Hacker News new | past | comments | ask | show | jobs | submit login
Why Valve? Or, what do we need corporations for.. (valvesoftware.com)
331 points by liquid_x on Aug 3, 2012 | hide | past | favorite | 159 comments



It is a dubious proposition, in my view, that a "boss-less" company can or should replace the traditional hierarchical corporation of today as the normative form of business organization.

Here are my reasons for saying this:

For all of the economist's sagacity that the author has and evidently brings to this piece, the author's underlying case against modern hierarchical corporations seems to boil down to a populist analysis that is not so much a proof as a set of ill-developed assertions.

The author claims (a) that hierarchical managements lead to "corporate serfdom" and to "Soviet-like" dominance within the framework of the corporation itself, thereby crushing creativity and wasting resources, (b) that all this is made by possible by "toxic finance," and (c) that it is all "co-dependent with political structures that are losing democratic legitimacy fast."

Corporate serfdom? Toxic finance? Co-dependent on illegitimate political structures? This lumps every early-stage startup with every mega-corporation that has ever existed and, in effect, calls them all illegitimate. And that is a political assumption about "corporations" in the abstract, not an empirical analysis, because it cannot possibly be defended as an empirical analysis. Is it serfdom to join a YC company as a founder or an employee? Is YC a toxic funder? If the answer to both is no, does all this change once startups get bigger? How about a startup that purports to offer a different form of corporate culture ala Google? Are there serfs working at Google? Is their funding toxic? Or does this just apply to a Walmart or a Standard Oil or other mega-corporation that does not specifically do creative work in the tech field? Does hierarchical management consist of simply having the normal forms of corporate government - a board of directors and corporate officers - or does it come about only when people are given authority to hire and fire, to supervise the employment of others, and to direct them in what to do in their jobs? Is this all good, efficient, and respectful of human talent and creativity when the organization is small but soulless and deadening and even "Soviet-like" only when the corporation becomes large? If there is such a distinction, where is that line crossed? And does this mean that the corporate form is not innately evil but that a large organization of whatever type, organized hierarchically in its management structure, is what brings in the evil.

What, then, does Valve offer that makes it different? It too is a corporation. It is privately owned by a few persons who have had the luxury of screening all employees so as to hire only very bright, highly self-motivated persons to do predominantly creative forms of work. Working with such employees, Valve has been able to build a successful model by which these bright, motivated employees get to choose 100% of their projects and have complete freedom on how they manage their own time and on what results they seek to achieve. It all sounds like an amazing work environment but how many businesses get to focus in this way on creative forms of work or get to screen carefully to make sure they only hire self-motivated employees? And how many businesses have the luxury of doing this without needing to raise outside capital through their early stages? Moreover (and the author himself raises this point), to what extent can this scale? Can such a model work if the company grows a thousandfold and suddenly has 40,000 employees? Of course, the model inevitably breaks down at some point along the way because the environment in which the Valve employees currently function is highly unusual if not unique.

Unless human nature should radically change owing to technological progress (a dubious assumption in my view), we can continue to expect that, in any large group, there will always be those who fail to carry their weight, those who seek to take advantage, those who are incompetent, and those who are plain bad actors making life difficult for those around them or trying to cheat the company or steal from it or whatever. A hands-off management that lets all such persons do whatever they want will very quickly find itself immersed in problems and, ultimately, some mechanism needs to be put in place by which employees are managed, are disciplined, are rewarded, are redeployed, etc. in ways that conform to the goals of the organization and not necessarily with those of each individual actor within that organization.

Every form of business organization needs people with a vision to set its model and its goals and to direct people and resources in a way that maximizes the opportunities of successfully reaching those goals. In some situations, some or even all of the impetus for this can come from those who work in common without an overriding authority. Those situations, though, are by definition highly unusual at best. Valve may be one of them and even then it has to managed at some level even by its benign oligarchs who own it. And, even if technological progress could someday supplant the need for corporations, this piece does not make the case for how that will ever be possible. It is, then, an intriguing piece (with thought-provoking elements) but suggestive and incomplete at best in its main argument about Valve and marred by populist assumptions in its broader themes about corporations generally.

Yes, this is written from the standpoint of a startup business lawyer who has dealt with corporate forms of organization for some three decades now. This may give my views an inevitable bias in that direction but it also gives me a close familiarity with how such corporate forms work. From that perspective, what the author says just doesn't ring true. Business organizations generally aren't places where free-flowing creativity will hold sway above all else. The Valve model may be great but I just don't see it being made broadly applicable to the vast majority of businesses as they operate today or as I can even conceive of them operating in the future.


I think that bossless firms can and will replace traditional hierarchies in many areas. However the two are more different in their limitations than they are the same. Bossless systems are not drop-in replacements for hierarchical systems.

If you look at bossless firms that are successful, they come in sizes ranging from GitHub at founding to WL Gore (3B in revenue, 30k employees). Scale in terms of firm size is not an issue. However every one of these companies, is narrowly focused. Every one has a narrow mission. There is no way to replace Microsoft or Dupont with a single bossless organization any more than Gore has replaced Dupont. The scale limitation for bossless organizations is that of bredth of focus. Without a narrow focus, whether it is a specific kind of plastic or video games, there is nothing to organize around. Without this, there cannot be spontaneous order.

A way to think about this would be the difference between agriculture and permaculture. Large firms of today are like giant factory farms. They are highly ordered by human standards, but in so doing, the complexity cost is high and so productivity is reduced. In a permacultural environment, the scale may be smaller, the order is organic rather than artificial (curves instead of straight lines, multiple species together instead of monocultures, etc), but they are more productive with fewer human inputs. The complexity cost is low.

A bossless organization that works well does so because it channels human nature in the right directions. As William Gore, founder of WL Gore, said, every successful organization has a lattice structure, which may or may not be aligned with its hierarchy.

It is true that these organizations have two major challenges, namely hiring and firing, but these are not problems which don't admit of solutions. The solutions are just different than they are in a hierarchical model.

I do think that bossless organizations will rise up and take over large sections of the economy, but if they do they will be smaller than the companies they are replacing. We will have a more just economy in large part because it will be dominated by a larger number of players and hence it will be more competitive. Ecosystems will win out over machines in the end ;-)


Thank you for this comment, this explains quite well something I have not been able to put clearly myself. Whatever organization a firm decides to follow, it needs 'something' to coordinate efforts within itself. Hierarchy is one way, having very clear focus (and a culture shared by all) another one.

That's also one of the reason most big and successful open source projects manage to get away with much less explicit organisation.


i hadn't heard of wl gore. thanks for the pointer, they look like a very interesting case study.


I got their number of employees wrong. They "only" employ 9000 or so.

You may not have heard of them but I am sure you have heard of at least one of their products: Gore-Tex(R).


Hierarchy kills creativity.

I just started working a few entry level IT jobs and I am seeing how sad this really is. A change request to restart a server takes 4-weeks and is turned down by a CTO who doesn't understand IT needs. The CTO implements a new policy without discussion with IT which renders a passion project by members of the company to be scrapped.

Startups are corporations. They work just like one. Sure they offer perks like free pop, food, paid lunches, breaks, naptime, pool table, ping pong table and so on. This isn't a management style, its a employee benefit.

When it comes to management, startup companies seem to approach business in the same way. A CEO who has the CTO/CFO/CIO reporting to them. Then there is someone else who reports to someone else who reports to someone else. It's a cycle of killing ideas. If one person doesn't like the idea somewhere along that chain, probably at the top somewhere, it won't get any traction. If ideas keep dying employees become discouraged to use their spare time to implement new procedures to decrease inefficiencies.

[i]"How many businesses get to focus in this way on creative forms of work or get to screen carefully to make sure they only hire self-motivated employees?"[/i]

I am sure Valve hires its fair share of garbage. By garbage I mean employees who last 3,6, 12 months before they find themselves looking for a job. In the start it must take long to adapt and understand the operating procedures and workflow. But if you are a person dedicated to getting something done you will implement something or join someone who is working on something awesome. In that first year you will show value and if you don't the rest of your employees will tune you out. Another thing is their hiring procedures are probably much better than an organization. Instead of having HR and then a manager from the department for hiring or a selected individual it's probably more spread out and allows all employees to interview new employees to see if they would fit their teams.

[i]Every form of business organization needs people with a vision to set its model and its goals and to direct people and resources in a way that maximizes the opportunities of successfully reaching those goals. [/i]

This is the traditional thinking where there is one person who knows more than everyone. If this was the case we wouldn't be inventing and developing all these new tools. One person doesn't have the answer.


