"Disagree and Commit" considered harmful
Last week I wrote a post (attempting to) deconstruct the well-known tech maxim to "move fast and break things". It definitely generated some good conversation! I want to follow up on a few discussion points that came out of the post. I had several readers reference "Slow is smooth and smooth is fast", which I hadn't heard before but is apparently a US Navy Seals slogan. I was also reminded of a similar slogan from The Lady Astronaut series by Mary Robinette Kowal1, where the main characters' motto is "slow is fast". @jawnsy on Mastodon followed up with the following quote which I thought was very insightful:
I think Silicon Valley often prefers the sugar high of early progress, but it doesn't take much experience to realize that it's often a false economy. It can work well for early-stage companies where the priority is finding product-market fit, but once there's traction, company needs change a lot
@mweagle (also on Mastodon) pointed out that the "broken windows" analogy (at least as it pertains to urban planning and crime) is a myth: a study from Northeastern University debunks this claim. Another colleague pointed out that the "broken windows" theory has led to a number of racist and horrible policing actions in New York City and elsewhere. It definitely wasn't my intention in that post to rehash traumatic and racist policies for any of my readers, so apologies if that happened. I've updated that post with this information as well.
The most common complaint/objection to last week's article, however, was about "disagree and commit" – I kinda just dropped it into the article as an aside, said it was a bad idea that many people think is good, and then moved on without really discussing it. I think this mention probably should have gone in a footnote, but it didn't, so we're going to do the next best thing and break it down in this blog post.
Thanks, Bezos (redux)
So just as a reminder: the "disagree and commit" concept (D&C for short), while not created by Jeff Bezos, was popularized by him in a Amazon shareholder letter back in 2016. The broader context of the letter was discussing how to make good decisions quickly, and one of the strategies he describes is D&C. The intended way this works is that for decisions that are "relatively" low impact and/or easy to course-correct, if there is disagreement between parties on how to proceed, one party should say "In the interest of making faster decisions, I disagree with your desired approach but I will commit to it anyways." This unblocks forward progress and allows the team to collect more data that either supports or disproves the decision that was made.
On paper, this is a great idea, and is actually a reasonable strategy for conflict resolution. It's a form of compromise: one party (or sometimes multiple parties) say "I don't like the chosen approach but I'm going to set aside my desires or opinions in favor of what's best for the team or organization." A related maxim is that you should have "strong opinions, weakly held"2. Again, the idea is, "you should care about things, but don't care so much that it gets in the way of organizational forward progress".
Again, this sounds like a great idea, but my thesis is that it doesn't actually work that well in practice. Let's look at why not.
Two words: power dynamics
There're a bunch of reasonable critiques you can make about D&C, but in my opinion, the most fundamental is the role of power and decision-making. My claim is that "disagree and commit" only works well in the absence of a power differential between the parties. If there is a power dynamic at play, D&C almost always turns into "disagree and commit, OR something really bad is going to happen to you."3 Consider, for example, the somewhat ironic "request" by current Amazon CEO Andy Jassy to "disagree and commit" on Amazon's return-to-office policy. The subtext4 there was "disagree and commit to returning to the office, or you'll be fired".
Not every request to "disagree and commit" is going to have as blatant of a consequence attached. But in a culture where D&C is normalized, in any sort of conflict it's common for someone involved to eventually say "We need to make a decision on this, will you disagree with me and commit to my approach?" If the person saying this is in a position of power (examples: maybe they're your boss, maybe they're a more senior engineer, maybe they're a white male and you are not), the subtext will always be "do it my way or else". Always. EVEN IF they don't mean it that way, that is how the person in less power will interpret it.
But let's say that never happens. Suppose you're a new engineer at TechBroCo and the director of your organization tells you on your first day "We have a 'disagree and commit' culture here," but it never comes up again. How will that shape your perception of the organization? Will you ever feel comfortable standing your ground on an issue if you think it's important enough? I mean, maybe, but my guess is that if you feel comfortable standing your ground it's because you've been afforded a lot of privilege.
But it gets worse. Let's say that the power dynamics are inverted. Let's say that the people who are in positions of authority always defer to the people with less authority. The boss lets the engineer do what they want instead of putting their foot down. Even in this scenario, there's an implied threat with D&C: "I'll disagree and commit to your approach… but you better not fuck it up." Again, the person in power may be the nicest, most well-intentioned person on the planet, and this will still be the message that the person with less power hears5.
