At first, I was apprehensive reading this. I had to work 3 years with a "cowboy coder" colleague who I often found difficult to work with, and who rarely seemed to listen to whatever little feedback he occasionally received. I, however, have tended towards the opposite extreme - frequently asking my colleagues, "should I do it this way?", or "lemme bounce this off you", often when it would have been simpler and more efficient to charge forward and execute, then get feedback later. This article is a good reminder that there _is_ such a thing as too much collaboration. :)
Felt apprehensive, confused and nervous my favorite org was losing their marbles. Read the whole thing and began to realize this is exactly what I was trying to explain to my team to reduce bottlenecks in communication and deliverables. This is true interdependence. Meetings and feedback aren't inherently bad -- but if you feel blocked or dependent on feedback it's literally a bottleneck. Too much noise, just release then get the feedback. Now that I think of it so many of my engineers await my thumbs on a simple design/logic feature, like dang just ship it man!! "It's your call" 💣
There were many great pieces of advice in this article, and this one was a real gem for me personally: [Say “there are too many people involved. X, you are the driver, you decide.” (This is a great way to make friends btw)]
It’s not the fact that you collaborate, it’s how you collaborate, you are either working in an echo chamber or you are an authoritarian manager/boss/leader.
This is why diversity is important and it’s why including people with different backgrounds and ways of thinking is important.
The other thing that matters is what you are collaborating over, if you show up 75% done with something no one will really participate because at that point you are not looking for collaboration you are looking for confirmation.
And lastly this is why equity is important, you call it feedback, but feedback is the output of a review process not collaboration, you ask for feedback from people who don’t have as much as a stake in something as you do, that’s why it’s called feedback and not collaboration. Collaboration happens from the start, and you don’t provide feedback you discuss and come to an understanding, and move on from there.
So reviews suck, because everyone has an opinion and most opinions are like assholes, just full of shit.
You want to knock collaboration, try actually collaborating first, collaboration has been the source for every human innovation since we crawled out of the primordial soup, we are social animals, we need each other and few of us are smart enough or strong enough to survive alone, and the fact that you work in this industry and are complaining about working with people is incontrovertible proof you are not one of those people
This honestly sounds like more of an organizational problem than anything, a lack of guardrails and channels for effective collaboration. Don’t suppress collaboration. Govern it.
There are a few questions that pop up in my mind after reading this.
I understand your message: collaboration by default or “because it’s always better” can backfire and make people go slow when they could just ship and see how it works out. And I agree that we need to find a good balance there.
But then, there’s a clear focus on speed here while you work with PRs. I understand PostHog is open source, but aren’t PRs a big blocker to ship fast? Do you ever work in pairs / mobs when working on new product initiative?
Also, could there be the risk of creating a culture of “heroes” and “lone wolves” working alone most of the time?
Finally, how do you make sure that there’s diversity in perspective in the things you’re adding at PostHog?
This seems to avoid the issues that come from over collaboration. Open and balanced collaboration creates space for exploration, diverse perspectives, and creative risk-taking. It brings the best decision-making, problem-solving, conflict resolution, innovation, motivation, performance, and stronger relationships.
At Amazon we have the leadership principles. Everyone are leaders, owners, and we have the principles of Backbone Disagree and Commit, Bias For Action, and Deliver Results that contribute to balance the problems from too much collaboration. If something is not a one way door and is baked with data it unblocks those feedback loops.
You’re mistaking the broad scope of collaboration for a very particular style of collaboration that comes from a very particular culture, and the baby has gone out the window with the bath water.
The thing you’re diagnosing as a problem, which you call “collaboration”, is certainly a problem. But that’s not the only way to collaborate. There are so many models of collaboration, coordination, and control that could be chosen from. Not just this one.
What I can see from your analysis is that you’ve honed in on the (probably) dominant form of collaboration (or attempted collaboration) that’s taken hold certainly in the US and UK - and maybe others but I’m less familiar - and identified it as completely fucking terrible. I think you’re spot on there.
But what about the deeper question of how this came about? What cultural forces, economic incentives, political factors led to it, and can it be fixed?
I understand the temptation to see X, observe that X is bad, and so try -X, but usually -X is equally bad (or close to it) because it’s on the same dimension and shares the same source.
What you’ve pointed out in your article is the classic design by committee problem. It reminds me of all the British comedy from the 70s/80s onwards that came from making fun of communists and Socialist Worker Party people sitting around for days discussing how to begin a discussion about where to hold a meeting establishing the agenda for the upcoming summit.
So get rid of the committee structure for something else.
Looking at this issue from a network science point of view, you have an over-connectedness problem in the graph of employees. There is an optimum (which varies based on the circumstances), and it sounds like you’re far over it. But reducing connectivity to approaching 0% is also incredibly suboptimal.
Saying all that, perhaps your “burn it all down” strategy isn’t a bad way to build bad up to optimal (or nearer to optimal) connectivity. You just have to ensure that the systems and processes you create support a better network structure than the one that’s formed.