Two basic truths:

1) An ownership society, one where the workers direct their own production as if they own the business, increases a sense of responsibility.

2) if you tell people what to do, most will do not much more than the minimum they think is necessary.


In addition to what you are saying, "boss-less" organizational structures are not exactly new. This structure has been kicked around and implemented with various levels of success since at least the 80's. It is a safe bet that most people's new ideas are some other person's old ideas. GE's Durham aircraft engine plant has a single manager for the 170+ employees that work in the plant, and the employees are responsible for almost all decisions, since 1993. In fact, the first plant manager believed she made no more than 12 unilateral decisions a year. The result was significant cost savings over other GE plants, with the entire plant working to reduce costs so they could win more contracts in the internal bidding. In other words, GE plants would bid within GE to manufacture new engines, so if the plant performed better, it was less at risk of being shutdown. Within 2 years defect rates fell by over 75%, and costs over 10% every year. Simple optimizations were made constantly; in one case an employee showed up to work with a bunch of fishing tackle boxes he bought, then spent the morning packing them full of all the small hardware necessary to assemble part of an engine. In another, the plant employees designed and erected specialized scaffolding, reducing time to assemble each engine by about 8 hours. Everyone is also required to change teams constantly so that there is cross-pollination of ideas, with the result being almost everyone in the plant can do every step in the process.

There are other examples too, like Honeywell's honeycomb structure, which works very much like the GE plant, except in the field of power plant operations. At one point, Honeywell took over a plant in Thames, and the executives observed the employees. They would watch them do things in a hopelessly inefficient way, and ask them why they do it that way. The response was almost always because "that's the way 'they' make us do it". Honeywell executives declared war on 'they' in the plant, literally having actors come in dressed up in 'they' shirts, who they hunted down and kicked out of the plant. The point was to drive home that there is no 'they' anymore, the people that work there are responsible for everything that happens. Case in point, some time after that, a regular bottom of the rung employee fired a contractor on the spot for severe safety violations. The contractor went to the plant manager in protest, asserting that the employee couldn't do that. The manager's response was along the lines of, "it appears that they did." Of course, all these good things must have some downsides. In the case of Honeywell, there was a disastrous coverup that went undetected for some time caused by oversight failures related to nuclear waste reading. An employee took full ownership of the reading process, and wouldn't let anyone else near it, after they detected a bad reading, because they believed they may be fired for it. There were massive NRC fines, though the situation never really became dangerous. To be blunt, not only are these not new ideas, there are well known major issues that can arise from them, and there is a lot of information available about the potential issues and there mitigations.

Another fun example from my own life, this time from NASA, with some key details changed without altering the narrative. One day, my mother, who works at NASA on the assembly of spacecraft, called me and started asking me engineering questions related to the heating and cooling cycle of different kinds of compounds, and the effect they could potential have on bonding. She had some concerns about a practice that had been going on at work. The assemblers were freezing a thermal epoxy compound before applying it because it was a bit easier to apply that way. She believed it could be an issue because of expansion and contraction, and asked them to stop. They didn't, so she went to their supervisor, who did not believe it was an issue. She also went to her boss, and her bosses boss, both of whom failed to respond. She also summarized her concerns to both in writing. From what she was describing to me, I told her it did not sound like she was overreacting, and that bonds could potentially fail in space, but that I couldn't really say what the risk actually was without more information.

At this point, she had three options. She could do nothing, since she had already covered herself and reported it, she could go to the head of the program, or I suggested she could go tell the lead thermal engineer exactly what she told me. She told the lead engineer first thing the next day, and he flipped out. The compound was only certified for application at room temperature. Work stopped, and he went with her directly to the head of the program. Each tube of paste about the size of a tube of toothpaste costs about $10,000, and there were about 10 tubes in the freezer alone that would have to be discarded. In addition, the practice had been going on for months, all the work might have to be undone. In this case, the resolution was not negative. The engineer was able to certify the compound for application at the reduced temperature, after destructive testing, so there was no significant loss. That won't always be the outcome, it could have gone very differently, with either a failure in re-certification or in space. NASA has had other failures from similar issues, most prominent being Challenger. In this case, it's not a boss-less environment, but a complete management failure, approximating a boss-less environment (supervisor was waiting to retire, boss was not coming to work, but having secretary turn on computer, open office door, and place coffee on desk so it would appear he was there, etc).

The overall point is, this stuff isn't new. Just because it works for Valve, doesn't mean it will work for everyone else. Valve does not have a huge need for internal controls because of the nature of the business, and ultimately of the product. Smart and motivated people want to work at Valve because they want to make games. No manager at Valve, and somebody gives a friend a bunch of free games. Worst case, they steal the password database and customer credit card numbers. If the same structure fails at Honeywell, there is a major safety violation and a bunch of people die. It could be downright dangerous to apply these ideas without fully exploring the potential consequences of doing so. Further, these types of management structures are well known, and there has been some significant research into the types of problems that can come up, and what can be done about. GE, for example, does extensive mandatory training at Durham on compromise, conflict resolution, and group decision making.


Great article. Anyone who has worked for a large, established corporation knows from personal experience that internally they are a lot like the Soviet Union, with a hierarchical structure of bureaucrats and their apparatchiks making decisions for everyone at the company.

Despite having grown to around 400 employees, Valve is evidently not like that. The author, Yanis Varoufakis, makes a compelling case that future companies may look more like it than like the traditional hierarchical corporations of today.

--

PS. As someone who regularly reads the author's economics blog at http://yanisvaroufakis.eu, finding him on Valve's corporate blog was the source of quite a bit of cognitive dissonance for me. ("What the...? Why is Varoufakis showing up on Valve's corporate blog?" was my immediate thought.) I had to do multiple double-takes before it dawned on me that, yes, he somehow works at Valve!


One has to wonder why "large" companies even exist in the first place. Why a building full of freelancers could not do the exact same thing.

And then it dawned on me: Valve is providing a "firewall" to unfair competition, financial insecurity, brand awereness, and legal risks. And then it just gets out of the way.

Maybe, the pure existence of large corporations, is a sign of capitalism failing. Being large should not give an advantage: yet it allows for a reportaire of anti-competitive behavior.

Maybe they should do a nation wide experiment, in smaller country: what happens if you take away the legal standing of all "companies" and "employment". What if all economic relationships can only exist between individuals?

Im not suggesting its some kind of magic bullet, i just wonder how competitive it would be. Will such a system adapt (to changes in the market) and optimize (cutting out middlemans) more quickly? Will it be truly bottom-up?


Take a look at http://en.wikipedia.org/wiki/The_Nature_of_the_Firm

Given that "production could be carried on without any organization [that is, firm] at all", Coase asks, why and under what conditions should we expect firms to emerge? Since modern firms can only emerge when an entrepreneur of some sort begins to hire people, Coase's analysis proceeds by considering the conditions under which it makes sense for an entrepreneur to seek hired help instead of contracting out for some particular task.


This is an excellent answer. Peter Thiel uses some similar ideas in his answer about why people form companies: http://blakemasters.tumblr.com/post/20400301508/cs183class1 :

If we want technological development, why look to companies to do it? It’s possible, after all, to imagine a society in which everyone works for the government. Or, conversely, one in which everyone is an independent contractor. Why have some intermediate version consisting of at least two people but less than everyone on the planet?

The answer is straightforward application of the Coase Theorem. Companies exist because they optimally address internal and external coordination costs. In general, as an entity grows, so do its internal coordination costs. But its external coordination costs fall. Totalitarian government is entity writ large; external coordination is easy, since those costs are zero. But internal coordination, as Hayek and the Austrians showed, is hard and costly; central planning doesn’t work.

The flipside is that internal coordination costs for independent contractors are zero, but external coordination costs (uniquely contracting with absolutely everybody one deals with) are very high, possibly paralyzingly so. Optimality—firm size—is a matter of finding the right combination.


This is pre information revolution thinking combined with ideas generated by the fears of an age when socialism/communism was really was competing with democracy as a form of social organization. This muddled thinking makes no sense presently, especially given Corporations are purely creations of the government monopoly of power and should be anathema to any free thinker. For those not fighting the last war led by generals like Hayek, all coordination costs follow the technological cost curve as there is no longer a difference between internal and external actors.


I almost can't tell if this comment came from some sort of Markov text generator.


I understood it just fine. Guess different people have different parsers.


