Ted Nyman from GitHub recently gave a talk at Mozilla about GitHub's "manager free" work environment. Here is a video with slides: https://air.mozilla.org/scaling-happiness/

I was excited to read about how an "org-less" org could work. Then: "We've grown from three to thirteen employees in about a year, to shrink back to four in about two years." And: "About a year ago, we've filed for the French equivalent of Chapter 11 of Bankruptcy Code."

It's hard for me to have much faith in the author's argument.

I read it as the description of a first experiment. Note they still can improve and come back in the economic race. However, this is interesting to see some people have courage to try those things, whatever the risk of failure it represents.

I have a hard time equating a lack of all organization with progress. It's true that many organizations have needless layers of management (and the associated interference), but what the article describes could most easily be replaced with four (or thirteen) individual sole-proprietorships.

And does the company's views on organization extend to the agile training they provide? Agile isn't disorganized or even unorganized, it just recognizes that you can't know everything/enough when you start. If they can't organize themselves well enough to avoid bankruptcy, can they organize a software project?

I think what they've done is an interesting social project, but by his own description, it seems like they've proved it doesn't work.

There's a somewhat different risk profile in these models than with a collection of a dozen sole proprietorships, though. Some ways better, some ways worse. I'm interested in a model along those lines myself, if I can find the right collection of people, as a way of trading away some upside gains (if I have a huge hit, we all benefit instead of just me) for some downside protection (as long as some subset of us is, on average, doing ok, we all can eat & pay rent).

My initial reaction was that this couldn't possibly work at company level. After some thought, I could see how it could work if you have just enough diversity in personalities and temperaments. For example, it sounds like they need at least a handful of people who are willing to do those "critical, not-so-gratifying tasks may remain undone and hurt us in the long-term."

"It's been a year since we've filed for bankruptcy and we're still there." - I work in a company that strives to follow such a model, and it honestly struggles to achieve the kind of efficiency most companies take for granted. I dearly hope it doesn't lead to that end, but from what I can see it's a distinct possibility.

I suspect that the highest performance levels come from a group that self organizes around the vision of a leader.

That's the most common pattern indeed. Note they tried to find a vision by themselves without an "official" leader. They were unfortunatly unable to find one apparently.

However, I believe a kind of implicit vision has emerged which is simply to make sustainable such a place. Put another way, people who are attracked by this implicit vision (such as the author of the article) are still with them, and the others have probably left because this was not a vision they were inspired by.

One can outsource leadership in a variety of ways. A highly cohesive and capable group can do consulting projects and self-organize around the project, for example.

In my experience, the biggest hurdles to a truly self-organizing organization (at every level) are people and trust.

If you don't have really excellent, self-motivated people, you are going to have a lot of trouble with self-organization simply because some people need to be told what to do. People also have to be 100% bought in to the concept and don't try to create themselves management and authority roles. It really does take the right people.

Trust is probably the biggest problem. It takes a lot of trust from the people who start the business to let it run itself. It is probably easier at the start of a business than when it's big. Everybody has to have a high level of trust of everyone else, and when that trust is broken the organization probably needs to be willing to fire the person who took advantage of the system.

For many people, the hardest part would be to maintain that kind of system with ruthlessness enough to fire people who don't work well in the system. Knowing that it's not going to be for everybody and practicing that is hard.

Not only trust, but honesty. It's a lot harder (I would wager almost impossible!) to operate in a flat organization without everyone being willing to both be honest and receive honest criticism.

The company I work for is about 30 folks strong and there are, for all intents and purposes, only two levels of hierarchy. It could probably sustain being completely flat, but I'd fear that it'd fall apart relatively swiftly due to people being unwilling to both bring others to task and be brought to task themselves.

Plus, there's something to be said for decision by committee severely obfuscating progress if consensus can't be easily reached.

I think some service industry employers try to get around trust and employee quality by limiting the capabilities of the employees, and it's a nightmare for everyone. Often its done by software-- "sorry, the computer won't let me." It feels awful to be a part of; the lack of trust is hard not to take personally.

This seems to be a serious option, or at least becoming a viable experiment for some groups of developers. Fred George and Developer Anarchy is the other proponent of self-organised teams. My experience has been it is very very tough but rewarding.

Oddly I think there is an intellectual justification behind this movement - if we assume code literacy is as important now as real literacy was post 1451, then it's a rare group of people who write for a living that do not have few or no managers (who is the manager of authors at a publishing house)

I love reading about people trying new things like this.

As an Agile/XP coach, what I find is that I need to be very careful that I distinguish between things that work and things that feel like they ought to work. If you're not a pragmatist, always trying new things but always looking at results with a cold eye, you're truly lost in this business.

Good article, but the jury is still out here. Right now this is more of a sales pitch than a story. In about three years, I'd love to come back to these guys and see how it's going.

One of the problems with any organization is that people like doing what they like to do, not necessarily the things that need to be done. There are things we are really good at -- programming, artwork, group encounters, whatever -- and we tend to gravitate towards situations that let us do these things. The immediate danger of an organization without a tight feedback loop is that people drift into doing their own things and other important stuff doesn't get done.

But that's just the theory. The important thing is how it works in this situation, for these guys. Hopefully there will be an update a few years down the road.