PostHog dropping a bomb 💣
At first, I was apprehensive reading this. I had to work 3 years with a "cowboy coder" colleague who I often found difficult to work with, and who rarely seemed to listen to whatever little feedback he occasionally received. I, however, have tended towards the opposite extreme - frequently asking my colleagues, "should I do it this way?", or "lemme bounce this off you", often when it would have been simpler and more efficient to charge forward and execute, then get feedback later. This article is a good reminder that there _is_ such a thing as too much collaboration. :)
this is the kind of "if you were to read one thing today that should be it" content
Felt apprehensive, confused and nervous my favorite org was losing their marbles. Read the whole thing and began to realize this is exactly what I was trying to explain to my team to reduce bottlenecks in communication and deliverables. This is true interdependence. Meetings and feedback aren't inherently bad -- but if you feel blocked or dependent on feedback it's literally a bottleneck. Too much noise, just release then get the feedback. Now that I think of it so many of my engineers await my thumbs on a simple design/logic feature, like dang just ship it man!! "It's your call" 💣
Came prepared to disagree. Loved every single second of it. Agree with I think every single thing.
Well done.
Banger! You put words to my feelings. thank you!
presumably because the bubbles are too collaborative hagagahahaha
There were many great pieces of advice in this article, and this one was a real gem for me personally: [Say “there are too many people involved. X, you are the driver, you decide.” (This is a great way to make friends btw)]
Lookout, your inexperience is showing.
It’s not the fact that you collaborate, it’s how you collaborate, you are either working in an echo chamber or you are an authoritarian manager/boss/leader.
This is why diversity is important and it’s why including people with different backgrounds and ways of thinking is important.
The other thing that matters is what you are collaborating over, if you show up 75% done with something no one will really participate because at that point you are not looking for collaboration you are looking for confirmation.
And lastly this is why equity is important, you call it feedback, but feedback is the output of a review process not collaboration, you ask for feedback from people who don’t have as much as a stake in something as you do, that’s why it’s called feedback and not collaboration. Collaboration happens from the start, and you don’t provide feedback you discuss and come to an understanding, and move on from there.
So reviews suck, because everyone has an opinion and most opinions are like assholes, just full of shit.
You want to knock collaboration, try actually collaborating first, collaboration has been the source for every human innovation since we crawled out of the primordial soup, we are social animals, we need each other and few of us are smart enough or strong enough to survive alone, and the fact that you work in this industry and are complaining about working with people is incontrovertible proof you are not one of those people
This honestly sounds like more of an organizational problem than anything, a lack of guardrails and channels for effective collaboration. Don’t suppress collaboration. Govern it.
There are a few questions that pop up in my mind after reading this.
I understand your message: collaboration by default or “because it’s always better” can backfire and make people go slow when they could just ship and see how it works out. And I agree that we need to find a good balance there.
But then, there’s a clear focus on speed here while you work with PRs. I understand PostHog is open source, but aren’t PRs a big blocker to ship fast? Do you ever work in pairs / mobs when working on new product initiative?
Also, could there be the risk of creating a culture of “heroes” and “lone wolves” working alone most of the time?
Finally, how do you make sure that there’s diversity in perspective in the things you’re adding at PostHog?
Thanks :)
This seems to avoid the issues that come from over collaboration. Open and balanced collaboration creates space for exploration, diverse perspectives, and creative risk-taking. It brings the best decision-making, problem-solving, conflict resolution, innovation, motivation, performance, and stronger relationships.
A lot of my substack talks about collaboration. I recommend to read https://wonderlead.tech/p/breaking-down-silos-in-tech
At Amazon we have the leadership principles. Everyone are leaders, owners, and we have the principles of Backbone Disagree and Commit, Bias For Action, and Deliver Results that contribute to balance the problems from too much collaboration. If something is not a one way door and is baked with data it unblocks those feedback loops.
You’re mistaking the broad scope of collaboration for a very particular style of collaboration that comes from a very particular culture, and the baby has gone out the window with the bath water.
The thing you’re diagnosing as a problem, which you call “collaboration”, is certainly a problem. But that’s not the only way to collaborate. There are so many models of collaboration, coordination, and control that could be chosen from. Not just this one.
What I can see from your analysis is that you’ve honed in on the (probably) dominant form of collaboration (or attempted collaboration) that’s taken hold certainly in the US and UK - and maybe others but I’m less familiar - and identified it as completely fucking terrible. I think you’re spot on there.
But what about the deeper question of how this came about? What cultural forces, economic incentives, political factors led to it, and can it be fixed?
I understand the temptation to see X, observe that X is bad, and so try -X, but usually -X is equally bad (or close to it) because it’s on the same dimension and shares the same source.
What you’ve pointed out in your article is the classic design by committee problem. It reminds me of all the British comedy from the 70s/80s onwards that came from making fun of communists and Socialist Worker Party people sitting around for days discussing how to begin a discussion about where to hold a meeting establishing the agenda for the upcoming summit.
So get rid of the committee structure for something else.
Looking at this issue from a network science point of view, you have an over-connectedness problem in the graph of employees. There is an optimum (which varies based on the circumstances), and it sounds like you’re far over it. But reducing connectivity to approaching 0% is also incredibly suboptimal.
Saying all that, perhaps your “burn it all down” strategy isn’t a bad way to build bad up to optimal (or nearer to optimal) connectivity. You just have to ensure that the systems and processes you create support a better network structure than the one that’s formed.
Either way, best of luck to you!
Report back on how it goes
I love articles that share real situations and experiences from real humans like this. Since too many articles are generated by AI nowadays
Ive got a better solution. Kill all your co-workers and do everything yourself.
Sounds to me you just work at a shitty company and are a huge dickhead. I would quite working for you immediately.