> all coordination costs follow the technological cost curve as there is no longer a difference between internal and external actors.

That's not true.

For one, internal actors have pre-arranged deals.


This is great. Coase Theorem explains governments as naturally large corporations, or at least an entity similar to a for profit corporation. Brilliant. Yes, governments create laws; unlike other corporations, they have surpassed the threshold of size (not sure whether to measure that as number of employees, money, information, etc) and support to and been given one of the ultimate money makers in society. Laws exist to direct wealth, property, and rights. They produce laws and regulation for the voter population to consume.

However, the "inequality" on the Coase spectrum is slowly balanced out as we gain access to decentralized technology. I bet that pretty soon company set-ups, like Valve's, will become much more prominent. I just hope the same thing happens with governmental institutions.


It is already happening, though very slowly. Part of the city's budget here is decided online and via presential polls [1] since 1989, and the model has since spread to cities like Brussels and Barcelona.

[1] http://en.wikipedia.org/wiki/Participatory_budgeting


Would it be fair to say that "costs of transactions" are in some part due to government bureacracy and taxing?

And could it be, that certain structures are less optimal (profitable) for the single "firm", but more optimal for the economy as awhole. Could there even be a distinction?


Would it be fair to say that "costs of transactions" are in some part due to government bureacracy and taxing?

Some costs of transactions are due to the need to document compliance with government regulations. The existence of regulations with compliance burdens will tend to make firms bigger than they would otherwise be, but many and probably most transaction costs aren't due to government, and there would be many additional sorts of transactions costs if there weren't a government at all.

And could it be, that certain structures are less optimal (profitable) for the single "firm", but more optimal for the economy as awhole. Could there even be a distinction?

It probably could, depending on what you mean by "optimal". Its pretty well established[1] that in the absence of transaction costs you can get a Pareto efficient[2] outcome, but Pareto efficiency is probably not the same thing you would call optimal (though what you'd call optimal is almost certainly Pareto efficient).

I would tend to expect that, given that we live in a world with transaction costs, firms are not the ideal size. But its hard to say whether firms ought to be bigger in general or smaller. Countries with relatively efficient economies (Germany) tend to have larger firms than countries with less efficient economies (Italy) but its really hard to tease out in those cases which way causality is flowing. Most countries tend to have laws that favor small businesses, and those might end up more than making up for their proportionally higher regulatory burden.

[1]http://en.wikipedia.org/wiki/Fundamental_theorems_of_welfare... [2]http://en.wikipedia.org/wiki/Pareto_efficiency


And could it be, that certain structures are less optimal (profitable) for the single "firm", but more optimal for the economy as awhole. Could there even be a distinction?

Sure. The extreme example is competition vs a monopoly. It is certainly more profitable for a single firm to be a monopoly and be able to set prices. It is always less optimal to have real competition, but it is always better for the economy as a whole.

Taken to the extreme of course, the economy might be best if perfect competition ruled everywhere, which would mean largely a nation of small businesses, artisans, and shopkeepers. But this would be bad for Microsoft, as it would likely mean networks of open source developers controlling the software market.


Yes. When members of a firm create value for each other, they don't pay sales tax. It is huge unfair advantage for large firms.

There is also the huge savings where teams have an executive authority to settle disputes and prevent them from trying to rip each other off with unfair contracts and nonpaid bills.


> Yes. When members of a firm create value for each other, they don't pay sales tax. It is huge unfair advantage for large firms.

This is very, very close to the sales pitch of a VAT. In a proper VAT system, there is effectively no difference between purchasing an intermediary good or service or producing it yourself.


In the US services aren't taxed. Most internal value created is in the form of services (e.g., IT). Most companies don't have their own internal computer production facility, for instance.

Your second point makes sense. There are usually much lower coordination costs for internal agreements.


And in countries where services are taxed, the tax is generally rebated for intermediate goods.


in the US, I pay sales tax to the people who do repair work on my house and who fix my dad's computer, and for the software I get from Apple engineers, all of which are not taxed within a firm.

http://dor.wa.gov/content/findtaxesandrates/retailsalestax/

"Similarly, when a business purchases a retailing service for its own use, it must pay sales tax on the purchase." "


I have always interpreted http://apps.leg.wa.gov/wac/default.aspx?cite=458-20-155 regarding computer repair as the idea that if I am replacing a piece of faulty hardware that I must charge sales tax for the attached service, but if I am merely cleaning off viruses, that this is a professional service and therefore not subject to sales tax. I have not been able to find a clear authority for this however, and it isn't clear as to whether hooking up and configuring an external modem that the customer bought from someone else is taxable.

Repair however is a problematic term. A lawyer might be involved in sending a letter threatening action of a problem is not repaired but that service is not taxable itself.

In Washington State, professional services not directly tied to an improvement of tangible property are not generally taxable. So landsacping and lawn care is generally taxable, but open source software development is not (since it is not a sale of a license, or a one-of-a-kind development aimed at one customer only).


I think it's pretty clear from the pro-market critiques of Capitalism (from the Distributists, like Chesterton or Belloc) that pure, unadulterated Capitalism inherently unstable and prone to failure. Large corporations paradoxically need to pay their workers only what the market requires but also enough they have discretionary money to go out and buy the products they are manufacturing.

Socialists recognize this inherent paradox too, but their answer to it is different --- more government intervention to counter corporate control. To a Distributist, however it seems insane to try to address issues of centralized economic control by adding additional centralization.

The existence of large corporations introduces inherent instability into our system. We'd all be better off if there were more self-employed and more small corporations out there.


Large corporations paradoxically need to pay their workers only what the market requires but also enough they have discretionary money to go out and buy the products they are manufacturing.

This fails the test of basic arithmetic.

Suppose you increase your workers pay by $X and they spend the entire $X on your products. Suppose your profit margin is Y. Then for every $(1+Y) your workers spent, you earned $Y. I.e., you get back $XY/(1+Y) < $X.

So no, it's always better for the bottom line (holding all else equal) to pay your workers less. You might be able to attract better quality workers by paying more, or perhaps you can cook up incentive plans to get workers to work harder (both of these are what Henry Ford did, for example). But that has nothing whatsoever to do with workers buying your products.


But the point is that if everyone pays workers less there is no market for anything non-essential you can make. This works if you are making plastic stuff to export to the US and you are in China. It doesn't work so well if you are in the US.

In essence corporations individually do better if they pay their workers less, but collectively they die.


But the point is that if everyone pays workers less there is no market for anything non-essential you can make.

Why not?


The rise of the bureaucratic corporation and the rise of a white collar managerial class during the industrial revolution was a natural pre technology evolution of the corporation because lots of people we required to carry out all the necessary communication and managerial roles. With the information revolution that is no longer true, but old ways of thinking die hard and it will take a while for new organizations to develop and establish themselves. This will be a long a period of creative destruction.


I suggest you read this excellent piece my Cosma Shalizi. http://crookedtimber.org/2012/05/30/in-soviet-union-optimiza...

There he talks about the computational complexity and limits to planning. He also talks about markets, how non convexities leads to inefficiencies (allowing polarization of wealth) and why large firms form: difficulty in setting prices and being informed enough to make optimal decisions benefit concentration of complementing talent.

As computational power increases and communication friction decreases I expect we will start to see company size decrease and lifetimes to within an order of magnitude of the sceanrio in alluded to in Rainbows End where companies are old at 5 years and large ones have 3 people.


I think the ultimate anti-Valve is a military and everything else falls in-between.

Some businesses are required by nature to be military-like - logistics for example.


And the military itself, of course :P.

This actually reminds me of an interesting observation in Homage to Catalonia by Orwell. The book is about Orwell's personal experiences in the Spanish Civil War. The particularly interesting bit was about the organization of some of the anarchist forces fighting against Franco. Essentially, they did not have any centralized organization at all. Even a military force can be organized like Valve and not in an authoritarian manner.

Unfortunately I don't have the book handy to find the pertinent passage, but it really stood out. It makes me wonder how much of the military-style management is due to necessity and how much is due to tradition. Of course, the fascists won in Spain, so this is hardly a great example, but there were too many other factors at play (e.g. support from Germany) to make it conclusive.


Arguably, Anonymous acts like a military with Valve-style management as well.


I don't buy that the military is required to work that way.

Take a look at Iraq and Afghanistan to see how many losses are inflicted on the superior military by groups who are not organized in any traditional military sense.

Do you think they would be more effective if they wore uniforms and had a formal command structure?


