What it means to “disagree and commit” and how I do it (2016)
amazonianblog.com
amazonianblog.com
The good: it's taken relatively seriously, it's not just some empty PR line. It allows you to prove to managers and colleagues that you know what you are doing instead of following the heard. It encourages people to prove their ideas with prototypes instead of talk and more talk. It discourage "design by committee" and blame shifting. It's refreshing to be able to disagree openly and challenge popular views.
The bad: It takes a lot of self confidence. Some people might feel insecure due to impostor syndrome, upbringing or belonging to minorities and this creates a disadvantage. Also, it's really exhausting in the long term.
Neutral: if you disagree, commit and fail it becomes a very public failure. On the bright side it teaches people not to disagree too lightheartedly.
I don’t really agree, but I sort of understand the Darwinism of it.
In an organization where management isn’t good at supporting contributors, it’s probably a good indicator actually. In an organization with strong support structures, you can use good ideas from people who aren’t natural leaders. But organizations who can really support ideas originating from those people are rare.
In my experience self-confidence is a core attribute of a leader, lacking this means you doubt yourself or are uncomfortable owning your idea. If you cannot own your idea how do you expect the rest of your team or company to own it?
I don't understand this part - if you publicly disagree with a decision, and then that decision results in failure, weren't you just proved right? Or does it just get assumed that the failure was your fault due to you not committing sufficiently?
Been with the company 3 years and this has happened to me once.
I agree with op. I disagreed but committed in an issue I saw. It soon turned out to be a production issue for a different team, as I had correctly called out and warned about...
I got called out in 1 on 1 by my senior manager for not having enough backbone.
Not quite sure what you mean, there. Can you give an example, please? How does your disagreement make the failure more public?
Jeff Bezos makes the strong point that risk is important: "most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90%, in most cases, you’re probably being slow."
Bezos also emphasizes that disagree-and-commit is a hard requirement: "Leaders are obligated to respectfully challenge decisions when they disagree, even when doing so is uncomfortable or exhausting."
I agree on both terms. If you plan and test everything to death, you are wasting and in fact atrophying our ability to think on your feet. This is an important skill, especially during incident and emergency handling. It's a fine line to learn, and it changes with the team, and the company, and the projects, but it's important.
Challenging decisions is the best way to keep unproductive work away from the team, and unproductive work is one of the best ways to waste time and motivation of a team. Even if it exhausts me at the end of a week, it's a good thing to see 2-3 guys getting excited about a project they should do, because it doesn't contradict our values and it furthers our infrastructure.
Most colleagues in Amazon agree that the company encourages quick thinking versus deep/long-term way too much.
Then again Bezos wont be fired if he makes the wrong decision.
I also found that if you need to follow the principle too many times it's a good indicator the team/company goals/vision/ and yours don't align and - if circumstance permits doing so - leaving is better than keep up the fighting forever (which is extremely exhausting, from personal experience).
Following this principle too many times isn't necessarily an indication that the team doesn't align with you, it can also be a sign your communication skills are lacking, and you need to get better in how you convey your ideas (unfortunately talking from my own experience).
A good tie-breaker is usually picking the simpler idea, to avoid perfectionism or over-engineering that engineers sometimes fall into. It seems especially common with people straight out of uni, and lessens with experience.
Unfortunately the proposal can be a fait accompli. I used to have demoralizing arguments with my boss, a company VP, about his product directives that were blatantly unsupported by logic or data. In every case his response would be to shrug his shoulders and say "but it's a done deal". (Meaning his CEO wanted to do it.) I can look back & say with confidence that each change was a financial dud.
Go ahead & say that it's your option to leave the company in these situations. It's true - and not always a viable option.
That sounds quite similar to Lenin's principle of democratic centralism - once upon a time communist parties allowed for a brief period of internal discussion prior to accepting policy decisions (in theory), once the decision was made every party member was committed to the 'collective' decision.
https://en.wikipedia.org/wiki/Democratic_centralism
Now that one did not quite work out, because you always depend on the whims of who is heading the discussion; I wonder if the Bezos principle works any better, as there are no third party regulators that might constrain the one who is heading the discussions.
Once upon a time they had this theory of gradual convergence between a west dominated by big corporations and the Soviet Union (for example ' 2001: A Space Odyssey' made this assumption); this observation might have had a grain of truth in it.
That sounds like only something a resentful Westener could say, in feeling entitled to a society exactly as they wish it, dismiss the one they have as "practically similar to tyranny".
I think its fair to measure tyrannical states in units of genocide of their own citizens. By this measure, the western democratic capitalism stands at 0% tyranny.
Universal enfranchisement in the UK happened only after WW1, with more than 90% of the population unable to vote for the whole history of "voting". And germany was just simply a dictatorship.
In 1800 90% of all of europe were farmers. The political class to even create a political system was tiny.