The moral of the story here is that power dynamics make everything worse, and there's always a power dynamic.
The power of communication
I've maybe developed a reputation on this blog for complaining about things without offering solutions, but I think in this case I actually have a decent (but not perfect) solution: more transparency and better communication, particularly by those in power. Instead of saying "Will you disagree and commit to my approach?" your boss could instead have a frank conversation with you: "I am making this decision. These are the reasons why I am making this decision. I know you disagree with this approach, and I have listened closely to your concerns and objections, but this is the direction we are going." Or, alternately, "I am making the decision to let you try your approach. I have listened to your arguments and concerns, but I also need you to understand that these may be the consequences if your approach doesn't pan out."
These are much harder and more stressful conversations to have than throwing around catchy slogans like "disagree and commit". But it is possible to have these hard conversations in a way that communicates support and empathy, and, frankly, they're more honest. If you, as a person with more power, are able to consistently show the people around you that you a) listen to them, b) care about them, and c) are doing the best you can for them, you're going to have (in the long run) a much happier and more productive team.
But that's a hell of a lot harder than asking people to disagree and commit.
Postscript: social capital
The post is getting long but I have one last topic I want to cover before wrapping up. The previous section was targeted at people in positions of authority, but I also want to share a concept that is applicable to everyone, regardless of your current power dynamics. This concept is called "social capital"6. The idea behind social capital is that everyone has a certain amount of power or sway that they can spend in conflict scenarios. But when you run out of social capital, then you become ostracized, shunned, or have other negative consequences.
To make this a little more concrete: if you are on an engineering team, and you object to every single change or proposal that comes across your desk, eventually people are just going to stop asking for your opinion. It's the "I already know Jimmy7 is going to object to this idea so why bother asking him?" phenomenon.
The reason I like this framework is that it inherently captures the power dynamics at play: everyone has a different amount of social capital they can spend in a particular relationship. If you're the boss, you have a lot of social capital with your reports just by the nature of your position, but not very much with the CEO of your company: if the CEO knows your name, it's probably not a good thing. If you're a white male on the team, you inherently have more social capital than the Black woman. But here's the thing: regardless of your position, eventually your social capital will run out. You'll become the squeaky wheel that everyone ignores. Someone will complain about you to your boss. People will start going behind your back to get stuff done.
But here's the other thing about the social capital framework, is that you can build more of it! You can build social capital by compromising, by listening to those around you, by going along with their proposals even if you don't personally like them. In short, by becoming someone who is easy to work with and someone who cares about their colleagues.
From this perspective, if you are aware of the social capital you have, you can start making informed decisions: "I feel really strongly about this decision, so I'm going to spend some of my social capital to argue hard against it. This other decision doesn't actually matter that much, and I think I'm on thin ice with that person anyways, so I'm not going to push against it that much because I'd like to keep the relationship working."8
So there you have it: my thoughts on "disagree and commit". Next time, we're going to totally switch gears and talk about graph theory!
Thanks for reading,
~drmorr
This is a must-read, by the way, very well-crafted story and good writing.
As in last week's post, I have no idea what the origins of this saying are, and it's almost impossible to find because internet search has gotten so terrible. That Medium article claims, at least, that the idea was originally invented by Stanford University professor Paul Saffo.
I believe this is true even if the person in the position with more power has 100% the best intentions: there are so many stories of bad behaviour in tech that the person with less power will feel pressure to go along with the person in power even if there is legitimately no risk of retaliation in this specific circumstance.
Actually it wasn't even subtext… Jassy basically said the quiet part out loud here.
I once had an engineering leader tell me that I needed to "get over" this sort of knee-jerk panic response, which let me tell you, did not help me "get over" anything at all.
I'm not 100% sure I'm using the right term here; the Wikipedia article on social capital frames the concept by looking at network effects in a group or society, whereas I use the term from an individual perspective. Whatever, I'm not a sociologist, just go with it.
No Jimmy’s were harmed in the creation of this blog post; all similarities to any Jimmy in real life, whether real or imaginary, are purely unintentional.
If you've been paying attention, your response might be "Isn't this just D&C and/or 'strong opinions, weakly held' in disguise?" But, no, I don't think it is. It's not just some pithy slogan that buries all the relational aspects of the situation; instead, it's a recognition that most of your job is relationship management, and it frames your engagement with the people involved as a core part of the process.