Regarding your PS: Valve hired him as "cheif economist". This article explains it all:

http://blogs.valvesoftware.com/economics/it-all-began-with-a...

I seem to recall an HN discussion about it, but am too lazy to track it down :).


Thanks. Also a great post.


This part is particularly enlightening:

The idea of spontaneous order comes from the Scottish Enlightenment, and in particular David Hume who, famously, argued against Thomas Hobbes’ assumption that, without some Leviathan ruling over us (keeping us “all in awe”), we would end up in a hideous State of Nature in which life would be “nasty, brutish and short


you can read the story about how he started working with valve on this post: http://blogs.valvesoftware.com/economics/it-all-began-with-a...


I'd like to see if Valve's model scales to a much larger number of employees.


It may not need to; they could take the approach of splitting the company into several smaller companies that are each an ideal size for this management style. I don't recall the name, but there is a manufacturing company (clothing, I think) that does this. Collectively the company has thousands, maybe tens of thousands of employees, but they're organized into independently run shops with around 500 people each.


Making the parallelism of firms as federations (not sure if is a correct assumption), this can surely relate with Proudhon's ideas about federations and confederations: http://www.ditext.com/proudhon/federation/federation.html#7


Gore dOes this


Morningstar Farms does this as well.


How big is much larger? 9k employee businesses exist on this model.

The scale issue is NOT number of employees. It is breadth of focus. You could have an operating system company, or an office productivity business, or a video game console company, but it is hard to see people organizing Microsoft in a bossless company without breaking it up first.


Capitalism doesn't require corporations. Corporations insulate capital from liability.

In the early 1800's in the US, you needed a act of the state legislature to form a corporation. This was difficult and came with other baggage (putting politicians on payroll, etc) so many more businesses were partnerships. Being a partner means that as you invest, you gain more equity AND more liability. It also means that governance becomes difficult as the partnership grows. As the industrial revolution brought about massive, more capital intensive businesses (railroads, steel, etc), the corporation become necessary to function.

If Hacker News was around in 1880, we'd be talking about corporate bureaucracy as a great innovation. It made sense.

I think what you're really seeing with Valve is a sort of modern partnership. I've seen similar sounding small businesses (farms, mostly) where running of the business is more consensus-driven or there is a "spontaneous" hierarchy that develops over time.


There's not really much of a connection between the ownership structure--corporation vs. partnership--and the management structure of a company. In Valve's case it is probably organized as a corporation, but it actually has only a few owners. It could easily be organized as partnership. Either way, what makes it special is that the employees (who are not owners) are allowed to self-organize. But that could happen under any ownership structure.


Well, if you define Capitalism as Capitalists hiring Labor, then yes, it does. A market economy could exist in ways organized through divisions other than Capital/Labor and as such the market does not need Capitalism to continue.


I appreciate the insight in the OP, but I'm frustrated by the imprecise use of language. Valve is hierarchical. The hierarchy is simply informal and spontaneous based on the attributes of individual team members. (The OP hints at the distinction with the occasional modification "authoritarian hierarchy", but this line is often blurred.) In addition, the notion that Valve is "bossless" is disingenuous. For every person at Valve, there is some other person (or persons) who can fire him. If you worked at Valve, that person could in principle tell you what to do. That he doesn't have to is a product of the kind of people who work at Valve, not any radical innovation in corporate structure. Although its organization may be relatively flat, Valve is strictly hierarchical by any sensible definition of the term.

Valve's model reminds me of the old Costco vs. Walmart debates. Costco, we are told, get lots more productivity out of its workers by treating them well—good pay, generous benefits, etc.—while Walmart suffers by comparison. What this analysis ignores is that the people are different. Costco has discovered that it can thrive by compensating disciplined, productive people well. Google does the same. Apparently so does Valve. But not all people have such discipline or high productive capacity—you couldn't just swap all Costco employees for the same number of Walmart employees and expect to get the same results, any more than you could with Google or Valve. The miracle of Walmart (or one of them, at least) is that they manage to thrive using the labor of people whose productivity is often marginal by the standards of Costco or Valve. Arguably, that is even more impressive, and perhaps more laudable. Unsurprisingly, Walmart's corporate structure is very un-Valve-like.


Indeed, every person is different, and every job is different. Corporations already create a huge diversity of jobs that can't be equated across companies. I can hire 100 programmers and say "your job description is to do whatever you want," and there is nothing inherently anti-capitalist or post-capitalist about my company. If I can make it productive by virtue of cultural and social pressures alone rather than bossing people around, that's great.

As the OP points out himself, corporations are "islands" as far as the larger economic system. There aren't really inherent constraints on how you organize them internally. There are handed-down "best practices" for management that are probably long overdue for rethinking; for example, perhaps the notion of today's "boss" or middle manager will come to seem antiquated. But no company runs like a free market. The OP spends a lot of time identifying "traditional capitalist corporation" with bosses and hierarchy in order to paint Valve as less capitalist, but really it is just less "bossist."

To say Valve is "more radical" than a co-op with distributed ownership is ridiculous, and misses Valve's own hierarchy (as my parent points out). Owning and running a company is not just a matter of telling people what to do. It includes hiring and firing, leading through culture and other channels, making executive decisions and generally keeping the lights on. As a Valve employee, can you fire the janitor if he isn't doing a good job? If the building is hit by a tornado, is it your responsibility to get it rebuilt and eat the cost? Being an owner isn't all peaches and gravy.

Could you decide you want a non-wheelie desk or to telecommute full-time from the beach? Hopefully you don't want to. Could the owner decrease your compensation or relocate the company offices to Kalamazoo? Hopefully he doesn't want to.

The best management is invisible.


"You can do as you will but not will as you will"


I think you and a few others are selling Valve and the article a little short. Varoufakis does give a precise definition of what he means by non-hierarchical: free allocation of individual time. How is that not a radical innovation? Nearly every other corporation is a tyranny by comparison. And it's not just that they happened to hire people who work this way. It was clearly a conscious design decision and probably couldn't have gotten started any other way.

(BTW, I don't think you can have an "informal and spontaneous" hierarchy, or a "flat" one. Those are oxymorons. Hierarchy by definition is a formal gradation of ranks. No?)

Varoufakis is careful to distinguish this management structure from Valve's ownership structure, which is hierarchical. Many people have pointed out that ultimately there has to be a boss at Valve: the owner. He implies this too when he points out that its management structure depends on the "enlightenment" of the owners "not using their property rights to boss people around". So yes, whatever this experiment is, it isn't communism. That doesn't mean it isn't radical! I think we're hypnotized a bit by having been raised with "capitalism" and "communism" as the only two poles of a binary system, where communism was an obvious failure so therefore the status quo is the only possibility... (not implying that this is your view, just that it's in our water). Besides which it's not obvious that Valve's structure is anti-capitalist. The way Varoufakis describes it, it's an extension and generalization of capitalism: extension because market dynamics are being applied to what has traditionally been monolithic bureaucracy; generalization because something other than price – namely, time – is being used inside Valve to signal value. (Anti-corporatist, sure. And yay! But I digress.)

Ownership must have an impact on how the company functions, so there are unanswered questions here. But let's not naively assume that ownership trumps everything. Suppose the owners decided tomorrow, "We tire of this experiment. The experiment is hereby changed to neo-Stalinism." Would people at Valve glumly bow their heads and say, "oh, ok"? Likely not. Some would leave; others would resist. It's not clear what would happen. So yes, vertical ownership does imply power - but not absolute power, except perhaps to destroy the whole thing.

Seems to me the big unanswered questions are: (1) relation of ownership and management structures, (2) how hiring and firing really work. The latter is the most important use of management power, so it's the big test of how non-hierarchically Valve really operates. Especially firing. No matter how good their "filters for self-motivated people" are, they can't be perfect. When firing is necessary, how is it really done? A lot of people seem to be imagining that it must work like this: you get called into Gabe's office and Gabe says, "we're not that anarcho-syndicalist, you're fired." That would indeed be hierarchical. But we don't know this. And even if it were true, it wouldn't completely destroy the experiment.

I don't think their claims are quite as utopian as some of the more cynical/debunking commenters assume. Thinking about this as "utopia or not" makes it less interesting.


I don't think you can have an "informal and spontaneous" hierarchy, or a "flat" one. Those are oxymorons. Hierarchy by definition is a formal gradation of ranks. No?