Democracy is rule by the people. Votes among feudal lords for which is the most powerful is not democracy.
You made a claim about two political systems: USSR's communism and western democratic capitalism.
If you're going to equivocate any old political system with any other then the claim to "convergence" is incoherent. I'm not even sure what "covergence" means when we're equivocating 19th Germany dictatorship, democratic capitalism, and USSR communism.
"Convergence" is severely under-evidenced. And my original claim is only ever-more evidenced: it is only possible to imagine our present system of government "converging" to mass murder upon some resentful delusion.
I personally like the principles although I’m missing a statement about developing employees. Also, the combination of two values can have dire effects, e.g. “Insist on the Highest Standards” and “Deliver Results” was probably the reason why some amazon warehouse workers have peed in bottles, see https://news.ycombinator.com/item?id=16849520.
But whatever, the stock shares I kept when I left keep going up, so I'm all right
For example, one time I disagreed with my director, who spent his entire career navigating by the seat of his pants. I was generally considered an expert in the subject, after having written several whitepapers that had become canonical reference amongst higher level leaders. That didn't matter, he never conceded the point and he treated me like shit until his boss told him he was wrong (which was after I transfered to a different org). This demonstrates this problem on two levels: he wouldn't tolerate any disagreement from me right up until the point where he realized that he was the one with the unapproved opinion.
the other long-time intel—ism that has now been appropriated everywhere is the OKR system. it was really weird moving from intel to google and hearing someone from peopleops talk about “this part of google culture called OKRs.”
It's great if everyone understands and commits to it. You must be able to make your case (and possibly acquiesce) without suffering political damage.
Of course if you keep backing the wrong horse ....
The issue though is when you do bring it up and none of the leadership care and they plow ahead anyways. Headlong into another poor decision.
On to addressing points raised in the essay:
"If you have been following Amazon at all, you have probably heard what sounds like code language when we talk about how we get our work done. As I have mentioned before, our work like a kind of short-hand for the types of qualities that make people effective here. And they aren’t just for leaders of organizations, they are for everyone at Amazon. We are all leaders."
No, everyone at Amazon.com is not a "leader". Leaders get to set the terms by which they and others will do some job. Reports of worker exploitation (see https://stallman.org/amazon.html#exploiting for links to relevant stories to back up the claim) make it clear that not everyone has the freedom to determine how their own job should be structured. Even if those changes would result in allowing workers to live in reasonable conditions, earn a living wage, work under conditions that don't make them ill, work without "spout[ing] the ideology of devotion to the company" (as Stallman rightly put it on his personal website linked above), and work a full-length career not a short-stint part-time or "contractor" job that will leave the worker to have to find another way to make ends meet all while delivering goods and services in a reasonable time-frame. I'll give you a hint as to how this could play out for boxing goods at Amazon.com: it's fine if Amazon.com's customers waited an extra day or two in order to let the "pickers" take bathroom breaks, longer working breaks, and avoid on-the-job hazards like constantly hustling and fainting during the working day.
"You may have heard that Jeff Bezos dislikes social cohesion. It’s a detriment to business success because it causes people to stifle ideas and objections for the sake of keeping the peace. I’m the kind of person who can have a concern or objection sidetrack my attention and I need to get it out; at least have it heard."
I doubt many who work under someone else believes this. This is what the manager class tell each other and tries to believe themselves while the worker class (people with the least say in the relationship) knows that stifling one's ideas and objections is conducive to helping them keep their job. This applies at every level in the hierarchy -- mid-level managers suffer from this when talking to their bosses and exploit this when talking to their subordinates. But the largest set of people always exist in the least-empowered class.
Managerial glib puffery PR like this usually downplay or ignore why people work these jobs at all: they live in societies that don't pay enough to give the worker the flexibility of choosing where to work. Capitalism pushes people to trade their skills for sustenance and tries to make it seem like that's a right and proper tradeoff, while capitalism also ridiculously over-rewards a few at the top (most notably in the case of Amazon.com's Jeff Bezos). Bezos, who might be the wealthiest person, clearly makes enough money to where he could never work again and be fine (as so could many generations of his children). But Bezos is not alone in this over-reward scheme; the gap between rich and poor is the largest it has ever been and it is accelerating.
It's ironic that any discussion of this kind would be raised in a single-point-of-censorship discussion forum such as this; it's so easy for anonymous users to score someone's post low (which affects whether other readers see the post with difficulty or at all) instead of responding. All scoring systems that affect how others see the posts are censorship, without exception. They work that way because that's precisely what they were designed to do. They are a mechanised way of implementing the reality of talking to the boss and run directly counter to the self-deluding lie of "disagree and commit".
I read your entire post. Is there a point besides: Amazon = capitalism, capitalism = bad, Amazon = bad?