others.<p>It also means there's very little responsibility for our shared space, little responsibility for operations, other than what individuals take on.<p>It also means that avenues that we could do (work more with First Robotics, or review if what we're doing is the best for our mission) is shouted down by said "more equal" members.<p>Sadfully, to that end, changes that I believe would be better can't even be discussed. I've quit doing so, and just pay my dues and have my 3d printer there (bed size is 500mm x 500mm x 400mm) along with 3dp equipment. Admittedly, its a sad way for me to "interact", but I'm not one of the 'in' people...<p>I'd much prefer if we actually had some hierarchy. We don't need to have every position delineated, but a basic framework makes a ton of things clear. Right now, I can't even figure out who knows how to make a static IP in the hackerspace's dhcp table. <i>Someone</i> knows.
Whenever I hear about flat organizations the old quote comes to mind: "All animals are equal, but some animals are more equal than others". It seems to me that all these orgs are a few people at the top that have a say and everybody else needs to survive through a lot of political skill.
I'm still looking for someone who works at Valve who can tell me the nuts and bolts logistics of how things actually work there. If there are no bosses, how are decisions made? How are raises actually determined? They publish their handbook but the logistical details aren't in there.
Quoting my self from a comment on another article <a href="https://news.ycombinator.com/item?id=19468090#19468699" rel="nofollow">https://news.ycombinator.com/item?id=19468090#19468699</a> :<p>> AFAICT—and again, I'm just someone who read about this stuff in magazines as a child—this kind of management-theory/theory-of-the-firm study and experimentation really was much more prominent in the public eye twenty-five or more years ago, and has faded from popular awareness since; maybe because the work itself is less common, and less influential on institutions, nowadays too. I think that's a pity, because it seems relevant to some supposedly-new things you see generating excitement nowadays, especially in the vicinity of the tech industry. For example the whole Valve structureless-or-notionally-structureless-organisation effort is probably pretty well prefigured in the work of a generation of management consultants who looked at and tried out all kinds of institutional structures or absences of structure beside the traditional corporate model, generally in pursuit of much the same ends—more innovation, more amenability to change, better decision-making—and with similar mixed-to-disappointing results.<p>I find it weird to hear this kind of thing discussed without reference to previous research.
Have been thinking about this a lot lately. Not just decision making, but how to actually function in a healthy way in the day-to-day, in a flat org.<p>The place I work recently restructured from a more traditional hierarchical structure to SAFe - so the business has a well defined hierarchy, while engineering teams are completely flat. Almost 4 PI's on (~9 months) the promise of "self organizing teams" has not been realized.<p>We have a lot of communication problems, especially technical cross team communication. The development process wants to move much faster than the information actually flows, and if the information does actually flow what is communicated is wrong/incomplete. So what ends up happening often is that teams choose to operate in their own bubble.<p>Another issue is having common direction across all teams, while also allowing teams to operate autonomously. That common direction is missing so you end up with work from one team being very different from work by another team, and in a system that integrates a lot of different components it is awkward to navigate and work in.<p>The above, among other things, brings out "2nd order effects", like strained inter-team relations, to put it lightly (so there is an additional policing aspect to contend with there as well).<p>We have a System Architecture group (as defined by SAFe) but their time seems to be consumed by meetings and managing information flow and coordination across teams. There is also a "shadow hierarchy" (as others have mentioned thus far on this thread) that has emerged, and that helps a little, but it's really a poor solution.<p>Curious on anyone else's thoughts/experience here.
The Decred open source project approves spending on proposals based on stake holder vote, 60% yes for approval. <a href="https://dcrdata.decred.org" rel="nofollow">https://dcrdata.decred.org</a> shows $17m in the treasury currently. Votes are the final say. Current and past votes are viewable on <a href="https://proposals.decred.org" rel="nofollow">https://proposals.decred.org</a> . Its all open source, docs <a href="https://docs.decred.org/governance/politeia/overview/" rel="nofollow">https://docs.decred.org/governance/politeia/overview/</a> code <a href="https://github.com/decred/politeia/" rel="nofollow">https://github.com/decred/politeia/</a> voting charts <a href="https://alpha.dcrdata.org/proposals" rel="nofollow">https://alpha.dcrdata.org/proposals</a>
My work's been effectively operating as a flat organization for the last several months as our manager's been away on leave. It works for the most part as we all have our own jobs we do and all know what we should be doing. Until it doesn't work. There's no accountability with the schedule and a big communication problem between everybody. When things change, nobody really knows whats going on, when mistakes are made everybody's quick to point finger and problems take longer to solve. You really notice the lack of a manager when they're gone. All these little things that used to be taken care of now fall onto whoever happens to be around to deal with it and it's definitely been causing problems we never used to have before.