Humans naturally form dominance hierarchies even in the absence of formal rank. If Alice, Bob, and Charlie all work together on a project, typically one of them (say Alice) will quickly emerge as group leader. She will act (if necessary) with executive authority, even in the absence of the formal ability to hire and fire.

A "flat" organization is simply one that has few formal hierarchical levels. There is no essential conflict with a strict hierarchy. An army with one general and ten-thousand privates is both flat and perfectly hierarchical. Real militaries find it advisable to have a large number of formal ranks, but the essential notions of unified command and strict hierarchy are common to any well-managed corporation.


"General" and "private"? That's two levels, not flat. But ok, so we're obviously using our terms differently. There is still a radical difference between an organization with fixed hierarchical roles and one in which individuals allocate their own time and respect only such authority as emerges during the work itself.

There's another point. In my experience with emergent authority on teams, even if Alice emerges as group leader, that hardly gives her "executive authority". If she begins to insist on making arbitrary decisions, she will soon be knocked down a peg and carry less respect. In other words, these forms of authority (emergent vs. executive) are not the same.


They're somewhat separate, but there is substantial emphasis in current management research on trying to blur them. The worry is that classic executive authority looks increasingly heavy-handed, but the executives still want control. So the goal is to design a workplace such that they get what they want while appearing not to be in hierarchical control. That's a bit difficult since it involves solving a "reverse emergence" problem: how do I set up my workplace so that the emergent authority will tend to emerge in the ways that I want it to? This can involve all sorts of subtle and less-subtle levers: hiring practices, office layouts, perks, "gamification", performance reviews, promotions, etc.


I'm using "flat" in the standard sense: http://en.wikipedia.org/wiki/Flat_organization


So there are two things that remain unclear from this analysis:

1) Moving between Valve projects may be like a "market" where the "buzz" around a project is like its "price"... but then the analogy breaks down, because in a real market, you'd have to pay to be on the hot team.

But at Valve you don't, so in this case, why doesn't everyone just drop their project and move to the hottest, most interesting team? Obviously in real life not everyone will, but if a team only needs 5 people, and 50 people want to join, who determines who really joins? Well, the project manager will have to choose, and now we're just back to managers choosing. Or am I missing something?

2) There's a lot of grunt work in software development. Bugfixes to maintain a year-old product, writing documentation, etc. If nobody wants to do the grunt work, then who does it? Because there's usually more grunt work than people who want to do it. Everyone wants to work on the new exciting sexy stuff, but that's not always what generates revenue and pays people's salaries.


This culture was pretty active at Google and it had some interesting downsides. The fact was that people were recognized for shipping and not for 'finishing' and so a lot of shipped but unfinished things clogged things up. When a bazillion people were being hired, new hires would be dumped on a project as their 'starter' project, and the smart ones would figure out it was a dead end and move on quickly to either their own project or one that was about to ship.

During the crisis in 2009 when hiring froze, and the new hires dried up, this became unsustainable and folks had to be 'incented' to work on less glamorous jobs. Really good project managers made a group feel good and the community aspects kept people around, but they hadn't solved the problem before I left.

I would expect that for games or anything which has a relatively short life after being shipped, you could extend this sort of thing for a long time.


I'm interested in the distinction you draw between shipping and finishing. Care to elaborate on what that meant in a Google context?


Vertically integrated web based businesses like Google have a tremendous advantage that comes from being able to prototype something and then instantly have it available to all customers with a simple push [1] out to production.

There were of course different levels of thing, Gmail was a 'big' thing, updates to the calculator one box, were 'small' things. But everything has a certain amount of technical debt [2] associated with it. This was especially true of infrastructure changes as there were often many subsystems that might be affected in one way or another and each of them usually had some sort of 'porting' effort to get with the program as it were.

"Shipping" was getting something running in production without having it be kicked out by SRE [3]. "Finishing" was having everything that was affected by the change that had been in production before shipping fixed and/or updated to run with the new system. As with most large systems I've had exposure too there is a big chunk of obvious stuff that is affected and then it asymptotically approaches zero. People seemed to start leaving at the 3dB [4] point.

[1] "pushing" is the process whereby a new capability is delivered into the production clusters that are customer facing.

[2] "technical debt" is the requirement that additional work is going to be required later (the debt) to resolve issues that aren't show stoppers initially.

[3] SRE see "The Roads must Roll" by Robert Heinlen.

[4] "3dB" is 3 decibels. In engineering or control systems the 3dB point is where the signal has been reduced in strength by 1/2. Also known as the 'cutoff' point for a filter.


> Well, the project manager

There are no project managers. The closest thing Valve has to them are nothing but glorified receptionists. There is no-one who can do the decision to not accept a team member.

If you flood into a project that is already overstaffed, you'd be hard pressed to find any actual work to do. Which means that you're not producing much, if any, value visible to your co-workers, and those are exactly the people who will be evaluating you.


One thing to consider is that the teams may not be totally ordered by "buzz". So everybody won't naturally gravitate towards the team with the most "buzz" because no team has the most "buzz"! There are also other factors: everyone has different preferences and teams are interesting not just based on what they do but on how they do it and on the people comprising them.

There is, of course, a decent amount of grunt work. However, a lot of programmers don't just enjoy programming as such but also enjoy making something cool. I know I'm certainly willing to spend time ironing out the bugs in my hobby projects because having something that works well is just as fun as writing something novel. Also, I suspect that everyone at Valve puts effort into minimizing the grunt work as much as possible, meaning they probably have far less really boring things to do than most other companies. So a bit of cleverness and forethought combined with a passion of creating something of quality can get good programmers to do the grunt work.


> Also, I suspect that everyone at Valve puts effort into minimizing the grunt work as much as possible, meaning they probably have far less really boring things to do than most other companies. So a bit of cleverness and forethought combined with a passion of creating something of quality can get good programmers to do the grunt work.

This is a big difference between big and small organizations. People will sit in a big org chart and do the same maintenance code-monkeying for a paycheck, but people on small passionate teams have the freedom and motivation to automate their way out of mindless code-monkeying.


I've been wondering about the same questions as well. I hope someone can answer.


In many ways, I think that a worker's cooperative is a much better legal structure for a software firm than a corporation.

I assume that Valve generously grants stock options to its employees, like most other Silicon Valley firms. This, combined with its interesting management structure, makes Valve a worker's cooperative in practice but not by law.

I suspect that if Valve lived in a jurisdiction with legal protections for worker's cooperative, that it would be one.


If I would have finished reading to the end before commenting, I would have had my questions answered. Valve does not grant ownership to employees. However, excess profits are distributed to employees for the most part, which does make Valve a de facto worker's cooperative.

Yanis does discuss the standard arguments against a worker's coop: malfeasance, access to capital, and scalability.

I grew up in an area (Saskatchewan), where for a long time the largest "company" was a cooperative. Access to capital is probably the main reason why Silicon Valley firms are not worker's cooperatives. But that can be solved by granting worker's cooperatives the same legal privileges as corporations. Instead of selling shares, they sell "bonds" which do not pay interest but instead pay a fraction of profit. This is in practice no different than buying insubordinated shares, like those sold by Facebook or Google. Such shares do not give you a "proper" shareholder vote, voting shares are held by the founders.


missed the edit window. It's "subordinated shares", not "insubordinated shares".


I find it really interesting how communist many start ups are, especially considering so many of them involve libertarians.


I think the command economy part of communism (the governance form) is probably the biggest problem libertarians have with communism. I don't think they have anything against equal partnerships between large numbers of people.


Oddly, that part came mainly from capitalists; the early-20th-century Marxist idea of "scientific management" of a centralized industrial economy was closely modeled on scientific-management ideas from large industrial conglomerates.

There was a period when large industrialists were arguing that having many competing companies is inefficient, and instead managing a large, integrated industrial sector via scientifically/mathematically sophisticated methods was the future. That was one argument against anti-trust laws: that breaking up large companies would just decrease efficiency. The only thing the communists really disagreed with was what should be done with the profits of these trusts.


> biggest problem libertarians have with communism

is ignorance of communism et al actually means. At least for the masses of "don't tax me bro!" tea baggers and other modern bandwagon, pseudo libertarians.


Actually it is the use of force and coercion that we have a problem with.

If you want to go and found a company as a worker owned co-op, that is your right. If you want to force me into it, that would be wrong.


Valve is a single data point.

A single data point that has several massively profitable ventures feeding its bottom line.

I'd be wary of drawing sweeping conclusions about political economy from a single example.


