The end of bosses | Hacker News Reader

The end of bosses

fullstart.com

58 pointsmjshampine33 comments

I've always thought of most companies like machines. Some guy who is really passionate comes along and builds it, but eventually he/she leaves or loses interest and the machine grinds away until it eventually collapses. Some machines are very well built and last longer than others, but without maintenance they'll all eventually collapse.

Basically most companies, like machines, are just an extension of someone's will. Human will is the only thing that can stave off the inevitable decay, and when no cares any longer, it's over.

The only way to delay this is to hire people who are competent and passionate, and give them enough autonomy so that instead of functioning as part of a machine that will break down, they can exert their will to fight entropy and keep it going.

I think this point of view is biased from the perspective of an entrepeneur. You seem to be over-valuing the contribution of the founders, to the point where you're implying that everything that happens after the founder leaves the picture is all his/her doing.

Would you make a similar analogy of nation-states? The Catholic Church?

I'm sorry, I should have made my point more clear.

The gist of what I'm saying is that if you treat people like machines, they act like machines. If your organization is setup as a strict hierarchy with zero autonomy, it will function as a giant machine, only kept going by the person/people controlling it.

They alternative to this is to maximize autonomy for all your employees. Allow your people to be people. There is an innate human desire to create, to organize, to make an impact, to move from chaos to order--to fight entropy. Harnessing that desire is, in my opinion, the best way to run a company.

I'll give you an example. A McDonald's employee isn't happy with his job, but one day he decides to make the best of it and just do the best job he can. He walks into work intent on making the best damn hamburger he can make. But then he realizes...he can't do that. His job isn't to make the best hamburger he can make, it's to make the exact same hamburger made by every other employee at every other McDonald's. In this case what we've done is to take a human and do our best to turn him into a robot.

Blessed are the meek for they shall inherit the earth.

Relevant: http://www.smbc-comics.com/?id=2213

I liked/agreed with everything in this article except the Sunday test. I get that people need to be compatible in order to work together. I don't get why people need to want to hang out together on Sundays. It all but guarantees that people who are otherwise qualified and compatible with the team Mon-Sat will not be hired. Nothing terribly wrong with that, but why not just hire your techie frat brothers or sorority sisters?

Yep. this part of the piece only works if the company is a cooperative, or you're just back to "get them to drink the kool-aid and sell their lives to the company" psychological slavery.

I've seen this idea work well in very young companies (2 to 10 employees). It can boost morale at that size, but it would likely feel forced beyond a handful of employees.

Problem is, there are people who get off on authority. Some of them will eventually slip through the cracks and start to infect your company - unless you diligently keep the culture alive.

I think that's part of the whole idea. There are lots of talks at the moment about hiring the right people for the culture of your company. And obviously I believe that if you are creating a hierarchy-less culture then you should go for those people who show those traits.

I personally believe the whole idea of non-hierarchical society better such as in Western Universities where students call professors, lecturers by their first name. Whereas it is toally opposite in Japan where you have to address them by their titles (Dr or Professor so and so).

I like to think of companies like a computer program. A startup isnt written in anything it is completely implemented in hardware. As the company grows it is necessary to add these different layers of abstraction until you end up with the CEO and board of a huge company in the application layer.

I don't think you can maintain a flat structure forever. Even people within valve have written about its "flatness". I've at least read that there is just a different set of internal issues.

I was trying to think of a way to say this, but I think you summed it up fairly nicely.

I've read too many articles on HN on how managers shouldn't exist. IIRC, at Google it had become something of a general complaint that there was a lack of managers. That is, there might be 100 employees to one manager, and you'd only see this guy once a year when it came to be review cycle.

Managers actually do something, and it's not just to micro-manage, set arbitrary deadlines, etc. It's quite possible for teams to organize towards a task without one. But managers do a lot of other things besides yell at people.

I think the common view is too heavily informed by bad bosses. It's true bad management is worse than none, but that doesn't mean none is ideal.

> Mature startups like Valve and Github and smaller up-and-comers like Treehouse and Medium have done away with managers who boss people around and expect servile compliance.

Valve, the post child for this model within this community, was founded in 1996 [1], so this is not necessarily a new idea or something limited to startups.

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

I have a huge section in my talks all about how Valve makes this work. They're damned impressive.

Can I admit here that I hate "the Sunday test"? I wouldn't ever want to use that on somebody, nor would I want it applied to me.

Bingo. These types of litmus tests always get strained the more an organization grows, and eventually organizations outgrow them.

And as frustrating as these things are from the to-be-hired side, I have dozens of colleagues who strongly complain about these practices in their own companies. They always use the same words -- "false positives", meaning that someone was considered an ideal candidate but they failed the do-i-want-to-play-cards-with-them test.

What surprises me is how everyone claims to be data-driven, yet when faced with the aspect of hiring, the same choices are repeated over and over, which (shockingly) leads to the same results. In a field where creativity is important, we as an industry are forcing people to fit into a certain persona, all with the goal of being predictable, interchangeable parts. Great minds may think alike, but I've never heard "like minds think great."

As with other trends, this too shall pass.

