DHH: Un-Manage Your Employees
nfib.comTo the commenters here who notice there may be an upper limit on the number of employees David's approach is valid for, keep in mind that the subtitle of the article identifies "small business" as the scope. Thus the counterexample of Google seems unfair. That said, the legal definition of "small" in the U.S. for most nonmanufacturing businesses is that the business must have less than $7 million in annual receipts (http://en.wikipedia.org/wiki/Small_business#Size_definitions). It would be interesting to know if David feels there exists an upper limit on the number of employees in order for a flat structure to work, and if so, what that number might be. Since large organizations, if they are designed well, are just a latticework of small organizations maybe it's possible there is no upper limit.
Regarding David's larger point I would add that having a boss tends to make people feel slavish and subservient to a certain extent. This creates an unhealthy work environment compared to the ideal in which all employees are truly on equal footing as regards power. One of the tragedies of working for a manager is that nervous feeling you get when you ponder how much control this one person wields over your career. A superior organizational design is one that takes care to hire people who are great at managing themselves and entrusts them with the power to do so.
Sure I agree. But on the other hand, the Joel Spolsky article I posted noted that he started having problems once he got to 17 people, which is smaller than 37signals. I was simply using Google as an extreme to illustrate my point.
> Since large organizations, if they are designed well, are just a latticework of small organizations maybe it's possible there is no upper limit.
How do you make a smaller organization though? By giving someone authority over it who can really take control of it.
> Regarding David's larger point I would add that having a boss tends to make people feel slavish and subservient to a certain extent.
Tough. That's just the way it is. You're always going to have a boss. Whether it's a manager, a C-suite executive, a board of directors, or the customer. You're never going to be completely free from someone having authority over you. It's just a matter of how they use that authority over you.
Now, I will agree that bad bosses (which are unfortunately too common) make you feel slavish. If you have a good boss (and let's give DHH credit, he sounds like he is a good boss), then they make you feel powerful. A boss's primary job is to enable you to do things either with encouragement or by taking care of the administrative details for you. Someone once said that a boss is a "secretary who can fire you."
FWIW, I've heard that Google once had no managers. The person I heard it from started around 2002, when Google was already an order of magnitude bigger than 37signals. So while it obviously doesn't work for all size companies, it seems that you can get fairly big and still do okay without managers.
My only issue with that approach is that of holiday days.
I can see so many ways that that can go wrong. To me, the reason it works seems more like peer pressure than enjoyment of work.
I love what I do, I love coding, and I also know that I love taking a week out to not care about everything I do at work. If I was in a system like that, I'd feel kinda pressured to keep up with my colleagues and not let the team down. Not by my friends and officemates, but by the knowledge that if I take a week off, that's a week they've got to handle the work I'd do. There's rarely a 'good' time to take holidays, especially in smaller teams.
It could work, it is certainly possible, but it seems an unnecessary area where pressure can occur when a team is under stress. To be honest, I think it could work better with the caveat that everyone has to take a certain minimum level of holiday.
Sick days however make perfect sense.
People take a healthy amount of vacation at 37signals.
And if we spot someone burning out because they aren't taking enough vacation or time off, we remind them - and sometimes strongly encourage them - to take some time off.
But remember, they are already working a 4 day week, so the chance of being burnt out and forcing themselves to work is lower than normal.
IIRC, they don't do that anymore. Something about the weather in Chicago. No, really.
We do four day work weeks from May - October. Five day work weeks the rest of the year. We've found that's the right balance.
Sounds good to me. I would like a workplace where it is accepted culture to work 4 days, or equivalent on average, since I'd probably rather work a few of weeks and take a long weekend.
As much as I like the idea of not having middle managers, I think it's impossible at a certain point. I mean, Google would love nothing better than to do away with middle managers altogether, and they have them. I think the best counter-point comes from Joel Spolsky: http://www.inc.com/magazine/20080901/how-hard-could-it-be-ho...
Thank you for sharing that article, Jason. :) Paraphrasing Joel Spolsky in response to DHH:
A final thought: You have to be careful when it comes to embracing the latest business idea. A single anecdote filtered through the eyes of a founder about a new cool philosophy for running a company has to be considered in the light of other evidence, such as the way thousands of other companies are set up and operate.
I sometimes wonder why DHH and Spolsky are so different. DHH speaks as if what he says is law, will work for everyone, Spolsky seems to constantly pad his advice with warnings about context.
In this case, DHH seems to have provided the appropriate hedge:
> You might be thinking, "This is crazy -- it would never work at my company." And you may be right. But I think there’s a greater chance that it would work. If you’re apprehensive, try experimenting with one team or division.
I wouldn't consider that a hedge, I would consider it lip-service to the idea of "I might be wrong here." It seems almost completely disconnected from reality.
What do you think would happen at Acme Inc., a company with more than 1000 employees, if one division suddenly demoted/relieved the manager, gave each teammate a gratuitous sick/vacation policy, and rotated frontline employees through the management roles?
I work at a company of over 300 employees, where we implement many of these kinds of ideas: we don't count sick or vacation days, our management hierarchy is extremely flat, and we have no policy about what time people come in to work. Despite that, I think this method of management-via-non-management is not something that every company can try on. It has to be built into the culture from day one. DHH is really only describing the trappings of a fun company culture, not the core of it.
You can't take the outward symbols of that culture and pin them on a company which doesn't approach employee trust and management in the same way.
What's ironic is that you're doing the same thing. You pay lip service (using phrases like "I wouldn't consider" and "It seems") to the fact that you might be wrong about DHH paying lip service to the fact that he might be wrong. And then when you actually change over to the main topic, you do the same thing. The majority of this comment explains how DHH is undeniably and completely missing the point.