I think this is a key point. While the Steam cash cow is generating enormous profits, they have a lot of freedom to experiment. They have an almost complete monopoly of the digital game download world. It's to some extent reminiscent of the freedom that Google had early on when they were absolutely dominating search, and they weren't trying too hard to attack other fields.

It remains to be seen if that's the only reason why their current structure is working, but that's at the very least a valid hypothesis.


Just about any system works in times of great plenty. The real test of Valve's model will be when things aren't so good.


Gore and Morningstar farms are two more data points in very different markets.


Something about this article reminds me of a post-scarcity future star trek paradise where everyone works on whatever gives them joy. At Valve, your basic needs are taken care of (a paycheck, healthcare, etc), and you produce the things you love to produce. If this works on the level of a firm, would it work on the level of a whole society?


> If this works on the level of a firm, would it work on the level of a whole society?

Well, while Valve employees work the the things they want and like, they have been carefully selected to want and like the same thing as Valve. "Society" doesn't have a very hard hiring process designed to only allow high-quality individuals in, at the expense of a high rate of false negatives, and if a society had the same access to "fire" it's members, even Iran would think it was a bit over the top.


As it stands right now, society doesn't value certain kinds of output from people. In a scenario where scarcity is no longer a problem for basic human needs, the idea of having valuable kinds of output isn't so important any more. Valve still exists in a world where it needs to accumulate money in order to keep operating and producing software, so it does have a need to filter its hires based on certain metrics.


How about the Star Trek: TNG episode with the game that directly accessed the pleasure centers of the brain, and everyone quickly became addicted to throwing little virtual chips into little virtual holes...except for Wil Wheaton!


Don't forget his love interest, played by Ashley Judd, who also resisted the game: http://en.memory-alpha.org/wiki/The_Game_%28episode%29


I bet that episode is ripe with animated gif fodder. They always got that weird self-satisfied look on their face when they were playing.


I hope someone creates a version of that game for Google Glass. Or a Glass parody video acting out similar poses.


This is essentially the liberal wet dream. The main philosophical difference between classical liberals (libertarians) and today's liberals is largely a question of practicality. This is also why libertarians tend to see themselves as lone heroic dreamers while liberals tend to see themselves as working in the dirt to make the garden grow.


If this works on the level of a firm, would it work on the level of a whole society?

If you can find enough people that take joy in working in your post-scarcity sewers, yes.


Alternatively find some people who take joy in building robots to deal with the sewers and other joyless tasks. (Or some people who take joy in building robots that build robots that deal with such things.)


You missed another interesting option. Have robots (or pharmacology) which can reversibly alter a person's joy module(s) so as to take joy in dealing with sewers (or making/managing robots to do the same), and people who are fine with having the robots teak their joy-production neurons for a suitable duration of time.


I am sure this is everyone's ideal job, but I do think that there are many people that simply just _don't want to work_. In the case of Valve (and software development), I think creative juices tend to flow more easier and creates a harmonious vibe that you don't always find with other professions.


There is a serious flaw with this model. The flaw is that no one wants to fix anything because fixing things is boring and usually results in lower peer reviews than building a new feature or game.

Anyone who has daily exposure to Valve's infrastructure will notice the flaws.

- Credit card breach http://www.theregister.co.uk/2011/11/13/steam_confirms_credi... ending up in foolish security measures like encrypting your password with RSA in javascript on top of SSL.

- Power outage of a single datacenter leads to Steam going down. http://kotaku.com/5884430/power-outage-knocked-out-valve-ste...

- Weekly unplanned outages of Steam Community and the Valve master server.

- Crashes and game breaking updates in nearly every TF2 patch. http://www.mail-archive.com/hlds@list.valvesoftware.com/inde...

This interview from Gabe shows that he knows there is a problem with this, but he doesn't realize how bad it is:

Newell: A lot of times people will want to complain. The first time somebody complains, you say, “Okay, fix it.” You just say, “I don’t know what you expect to happen now, but you’ve just given yourself a job.”

Fries: Does that train them to complain less or to fix things more?

Newell: If you hired the right person, it trains them to fix stuff. If you hired the wrong person, they’ll say, “Oh, this is mean.”


> no one wants to fix anything

Is patently false. I like to fix things. I get way more satisfaction from cleaning up a mess, improving X, etc. Than I do from creating iteration 1 of something that may or may not be used.

I am not the only person in existence with this quality.


"This interview from Gabe shows that he knows there is a problem with this, but he doesn't realize how bad it is"

That is an incredibly arrogant statement.


Maybe there could be a rotation of the sacrificial fixer. Every month, one dev is given perks (private office, free food, a really big chair, ...) while they handle the back-log. They work under the guidance of the last fixer, to make the transition smooth.


When I toured Valve last summer, there were quite a lot of large empty offices. Everyone's desks were on wheels, and the workstations were set up to make them extremely easy to move. Naturally, with no management structure, anyone was free to take up an office, but hardly anyone wanted to.

The receptionist who was giving the tour, indicated that Valve had constructed more offices than people actually wanted to use, so they were considering tearing many of them down. Most employees preferred working in more open areas in the midst of the rest of their "cabal".

As far as other perks go, Valve employees already have a nice selection of free snacks, drinks, in-office massage sessions, etc... I'm sure there are ways to incentivize Valve employees to do less appealing work, and I wouldn't be surprised at all to hear that they already do this on an ad hoc basis, but they would have to go a little bit beyond your traditional office perks.


i'll bet something as simple as putting up a list of important if not very glamorous tasks somewhere prominently and continuously visible (e.g. large whiteboard on one wall), and then inviting people to choose a task and cross it off the list, would be motivating enough. note that the kind of self-motivated employee valve hires for usually wants to both make better things and make things better.


Can you explain what you feel is foolish about encrypting passwords on top of SSL?


If your SSL is compromised, the attacker can insert javascript to send the unencrypted password somewhere else.

That is why security experts like tpacek have repeatedly said js encryption schemes aren't secure.


Why do you say an SSL compromise necessarily means the attacker can manipulate the connection? And what about a promise between the SSL endpoint and the database?


"The flaw is that no one wants to fix anything because fixing things is boring"

I've always thought this was the funniest thing as I relish digging into a piece of open source software that I love to use and fixing little flaws. Just this morning I wrote a patch for guake the drop down terminal that allows the width to be resized on the fly as it was broken on Unity in Ubuntu 12.04. The sense of accomplishment at taking a piece of broken software and making it "right" is phenomenal for me. I guess different strokes for different folks.


Thanks for fixing that bug! It was really irritating. I guess I should try out guake again.


When your job is on the line due to stack ranking your opinion may change.


From what I've read Valve's use of stack ranking is for compensation bonuses only, not about who to fire.


That's what they say, but what do you think really happens when you are consistently rated lower than your peers?


You get paid less than you deserve but still more than you would get in most other less profitable firms? No one wants to fire the guy who takes out the garbage every day, literally or metaphorically.


When times are tough the guy who does the most menial tasks gets fired first and replaced with someone cheaper.

The thing about being the guy who fixes software, is that no one will hear about it if you are doing a good job at it.


I don't buy it. That's not what I've seen.

If you're fixing software, you're either fixing your own bugs or someone else's bugs. There's bug tracking software; these metrics are visible. If you're fixing your own code, you're also adding functionality (presumably you wrote the code you're fixing). If you're helping to fix other people's code, they will notice. People talk.

I've personally seen almost an entire team disbanded, the work outsourced. The people kept? The guys who fixed things, who made sure it shipped at the end.


You can see bug fixes on a tracker sure. But how many lines of code is that going to be compared to a new feature or game?


Awesome read. It's hard to find accessible well-reasoned anti-capitalist literature outside of fringe radical contexts. Although, as he himself admits, holding up Valve as the ideal model for post-capitalist society seems like a bridge too far.


It's also hard to find well-reasoned anti-mathematics literature or well-reasoned anti-genetics literature or well-reasoned anti-vaccination literature or well-reasoned anti-industry literature outside of the radical fringe. For the same reasons, probably - there just aren't much of persuasive arguments that weren't thoroughly disproved by both theory and practice.