I find that places that harp on their "flat" organization tend to be places where the people in charge are just trying to defer the hard work of putting structure in place or just afraid of it for whatever reason.<p>In my experience flat organizations just try to keep everyone happy by making them think they have decisive input on things that they honestly shouldn't, past a certain point.<p>I find (completely anecdotally of course) that they don't scale, and I think what the writer (who ironically is a lead engineer at this supposedly flat organisation) is running into are the limitations of this flatness once it gets to a certain point ... I bet the next post will be something along the lines of<p>"Managers: a necessary evil"
I'm sure you've all worked some place that talked about 'meritocracy'. I'm sure most of you have worked some place where every time a promotion happens, people joke about "I thought you were already a <new job title>".<p>Years ago someone put me onto this phenomenon. Some people in a business have no trouble making decisions and others are hesitant for a myriad of reasons, from self doubt to not wanting to be responsible for the consequences, to just not liking being randomized. These people will line up behind folks who seem confident, and as this manager half-joked, all he really had to do was 'show up' and start making decisions and people would follow him.<p>The best description I've ever heard of this phenomenon, and describes my own experience, is that people are looking for and will line up behind someone they can trust. If you sound confident they will give you the benefit of the doubt. But if things go wrong and you don't jump on the problem, they get burned and they won't listen to you as much in the future, even if you are ostensibly in charge. Strong opinions loosely held, own your mistakes, be good at troubleshooting, talk to other teams (even knowing things an hour before everybody else makes you look a hell of a lot smarter), and people will defer to you even without a title. Because even if the boat goes down they know you'll go down with it instead of running away. People will take your advice, rank be damned.<p>That said, there are darker patterns here too, that could easily dominate in a 'flat' organization. Hoarding knowledge forces people to defer to you. There are obvious flavors of this that I don't need to explain, but here's one I see all the time but others miss: modelling systems to exactly reflect your mental processes (instead of a simplified version, see Brian Kernighan on debugging) means nobody else can work on it except for trivial things. This makes people feel stupid, and imposter syndrome stops them from fighting back.<p>As one of my bosses said, you should fire all of the 'indispensable' people because they are holding up progress. An earlier boss started a new project and wouldn't let a bunch of us work on it because we were 'too important' to the old project. I think maybe 2 of us figured out this was a trap and worked to hand off our stuff.
> At Doist we believe that open and sincere communication improves our decision-making process. That’s why we’ve built a culture that encourages feedback at all levels of decision-making<p>Who is 'we' and doesn't that get to the nub of the problem? It's highly unlikely that they held a group conference and decided what 'we at Doist' believe.<p>That 'we' actually refers to a small subset of leaders who decide, and it is 'flatness' and stagnation for the rest.
The article touts transparency as being the virtue that led them to a flat organization. I'd argue that _it's within an explicit hierarchy where transparency (e.g. who makes decision, who earns more, who gets promoted) actually flourishes_ with the lowest cost. As other commenters have pointed out, a hierarchy will emerge regardless, however, transparency will be subdued beneath secret cabals and handshakes.
A good resource to learn more about scaling flat self-organizing organizations is the book <i>Reinventing Organizations</i>. It's a curious mix of dreamy hippie philosophy interspersed with very practical examples of flat orgs and how they remain flat while still getting things done.
In my experience, hierarchy (explicit/implicit) is necessary to conclude differences which are grey. In fact, there are cases where I have seen one side make their points logically enough that they are basically black & white and the other side sticking their guns saying "good points, but I think I would prefer to go ...". One might rightfully say the issue is deeper but we don't always have the luxury of choosing the most well fitting team. Somebody with the authority (granted/earned) has to step in and say "I have heard all sides and this is what we are going to do". Saves endless bickering over never ending e-mail chains.
Are there examples of flat organizations that have to meet make-or-break deadlines? I'm a little skeptical of a flat org's ability to hold itself accountable to externally-imposed deadlines/requirements/etc.
The answer is simple: Decide to leave the fantasy of flat organizations behind.<p>Really.<p>Flat organizations might work for a yet-to-be-discovered species. They do not work for humans. We can’t even have a picnic without leadership.
Answer: you get lucky by having a great flat organization. Great people and great cohesion.<p>Structure is a pricey insurance premium. You best case will never be as good, but neither will your worst be as bad.