This is probably one of the hardest things for many companies to grasp today. One of the things I do for a living is give full-day presentations to companies looking at Agile from a "strategic" position. In short, companies often do agile development, but they still lay down hard deadlines, micro-manage projects, and provide command-and-control rather than support-and-deliver. This problem is the elephant in the room of agile: what do we do with managers when we have self-directed teams?

If you look at the Feudal era (this is a gross oversimplification and varies wildly from from country to country), we had a king who gave land/authority to dukes, who gave land/authority to barons, who gave land/authority to knights, who had serfs work their fiefs. In return, the serfs had a place to live and gave food and goods to the knights, who pledged service to their liege, who pledged service to theirs, up to the king. The king also taxed every level of this and made orders of magnitude more money than the serfs.

There were a couple of interesting side-effects of this. First, while you were technically loyal to your king, he didn't have much day-to-day impact and your loyalty was to those immediately surrounding you. Communication throughout the realm was slow or non-existent, further isolating communities. Politics were local, not national, and these made made medieval intrigues all the more fascinating. Second, there were no price signals inside of a feudal society. As a result, you couldn't have a functioning economy in the modern sense.

And then about 300 years ago we had more and more international trade and many lords discovered that by displacing their peasants/serfs they could do things such as raise sheep and sell the wool for a far greater profit than having people harvesting food[1]. So you wound up with a bunch of free men traveling the land, selling their services where they could. In other words, you had one of the foundations to a modern economy: a mobile labor force.

Today, that feudal model still exists: inside of many large corporations. It's a mess of internal politics. People have loyalty to their teams and maybe their boss, but not so much to Steve Ballmer. There are few, if any, internal price signals. Internally. If you commit "treason" by suggesting Ballmer should be fired, you might get laid off, but not executed. You have hiring and firing instead of births and deaths to maintain the work force. In may respects, the modern corporation resembles the old feudal system, but sped up many times. And the king/CEO still earns orders of magnitude more money than the workers/serfs, completely out of proportion to actual value provided.

Agile itself had its roots back in the 50s, but as a modern movement in programming, agile is still just over a decade old. We don't know what's happening to the traditional role of management, but the feudal system couldn't compete with capitalism. I don't think that traditional corporations can, either.

1. http://en.wikipedia.org/wiki/History_of_capitalism#Enclosure...

Edit: A couple of tiny clarifications.

So, playing the devil's advocate for a bit here, are the agile companies without managers delivering consistently better results than companies with managers?

Companies grasp one thing quite easily and it's their bottom line. Maybe, agile is not delivering better enough to have an impact on the bottom line and all this discussion is, well, moot.

That's a long and complicated story. The short answer is "we don't know". One company that hired me complained that every agile trainer they've spoken with has done a terrible job of explaining the role of management and I have to explain that's because it's one area that agile really doesn't have a good grasp of. It's a time of tremendous experimentation and we don't yet know how this is going to play out.

That being said, in my training I give several examples of Agile companies that either have no management or very flat hierarchies, such as 37signals, github, Booking.com and Valve. There are many reasons why they're successful and it would be silly to simply say it's because they're jettisoning management. However, we know that large companies tend to be significantly worse at innovation (Steve Jobs is an anomaly) and they also often focus on risk avoidance rather than reward.

The primary strengths of many large companies is name recognition, financial reserves, and mature products. That's very hard for smaller companies to compete against, particularly given the old saw of "nobody got fired for buying IBM/Microsoft/$insert_large_corporation_here".

when shit hits the fan and you need a synchronized effort by many to reach a common goal, the hierarchical organization is the best. proven over thousands of years of human development.

past a certain size these hierarchies do not work, hence you break them down and then let them compete against each other.

look outside of software and see how humans have organised, what has worked and what not. the modern army is a great test bed for any form of team effort, they try out everything in large, controlled samples. one of the few organisations where management is a distinct set of skills you need to prove to have (officers, NCO, etc.).

strategy comes from the top, defines the goals, objectives. smaller teams then take those goals and have freedom to execute against them - but only in the how, not the when.

orchestration, synchronization of those teams is management, pure and simple.

leading by objectives. then measuring the outcomes and defining what to do better the next time, executed in AARs (after action reviews). making sure the right things are done at the right time. removing obstacles, managing resources (hiring, etc.) - all of that is management.

capitalist societies still have management as well, it's called government, selected representatives to manage a country. difference to feudalism is rule of law and a process to elect people in the ruling structure.

I'm happy you used the military as an example because I'm talking about markets. A military is precisely the place where a command-and-control structure is paramount. They do not have, and should not have "price signals" directing what they do. Their goals and motivations are entirely distinct from a profit-driven corporation. I think we should all feel grateful for that :)

Edit: as an aside, I did have a military analogy in my talk at one point and I used guerrilla warfare as the classic example of where a smaller force, with fewer resources, can take on a larger, well-organized force. In some ways guerrilla warfare can resemble Agile, but though I loved the analogy, I cut it from my training because it felt like too much of a stretch.

I think these are good arguments, but that said, the market mechanisms at most companies today are relatively weak.

Just because the organization exists within a market, doesn't mean that the organization itself acts like a market. I think the reality of companies today is much more like benevolent dictatorships than anything else.

Completely agreed. This is one of the main points I made in the original reply (https://news.ycombinator.com/item?id=6547517).

Are there video recordings of some of your talks?

Also, what is your opinion on Semco's system?

This is a one-hour "short version" of the talk that I gave as a keynote at a conference in Germany: http://www.youtube.com/watch?v=qnUD8ed3-ok

There are a couple of issues with that talk. It wasn't structured as well as it could be and it was only an hour long, so I had to gloss over or skip many topics. It also doesn't cover the topics I'm discussing here.

Also, I'm doing reading on Semco, but don't want to comment too much until I understand them better. They are, however, really fascinating. Some things would be hard for traditional companies to adopt if the company is bound into union contracts. I am, in theory, pro-union, but they're often counter-productive.

Update: for those who are into OO programming, here's a talk I give about traits (roles) versus inheritance. It's completely unrelated to this, but I love this talk so much that I want more people to know about it :) http://www.youtube.com/watch?v=cjoWu4eq1Tw

We are a combination of managed/professional services + platform development and this is exactly how our hierarchy has evolved. We have account managers that drive a lot of timelines but we don't want account managers to actually be the boss of the engineers. We had everyone "reporting" to the CTO for awhile but recently broke things up into teams. I've got a number of people who report to me but their work is not always driven by me. In the same way, I report to the CTO but it's very rare for him to give me something to do outside of maybe a high level priority. It's much more common for me to do work directly with our CEO.

Information is shared freely around the company, hiring decisions are made relatively democratically, we hand out tasks to people who have skills, expertise, time and the desire to do them, not giving cushy work to people who play the politics game.

At first I felt like I should be doing more as a manager, like I should be managing more, but there's really not a lot for me to do. I help out my team when they need it but we hire good people so everyone seems to enjoy working autonomously and being self-driven.

> With peer management, accountability isn’t to a boss or supervisor, it’s to each other, your team, your company, your customers, and yourself.

I'm frequently working for six different account managers, plus my own initiatives, plus something for the CEO. The person who knows the least about what I'm working on is probably my boss and it works out well that way.

I'm a huge advocate of ideas like this (repeat, huge), but like many things in life, it is dramatically more complicated when you dig in.

First, hierarchies always form. They might be informal, or quasi-transient, but in the absence of explicit, implicit forms.

Second, things like this are great when the teams are small, or even relatively small. Though once the company grows they start to become impractical. And when you get to a certain size company they would be impossible. I'd love to see a multiple thousand person company in this form with some details on how it operates. I'm nearly certain we'd see "bosses" (in some sense of the word) emerge from somewhere.

Lastly, who mediates? Who makes the hard call? Who pivots the company? Who pivots the product? Who starts a new product? Who axes an existing product? Do people/teams/products/projects have budgets?

The point being, "bossless" (what a terrible term) is great when things are going well, but not so great when hard calls need to be made. If you have ever been in a death spiral for lack of money or bad-product fit, I'm sure you'll know what I'm talking about.

I'd love to see some case studies on how it worked at very large companies as well as how it worked when times were really bad.

What about informal hierarchies? Do they form in these companies? If not, why not? If they do form, do they cause similar problems to a traditional hierarchy?

Internally Valve forms temporary hierarchies to shepherd projects along and there are also anecdotal accounts from ex-Valve employees that things aren't as "egalitarian" inside as they claim and that some people with seniority have more authority than is claimed.

That being said, I've had a hard time verifying any of this stuff and it's rather a given that a company would be reluctant to air their dirty laundry.

I would think that, no matter how hard a company tries to hire people of the same mindset, temporary hierarchies would develop - not necessarily along the lines of seniority, but along the lines of who had more experience/comfort/knowledge of best practices, etc. regarding a particular domain, technology, etc.

Ideally, the hierarchy would be truly temporary and whatever knowledge/experience/insight caused one person to become the project "shepherd" would be willingly shared amongst the other folks.

I've heard that this is the case. A former Valve employee described it as similar to high school. http://www.wired.com/gamelife/2013/07/wireduk-valve-jeri-ell...

IIRC, there's a longer write-up. And a lot of the problems seemed to be partly due to her being a hardware engineer trying to make a next-gen console, embedded in a software company.

Some of the issues were specific to Valve's style. I think she had trouble recruiting staff, because Valve wanted her to hire people who could be team players within Valve, and she couldn't find a technician who could also program first-person shooters (unsurprisingly).

But more generally, she was hamstrung by Valve not having quite the hardware instincts to manage building a console with the time and resources they gave her.

This article sounds like it's talking about a form of cooperative, without calling it a cooperative, which isn't a new idea, but certainly is one that is starting to become more mainstream. For example, the UN named 2012 the "Year of the Cooperative." I'd like to believe that people are starting to see cooperatives as a way of organizing businesses outside of only coffee shops and grocery stores.

Yes! Magazine came out with a good series not too long ago on cooperatives in operation today, their varieties, their benefits. http://www.yesmagazine.org/issues/how-cooperatives-are-drivi...

Reply on news.ycombinator.com