Valve is an interesting experiment in self-organisation, but I don't see how it's in any way anti-capitalist, unless one thinks capitalism is identical to Goldman Sachs. It is not. Capitalism is about freedom to pursue every option and to succeed and fail independently. It is completely wrong to assume capitalism mandates some kind of specific company structure - on the contrary, the whole point of it is to organize the economy and the society in a way that you can try any organisation and structure (or no structure at all, if you prefer) and let the evolutionary processes to lead the ones that are more successful and adaptive to survive, while wasteful and ineffective perish (yes, I know in reality it's not ideal at all, other factors often intervene and screw up the picture, but that is the basic idea). If it turns out Valve model is superior and brings immense success, people start to follow it and there will be hundreds and thousands of companies built like that. If it turns out it's not as great as it looks, Valve followers will fail and nobody would do it anymore. In both cases, it is entirely what capitalism is about!


Does this flat structure apply to the entire company? Who cleans the floors, does QA, decides when to release a product, does the accounting, waters the plants, answers support calls? Can all of these people move their desks or is it only the developers that are free to move within development to development related tasks? If the accountant decided that they wanted to do art design, would they do so and then likely be evaluated by their peers and possibly fired(assuming that they are not a talented art designer in addition to being good with the books)?


I suggest reading this previous blog post: http://blogs.valvesoftware.com/abrash/valve-how-i-got-here-w...


Valve is a "hit business" meaning that they need to produce hits to be profitable. In the short run they have been successful, but over the long run, they may run in to problems and if it does, this business model will look foolish (rightly or wrongly).

The reason, I and others on HN don't like BigCo, is because we feel stifled, but that doesn't mean that it isn't the most efficient business model.


Valve has moved quite far beyond "hit business". Consider how much commerce happens on Steam, and what their cut of that action is. They're really more of a distribution service with a sideline in game development to keep honest.

Also, re: BigCo--that also doesn't mean that it is the most efficient business model.


Sorry, I was ignorent about their distribution service. I completely forgot they did that.

Thanks for the correction.


No worries! :)


Valve is by no means in the "hit business" anymore. They moved on and provide service and content. From the stereotypical hat in games to deliver content from other creators. Not just other developers but players that create content. Parden my language, but srsly, in the long run? What the heck? Steam is now running since ~2004 while being a provider since 2005/6? That are six years of being successful and there is no sign of steam dying anytime soon. "In the long run", like what, 20 years?


This article was absolutely fascinating, in particular because it compared traditional worker co-ops with Valve's unusual horizontal-style organization. I'm a former worker-owner at a tech startup (we were organized exactly as he described traditional co-ops, basically we all owned the firm but were hierarchical) and based on my experiences there I think the only way a truly spontaneous structure such as Valve' could work is in the high-tech or "professional" sector; I think the average person (especially those lacking college education) has a very difficult time deciding how to best be productive. I'm not saying they couldn't co-own a business - on the contrary, I think it is the way many businesses should be run! But a strict managerial structure is essential in organization differently-minded individuals (in my experience).


Corporations reduce the transaction costs between many actors, allowing large projects to be done with greater economic efficiency. This is also true for Valve. For a company their size, my guess is that they've found a much better way of collaborating efficiently than hierarchies.


Reduce? The larger the corporation, the larger the overhead introduced, and less efficient it becomes for intellectual work. Your thinking only applies to manufacturing.


Well, note that the size of Think Tanks is a lot smaller than the auto companies. However, there are still some things to be gained from the pool of talent at Valve. Gearing up a new team doesn't involve hiring. There's no paperwork. There's a lot of communicating (a good thing) and moving some desks.


"Yanis Varoufakis is an academic economist, an author, and a prominent contributor to the debates on the recent economic crises in Europe and the United States. Born in Athens, 1961, he moved to England to read Mathematics and Statistics and holds a PhD in Economics from the University of Essex. He is currently Professor of Economic Theory at the University of Athens and Visiting Professor at the Lyndon B. Johnson Graduate School of Public Affairs at the University of Texas at Austin. His previous academic appointments include the Universities of Essex, East Anglia, Cambridge, Sydney and Glasgow. His books include:"

Let me boil this down a bit.

Heads in the clouds. So using Valve, with it's particular product (games) and it's type of employee (young) we are going to construct an argument that ends in:

"and it so happens that it constitutes the reason why I am personally excited to be part of Valve: The current system of corporate governance is bunk. Capitalist corporations are on the way to certain extinction. Replete with hierarchies that are exceedingly wasteful of human talent and energies, intertwined with toxic finance, co-dependent with political structures that are losing democratic legitimacy fast, a form of post-capitalist, decentralised corporation will, sooner or later, emerge."

One wonders if people who write things such as Yanis, well, if they've every done anything outside the academic world and pure theory relying on what appears to be on the surface well written arguments that would probably go over the head of Sam Walton or Warren Buffett.


valve's people aren't very young, mostly they hire senior people. the narbacular drop kids were an exception.


Indeed. I know several young developers and designers whom any games company worth its salt would hire on the spot (and actually, they're all employed by such) who have "come back and talk to us in two years time" memos from Valve. They don't just want great people, they want great people who can self-direct.


This reminds me of Iain M Bank's Player of Games in which the hero finds a winning strategy in an elaborate StarCraft style board game by decentralising the production and strength of his ranks, so that his territory was a sprawl of self sufficient mini-empires. I think Banks meant it as an allegory for Islamic/Persian/Asian empires being eventually swept aside by Western civilisation, but it seems relevant here.

However one factor I generally find missing in a lot of economic analyses is that no matter how 'feudal' or 'hierarchical' an organisation may be, workers choose whether to work there or not. The classic characteristic of soviet states was that you couldn't leave even if you wanted to...


Someone that is somehow renegade is Pierre-Joseph Proudhon. And it's quite possible that his thought is way better fitted here.

Of course, he's another non-authority thinker but with a quite different perspective and way to think the world, the society and men.

Take some of your time at gutenberg.org: - What is Property? by P.-J. Proudhon http://www.gutenberg.org/ebooks/360 - System of Economical Contradictions; or, the Philosophy of Misery by P.-J. Proudhon http://www.gutenberg.org/ebooks/444


Valve's organizational structure has always been of great interest to me from an idealistic point of view. There have been a lot of articles written recently that cover nearly all aspects of this interest-based economy for developers, artists and so on. Other companies, such as Github, seem to have adapted similar models.

Regardless of the article though, one detail always seems to slip past me. Who answers the phones?


The person who was called?


If anyone is interested in hearing more from Ronald Coase (and you should be), he was on the excellent Econtalk Podcast earlier in the year: http://www.econtalk.org/archives/_featuring/ronald_coase/

The topic is "Coase on Externalities, the Firm, and the State of Economics", well worth the time if this topic interests you.


I just did a quick skim of the article, and call me naive if you want, but in the theory of spontaneous order why is Valve necessary? Obviously, the time and labour of Valve's employees is worth more than what Valve is paying them, otherwise Valve would not be turning a profit. I'm assuming that employees really have 100% work time for their projects as you claim. Then in essence Valve is providing facilities and connections (to other smart employees) in return for the lion's share of the labour profit. Is that really worth it for the employees? Once they have the connections wouldn't it benefit them more to split off their own companies, in a co-op horizontally structured company?

Don't get me wrong, I think traditional corporate structures are often abominations and I think Valve is a great company. It just seems to me that if you take the spontaneous order philosophy to the extreme than corporations in general just become unnecessary overhead.


Economic profit is a function of land, labor and capital. Taking out profit from labor still leaves the other two; You still need shareholders to finance the initial purchase of land and capital, and the shareholders still demands compensation for risk.

The corporation becomes unnecessary overhead if the company is owned by the employees, but then you have issues of what happens to the equity when an employee retires, and thus having an employee become a shareholder.

I think the word I'm looking for is a collective, and looking into the past, despite Valve I still don't think it works. The very idea has caused the deaths of millions of people implementing it.

I think Valve is a great idea; the idea being, the employees of the company get to decide everything in the company besides a few hard rules. Example:

Rule 1: shareholders get 50% of the profit.

Employees can decide how much is spent on profit-sharing schemes and salary versus how much to save for the company, to invest in new equipment, whilst following the rule. Employees dictate all aspects of business strategy.

Failure to abide the rule simply ends in the shareholders dissolving the corporation.

Ideologically it doesn't differ much from corporations you see today. The CEO, an employee, typically has free rein over the company until he does something really stupid, in which case he gets fired by the board. This "new idea" is simply making a hierarchical organisation a horizontal one, i.e. all employees are now co-CEO's. You still have leaders but they are distinguished neither by status or wealth.