My org uses a similar ranking method for hiring panels-- your feedback can be anywhere between "I'd quit if we hired the candidate" to "I'd quit if we don't hire this candidate" with a couple options in between, and no option for "maybe" or "I don't know". Of course people try to get around that by saying things like "soft yes"...
A few related Zappos discussions:<p><a href="https://news.ycombinator.com/item?id=10964404" rel="nofollow">https://news.ycombinator.com/item?id=10964404</a><p><a href="https://news.ycombinator.com/item?id=11239471" rel="nofollow">https://news.ycombinator.com/item?id=11239471</a>
I find the Roman concepts of auctoritas, dignitas, and potestas to be quite helpful for understanding how allegedly flat organizations self organize. Please note that I’m not saying they apply directly, they don’t, but it helps to understand social status is multidimensional.
Sounds a lot like consent (not consensus) as practiced in Sociocracy or Holacracy.<p><a href="https://patterns.sociocracy30.org/consent-decision-making.html" rel="nofollow">https://patterns.sociocracy30.org/consent-decision-making.ht...</a>
My hope is that you can begin by dropping concerns over efficiency.<p>For that matter, an slight inverse of the headline is way more important, to my mind. Avoid slowly making bad decisions. An easy way to help there, is to not turn every decision into a debate.
The article very much needs a link to this:
<a href="https://en.wikipedia.org/wiki/Law_of_triviality" rel="nofollow">https://en.wikipedia.org/wiki/Law_of_triviality</a>
Decisions in a flat organization are made by the informal hierarchy that you have to discover through trial and error. Be careful not to antagonize anyone important while you do it!
I think in "flat" organizations, there must still be a responsible person, but he should convince people to do what is needed to achieve the goal, not force them.
An organization can only be flat if you can’t be fired by someone else or you can fire anyone else. As long someone can do that there’s always that inequality.
Group decision-making > Formal systems:
<a href="https://en.wikipedia.org/wiki/Group_decision-making#Formal_systems" rel="nofollow">https://en.wikipedia.org/wiki/Group_decision-making#Formal_s...</a><p>> Consensus decision-making,
Voting-based methods,
Delphi method,
Dotmocracy<p>Consensus decision-making:
<a href="https://en.wikipedia.org/wiki/Consensus_decision-making" rel="nofollow">https://en.wikipedia.org/wiki/Consensus_decision-making</a><p>There's a field that <i>some people</i> are all calling "Collaboration Engineering". I learned about this from a university course in Collaboration.<p>6 Patterns of Collaboration [GRCOEB] — Generate, Reduce, Clarify, Organize, Evaluate, Build Consensus<p>7 Layers of Collaboration [GPrAPTeToS] — Goals, Products, Activities, Patterns of Collaboration, Techniques, Tools, Scripts<p>The group decision making processes described in the article may already be defined with the <i>thinkLets</i> design pattern language.<p>A person could argue against humming for various unspecified reasons.<p>I'll just CC this here from my notes, which everyone can read here [1]:<p>“Collaboration Engineering: Foundations and Opportunities” de Vreede (2009) <a href="http://aisel.aisnet.org/jais/vol10/iss3/7/" rel="nofollow">http://aisel.aisnet.org/jais/vol10/iss3/7/</a><p>“A Seven-Layer Model of Collaboration: Separation of Concerns for Designers of Collaboration Systems” Briggs (2009) <a href="http://aisel.aisnet.org/icis2009/26/" rel="nofollow">http://aisel.aisnet.org/icis2009/26/</a><p>Six Patterns of Collaboration
“Defining Key Concepts for Collaboration Engineering” Briggs (2006) <a href="http://aisel.aisnet.org/amcis2006/17/" rel="nofollow">http://aisel.aisnet.org/amcis2006/17/</a><p>“ThinkLets: Achieving Predictable, Repeatable Patterns of Group Interaction with Group Support Systems (GSS)” <a href="http://www.academia.edu/259943/ThinkLets_Achieving_Predictable_Repeatable_Patterns_of_Group_Interaction_With_Group_Support_Systems_GSS_" rel="nofollow">http://www.academia.edu/259943/ThinkLets_Achieving_Predictab...</a><p><a href="https://scholar.google.com/scholar?q=thinklets" rel="nofollow">https://scholar.google.com/scholar?q=thinklets</a><p>[1] <a href="https://wrdrd.github.io/docs/consulting/team-building#collaboration-engineering" rel="nofollow">https://wrdrd.github.io/docs/consulting/team-building#collab...</a>
Consensus is the path, not the destination. Implemented it in a project such as:<p>Start using the tech. If it explodes it's on you; if other people pick it up it was good; If nobody picks up and you get tired of fixing it, abandon.