If you want to criticize DHH, be my guest. I might agree with what you're saying. Just don't criticize him and then turn around and do the exact same thing as you accuse him of doing.
Not really. The implied point is that DHH has an obligation to explain not only why a company should consider his radical changes. But also, because of the magnitude of the changes he proposes, he has an obligation to explain seriously why a company shouldn't. Otherwise, it's just cheerleading.
I was calling out that he failed to do the latter, and my obligation is to explain why a company might not want to, or be able to, implement those changes.
The fact is that I'm not wrong about him glancing over an important part of his proposal. If you write an article advocating companies undergo radical shifts in their organizational structures, then as someone who is regarded as an informed commenter, you have an obligation to tell people the pitfalls of such a large change. Again, when you don't do that, it's called cheerleading.
He didn't include any discussion of the downsides, or the upsides of hierarchical organization, so I'm not wrong. I should have used a more tactful phrase than "lip-service," however: it's more loaded -- perhaps even derogatory -- than I intended.
I wasn't trying to say that you were being disrespectful. I was just saying that the way I interpreted it was logically inconsistent. It came off that you were saying "DHH isn't admitting that there might be a downside to his plan. He needs to provide a balanced argument. Oh, and by the way he's completely wrong about everything too."
Whereas after reading this post (and rereading the last one), it sounds like you might have meant to say "DHH isn't admitting that there might be a downside to his plan. He needs to provide a balanced argument. For example, here's one potential hole that I see" which is a reasonable thing to say. Remember, we humans are dumb animals who are prone to completely misinterpreting your argument. Sometimes you have to beat a dead horse to make your point. :-)
It's all good; I'm glad we came to an understanding. Thanks for the feedback and perspective.
> You can't take the outward symbols of that culture and pin them on a company which doesn't approach employee trust and management in the same way.
My first reaction is that you are very much correct. In fact, it's clear that the article's advice applies moreso to smaller companies than larger. I don't think that anyone would question that it doesn't scale.
> What do you think would happen at Acme Inc., a company with more than 1000 employees, if one division suddenly demoted/relieved the manager, gave each teammate a gratuitous sick/vacation policy, and rotated frontline employees through the management roles?
Divisional consistency is a concern for a large company. However, it is simultaneously possible to maintain a "skunkworks". It comes down to (unsurprisingly) the overall culture / mandate.
At twenty employees with several partners, a flat model works. Typically however, that approach only scales so far. As an organization starts hitting 30-40 people it starts to become hard to maintain informality and avoid territoriality. At 150 or so, you hit Dunbar's number.
http://en.wikipedia.org/wiki/Dunbars_number
Don't get me wrong, it sounds like a great place to work and an admirable way of managing people.
If the company does start to become bigger, and groups in charge of certain aspects of the products start to form, is it possible to still apply this model on the smaller formed groups? Then the weekly leader of each group meet and therefore encourages employees to maintain the "bigger picture" of the product.
Is this doomed to fail? what do you guys think?
Yes, that's what the "Scrum of Scrums" idea is meant to solve, in exactly that way:
http://www.xqa.com.ar/visualmanagement/2009/08/scrum-of-scru...
That's one of the reasons to keep your company small.
It's also a reason to make sure your middle managers are not paper-pushing bozos.
The idea that a middle manager has to be an entity that generates meetings, paperwork, stupid rules and hot air is prevalent, but by no means necessary. Hard to believe, but trust me - the good ones see their role as exactly the reverse.
There are tons of bad managers, but there are also a few excellent managers. I think it mostly depends on their focus.
Most middle managers (and most people in general) are focused on what they need to get done--the reports that need to be filed or the next status meeting to be scheduled. What good managers realize is that their work contributes absolutely nothing to the company (directly.) They are overhead in the purest sense of the word. All those i's to dot and t's to cross don't add a single cent to revenue. The only way they can make any contribution to the company is by making their people more effective. A great manager should be a hacker focused on her people's time rather than on code. She should be anticipating problems and annoyances and dealing with them before they blow up. If there's a fire hose of distractions, she should be the valve that slows the flow down to a trickle and routes the real issues to the appropriate people. Her goal should always be, "How can I make my people 1% more efficient?"
Too bad that most managers appear to be little more than a secretary with a checklist. (Actually, most secretaries I know are more useful than most managers.)
I agree, good managers are supposed to manage people, not tasks.
It is possible to keep this model working at a scale with a lot more employees. That being said it requires not only the right culture within the company, but the right people. Keep only the creme of the crop employees; stop 'fiscalizing' sick days, time sheets, and clocking hours; get the developers to get together at the start of an iteration (say every two weeks for example) and together come to a definition of priorities and tasks; measure results, not time in front of a computer. You as a business owner have - of course - the last word on any issue or priority; however empowering your employees, while keeping a friendly, social, and liberal company culture, will most probably have a positive effect on your workforce's drive and productivity. Hell everyone wants to the see the company grow when you really feel a part of it, as opposed to just another employee.
As a counterpoint though, I feel this approach only works with IT, Development and Design Departments. Lot's of other departments (Sales, Marketing, Customer Service, Implementation) require a certain structure and hierarchy to function correctly. In my experience, the more specialized the skill set, the more are liberty and culture important to further the drive of the team/teams.
Here's what I got out of this: employees are extremely obedient when you don't tell them the rules. Because they have to guess what's expected of them, they come up even more restrictive rules just to be on the safe side. Example: if you don't count vacation days, people take less time off. That's your employees putting money in your pocket!
Bonus: you can tell them you're giving them "autonomy", and also fire some managers.