I completely agree a horizontal structure is a better working environment for employees. I don't know how the compensation packages at Valve works, but I'm betting the shareholders (and executives at the top who are given shares) are getting the vast majority of the profits, just like most corporations. I bet a lot of the most talented employees at Valve are vastly overproducing the amount they are paid, and could easily gain much more personal equity even after trading some for enough land and capital to start a company themselves. Not saying everyone should start a company, I haven't yet so I guess I'm being a bit hypocritical here.

Good reply. Like I said I figured I was being a bit naive in my post.


Sounds like a worker's cooperative to me. Many examples of such exist, but they are shackled in the United States because they do not enjoy the legal protections that corporations do.


It's a long comment so I'm breaking to a second reply.

A horizontal company to a hierarchical one is akin to a group of amoeba compared to an animal.

As a group of amoeba grow, it simply splits into multiple groups. An animal will grow very big, much bigger than any single group of amoeba, but it is vulnerable. Stab it and it dies; once the central nervous system is dead, all other cells will die off eventually. You can't stab a group of amoeba. To kill amoeba you have to kill every single individual one.

Large companies today are much more likely to make mergers and acquisitions than to make divestments. Divestments are only made when a group within the company appears to be unlike any other. I think horizontal companies are more likely than large companies to make divestments. The only other horizontal company I know of, Semco, is an example that is evidence supporting this; it is a company controlled only by employees. At times it has split off its services as separate companies, with the separated companies remaining horizontal, and most of them are very profitable.

http://en.wikipedia.org/wiki/Ricardo_Semler#Semco_1990.E2.80...


W. L. Gore is a fairly well known example of a more horizontally structured company.

Some conglomerates, like Berkshire Hathaway, also eschew structure, at least to some extent.


I am not an Economist, but what you're discussing reminds me of John Lewis, a British department store & supermarket chain. It's owned by its employees, through a trust, and they all get a share of the profits as well as some say in the running of the company.


Valve provides the connections and environment of trust without which spontaneous order can't occur. It also flattens peoples risk profile (lower risk, lower reward), which is actually quite desirable for most human beings.


> Obviously, the time and labour of Valve's employees is worth more than what Valve is paying them, otherwise Valve would not be turning a profit. I'm assuming that employees really have 100% work time for their projects as you claim. Then in essence Valve is providing facilities and connections (to other smart employees) in return for the lion's share of the labour profit.

I think you're underestimating the value of those facilities and connections. That is the overhead.

Also, the time and labor is valued higher because they're employed together; that same time and labor would not be as valued outside the company.


Valve sounds like it might be the only other company I'd ever want to work for.


There are some activities that always needs to be taken care of in time. Such as customer support. I wonder how do they do those tasks. If there are no excepted level of work, some tasks might go undone.


This is the mild version of the "If there wasn't a God and a Bible and a Hell, we'd all kill and steal from each other all the time" argument.

1) People do exist who like to do customer work and want to do customer work. If they're sought out and given internal support--and it's not really that hard to convince decent engineers to provide that on their own initiative--they will handle it fine. Same goes for things like office management.

2) If some tasks go undone, it's because no one was interested in producing a finished, polished version of the product into which those tasks contribute. If no one is interested, then the product is far less likely to be worth producing.


I am one of those people. I like helping our customers. But, liking something and working on something when you feel like it is very different than having a consistent support response time of 15 minutes.


Does Valve make a promise of having a consistent support response time of 15 minutes?


That's just an example of a task which would be difficult to implement without any management.


1) Sounds like a challenge.

2) Then maybe that's a good reason not to do it.


Fascinating stuff! I am very curious as to whether Valve is scalable to either very large companies, or companies composed of mere mortals.


Does anyone else hear the sabers?

There seems to be a lot of noise about Valve these days. Some examples: the "Windows 8 is going to be a catastrophe" thing; the "we're hiring super elite Linux developers" thing; the "our engine is faster on Linux?!" thing; and now this navel-gazing self-congratulatory gobbledygook.

Sudden Clarity Clarence asks, "Is Windows 8 going to have an app store? Could that be a threat to Steam? Is all this noise just saber-rattling?"


Valve's statements about Windows 8 usability were mirrored by Blizzard.

The source engine is faster under OpenGL on Windows as well, not just Linux. So It's an OpenGL vs Direct3D debate, not Windows vs Linux.

"Is all this noise just saber-rattling?"

All this noise is branding/promotion of the Valve brand, and would likely be going on even without Windows 8.

But I agree that Windows 8 will likely try to copy Apple's iTunes success and that Valve needs diversify.

I'm hoping they do a console running Linux at some point.


Thanks for not calling my comment "crap".

"Valve's statements about Windows 8 usability were mirrored by Blizzard."

Might Blizzard not feel similarly threatened by Windows 8?

"So It's an OpenGL vs Direct3D debate"

I slightly misspoke, but the example still holds: an attack on Direct3D is an attack on Microsoft. (Do I need to back that up? Could anyone conceivably disagree?)

"All this noise is branding/promotion of the Valve brand, and would likely be going on even without Windows 8."

Good point, and I agree. But what would it be like if Windows wasn't pushing an app store? Would they still be calling their main platform a "catastrophe"? (Maybe, maybe not.)

So I guess I should have asked, "How much does Valve fear a Windows app store?" And also, "Doesn't it kinda seem like Valve is attacking Microsoft?"

Sometimes it seems like "Hooray for Valve/Steam/Gabe" are three of the internet's favorite topics. I get the impression that the Venn diagram of people who love memes and people who love Valve probably looks like a circle.

How does Valve manage to get so much love from the internet? Sometimes I think they must have a small team dedicated to feeding social news sites. For example, "Hey look guys, a new Team Fortress comic!" Or, "Here's some fancy talk about how enlightened we are! P.S.: Give us your money."

Aren't there better-loved companies that receive much less internet-love?

According to the internet, Valve is not just a company, it's a hero. And Valve clearly feeds this idea. For example, this very article. Valve doesn't just want your money, it wants your love.

Is this dangerous? Is this effective? How much does internet-love translate to real dollars?


Keep the meme crap out of HN.


Whales are big and fleas are small. The optimum size of an organi(sm|zaton) depends on environment and niche.


tl;dr synopsis of this management style:

* Hire great people,

* Give them goals and the resources they need, and

* Get the hell out of their way.


I really, really hope Valve succeeds and gets this vision through. It sounds like a great company.

His insight about corporations being Soviet in nature is spot-on. Corporatism is neither capitalism nor socialism, but a hybrid system to give a well-connected elite (~0.5%) the best of both systems and the other 99.5% the worst of both worlds. Look at air travel; that's about as Soviet an experience as one gets, but the pricing is aggressively and mean-spiritedly capitalistic. Or consider suburbia as a microcosm. The rich live in places like the Hamptons and have both rural and urban amenities, while the poor live in depressed, polluted exurbs that combine the worst of city and country life.

What Valve sounds like to me is a post-scarcity capitalistic model where there's still inequality of results (as, IMO, there should be) but there isn't pain or poverty.

In our current world where there is a lot of scarcity (even though it's outmoded and artificial in the US) people work a certain way, and give up too much power, because the alternative is risk of economic misery. In a post-scarcity world with more of a safety net, people probably would "wheel their desks" to other projects, companies, and opportunities (or split their time among more than one, rather than lingering in this undiversified full-time thing) more freely. That's what we're starting to see in technology, as the demand for programmers increases and the stigma against changing jobs frequently (assuming there's upward progress and learning) goes away.


Valve's move to Linux was inspiring. He referenced Marx, I'm now starting to wonder about their wisdom in general. Why can't they just make a great game platform and stay out of politics.


You're an idiot if you take a completely reasonable assessment of various economic theories as "politics". This article is an exploration of where Valve's corporate structure and soul fit into the spectrum of ways to interact in a market.

Don't let the names of leading thinkers in a field scare you off because of some bullshit brainwashing.


Why? Marx is one of the more influential figures in modern history.

In this case, the reference to Marx was basically saying that corporations pay you a fixed rate for your time, but must extract a larger value than what they pay you. Reading this won't make you start hymning the Soviet national anthem.


Economics used to be called political economics. Why Valve can't stay out of politics is because a lot of things worth doing are political in nature or related to politics. Politics is not a different sphere from everyday life.


In the context, Marx isn't about politics. Marx is about economics.

Talking about economics without referring Marx at all seems disingenuous.

The article is a post from the Valve's economies blog.


I don't see any place in the article that places a value judgement on Marx's beliefs or positions, just sections that outline what questions he asked and what his ideas were. Can you point out what I am missing?




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: