It's too bad, because the best managers I've worked with were not good ICs, but they did multiply the effectiveness of the ICs they worked with, and so were absolutely invaluable to the company in a way that may not have shown up on paper. If those people exist in FB, as I'm sure they must, then they'd presumably get jettisoned as a result of this choice. That would be bad long term.
I used to despise managers, until I met two really good ones. This is after working with hundreds of them as a consultant. Actual, natural-born managers are such an incredibly rare species that you can go a long time without seeing any in the wild. It's basically the equivalent of the legendary 10x programmer — you hear about them, but it's rare to actually meet one, and there are sure a lot more people claiming to be one than who actually are.
My theory is that there are a set number of great managers in the world, and that number doesn't scale with the number of people who take on the role. They're just out there, being awesome, while a bunch of pretty lousy ones share their job title.
IMO the easiest way to solve this issue would be to normalize ICs making more than their managers. That way, only the people who really want to be managers are managing.
Not going to happen, fundamentally just due to supply and demand.
I have been a senior/principal engineer, as well as a director/senior director. The fact is that being a manager or director is just fundamentally a much harder job than being an IC. It's not that it's inherently more difficult, it's just that the day-to-day is much more of a grind than being an IC. For people wondering why engineering interviews can be so obscure/difficult, it's often because the cost of a bad hire can be catastrophic to a manager. I had a great team of about 30 people, except for 1 person who just couldn't get along with others. I spent about 80% of my energy on that person, and it sucked.
So for people wondering why managers get paid more, it's just that it's a shittier job that fewer people want to do than program.
I agree. Too often people feel becoming manager is the easiest way to advance their career (which most of the time equates to making more money). In many cases becoming manager is the only way to advance their career in the current org.
Yes, absolutely this. Once you hit senior ranks in engineering at a big tech firm, it's actually very difficult to move upwards. Very few people are ever made principal engineers or senior staff or even staff level for that matter. The jump in technical ability for those levels is quite large as well and can take many years to get a single promotion since openings are pretty slim as well (some orgs will only get 1-2 staff engineers).
If you switch to becoming a manager, your salary increases a lot and you have many more opportunities for promotion. So a lot of engineers hit senior and say "screw that, I can just be a manager" which is a much easier path to high salary.
It's very common and we need to incentivize people to stay as ICs.
But do you want managers who are people who want to be managers? It seems to me that many people who want to be managers are motivated by wanting to boss people around, i.e. they want power, not necessarily money. And some don’t want to do “the work” whatever that work might be.
I really think we need more “reluctant leaders”: people who are good at leadership but don’t want to lead, and yet can be relied upon to lead because of their sense of duty, especially their duty to the team. People who say, shit, I really don’t want to do this but damnit if I don’t step up, this will be a failure.
That already happens a fair amount. I've had two orgs where for stretches of time my top guys cleared more than I did, and I've known others in the same boat (they ran "rockstar" SME teams of various types). I wasn't offended by it either. I gave them air cover and they did exceptional work largely undisturbed, and periodically asked questions that made them not get too comfortable in the status quo. It was a fair trade off.
But ICs making more than managers is not what should be normalized; building teams with structures that optimize for performance rather than "management by spreadsheet" will get you further than just one dimensional rules about comp, especially in larger orgs where the comp tiers are deep.
That's an easy way to have zero managers and more problems than you started with. The vast majority of ICs are like children. It's not fun dealing with them. The only way one voluntarily takes on that role is for equal or greater pay.
That is pretty much the whole idea of the Staff+ Engineer track that mirrors the management track. The divergence normally happens after senior engineer level.
Smart companies have been doing things like this for decades. The first company I worked for 20 years ago had distinct tracks for management and engineering staff once people got to a certain level.
1. All people who really want to be managers are good managers
Or even
2. Best managers really want to be managers
To put it another way, a lot of people would rather be IC than managers. Not all of those people would be bad managers. In fact some of those might be the best managers.
Or to put it yet another way. Everybody wants to tinker. Managing is a pain. And the higher up, the fewer of the sane people would want it :-)
I think you are partially right, in that it should be normalized that some IC's may be delivering more value than their managers, and compensation should reflect that.
The broad version, that managers just don't contribute as much as the IC's they manage, doesn't hold much water in general - but may in a dysfunctional organization.
The problem with great managers is that you don't notice them. Things just work, no drama, no fires to extinguish, ... . It's also not clear why things are going that smooth, it just seems to work that way (clue: it's actually the manager that creates this environment).
I've worked in a few environments where managers were replaced. Only then can you really make the distinction between great managers and bad ones. With one you think "wow this is such a tough business, the manager is doing crazy effort to keep things going", and then another manager jumps in and "wow, this is an easy project. Manager is taking a walk through the park. All seems like a piece of cake."
100 times this. If you have an incredible manager, most likely you as an engineer feel little pressure and a lot of freedom. That is because that manager has his back full of arrows he intercepted before they had a chance to hit you.
I am a massive sufferer of imposter syndrome, I do my best to help everyone and smooth over things and set direction in a clear way with as much freedom in execution as possible.
but i am always second guessing myself. I know I am almost certainly a middling manager (if not a bad one), but I would like to improve.
Not because I love doing it, but because I don't trust that anyone else in my position (or who would take my position) actually wants to improve.
It's remarkably easy. For some reason it's simply not taught at schools.
1. Psychological security.
2. Crystal Clear targets, planning, goals, written down. To paraphrase this, a top manager will write their ICs an 8-12 week JIRA backlog. Obviously, these goals will iterate. However, the ironclad clarity is essential.
3. Positive, effective, coaching and feedback at the principle skill the IC performs. Aka, coding.
4. Lead by example. Conspicuous, hard, effort, devoting majority of time and energy to managing down and shielding.
The issue is simple. All of the above requires sweat from the manager.
Manager's are lazy, and nobody is watching them manage their IC's.
>the best managers I've worked with were not good ICs
Reminds me of baseball. Many of the best managers weren't the best when they were players. Some never even made it to the major leagues. But as a manager/coach they're incredibly well-regarded.
Managers don't have to have the hands on skills. What's more important is their understanding of the players, and making players better than they would be otherwise.
Those well-regarded managers are often so because they create a whole greater than the parts. Average managers break even. Bad managers decrease value.
Oh I completely agree with you. Good managers have an amazing multiplying effect on their team. People underestimate how important it is for a team to gel to be productive - although it's been said since the Peopleware days.
On the other hand, I've had maybe 2 great managers my entire career. The majority were either bean counters or empire builders. The latter is particularly toxic - they don't manage down at all beyond the minimum required. They focus almost exclusively on external optics and climbing the ladder.
So given that, I can see an appeal to trim down on managers. Of course, choosing the right managers to let go or convert is incredibly challenging.
You lament the decision by FB yet you say out of the hundreds of managers you worked with only two were good. You are saying most managers aren't good and it sounds like FB agrees with you and hence decision. I also think this might be a good decision because I do think managers need to be able to set up programming environment and at least be able to navigate the code. Kinder of like how some retail companies make head office staff spend some time at a real store so that they are aware of operating environment at the coal face.
Being a manager is hard and comes with a boatload of emotional responsibilities[1]. It takes a long term to learn, and our process of training managers is fundamentally broken.
It's not that there's a "set number of great managers", but we fail to train people, and many who could've been simply wash out because they were thrown into the deep end without any support. ("Hey, you're a good IC, you must be a manager")
And before somebody suggests it, no, MBAs are not the answer[2]. And I don't think we're fixing that if we don't manage to move away from a world where the ends justify the means. As a manager, your fundamental job is to care. Not to goose numbers, but to care enough about people that they feel safe enough to take risks and grow. In an environment that lives and dies by "what did you do for the metrics/the quarterly numbers", developing that care without losing the job or becoming a deep cynic is really, really hard.
If you have a good manager, hang on to them. If you're reporting to one, think hard about following them if they leave, because they are rare, and you might not get the experience again. And if you have the opportunity to become a manager, think carefully if the environment will allow you to care enough.
Hard agree, it's incredible that there's no systematic way to train such managers.
When I was younger I thought maybe that's what business schools did, but they produce a completely different kind of manager who's probably worse at managing people than they would've been had they not gone to business school in the first place.
It's remarkably easy. For some reason it's simply not taught at schools.
1. Psychological security.
2. Crystal Clear targets, planning, goals, written down. To paraphrase this, a top manager will write their ICs an 8-12 week JIRA backlog. Obviously, these goals will iterate. However, the ironclad clarity is essential.
3. Positive, effective, coaching and feedback at the principle skill the IC performs. Aka, coding.
4. Lead by example. Conspicuous, hard, effort, devoting majority of time and energy to managing down and shielding.
The issue is simple. All of the above requires sweat from the manager.
Managers are lazy, and nobody is watching them manage their ICs.
> were not good ICs, but they did multiply the effectiveness of the ICs they worked with
that sounds nice (meaning, "I know what you are saying") but it's not really true. As organizations grow, they become less efficient. If a manager can keep a team of N working at N*1 productivity, that's amazing, but it's not a multiple. Now, people can be dragged down by poor communication, poor leadership, changing targets, stepping on each other's feet, bad attitudes, etc. and a good manager can get that 0.1 efficiency multiplier back up to something approaching 1, so it might feel like a multiple.
> Actual, natural-born managers are such an incredibly rare species that you can go a long time without seeing any in the wild.
So how do we replicate this ability in people? For sports, there are academies to surface and nurture sporting talent. As a society, at least to me, it seems we have not been able to do something similar for "management" at societal scale. Maybe because the whole domain requires a broader set of competencies in many fields, including the specific technical domain, the humanities, business skills, and political savvy, amongst others.
Maybe we no longer need to replicate this with people. Instead, repeating what I said elsewhere in the thread: consider how there are natural-born managers, and how there are natural-born artists, and how generative AI like Stable Diffusion can deliver good results for those that are not natural artists, we may actually need to turn to generative AI to solve for managerial tasks in an optimal way across large orgs. That might even resolve issues with misaligned incentives.
Training such an AI could be done on a curated corpus of HBR-type case studies. And you could actually test performance of different corpuses (reflecting different management styles, for example) by measuring real-world performance of the corresponding business units in which these AIs operate.
There is no such thing as a natural born manager. Like there isn't a natural born electrician, politician or nurse.
There are combinations of skills that make one person far far better than everybody else, for that job. Some adapt existing skills, some learn new ones. But no one is natural born great at anything useful.
Those skills are acquired, either with willed direction from internal intuition or very rarely forced via external environment. But still acquired, not handed down, not born with it.
Can anyone acquire skills and be great? No! No one knows the winning combination of skills. It takes a great factor of luck.
Evidence of this is found in abundance. Ask someone great "How can I make my son be as great as you" and see how tangential an answer pops out. You'd think its because they want to keep it a secret. Its because they don't know.
You're taking OP's claim too literally. Obviously babies don't pop out of the womb ready to competently manage a team of 30 engineers.
The claim is that people are born with different aptitudes for management that allow them to progress faster along the management skill curve per unit of effort.
Managing is like any skill, if you want to develop it you have to work at it. For it to come naturally, is luck. That is, some collection of personal / personality traits, in the right place, at the right time. The problem is, few (are allowed / encouraged) to develop their manager-ness.
That aside, managers are a function of leadership. Got a bad manager? That means some leader is not doing their job.
Yes, there are few great managers, just as their are few great leaders.
My theory is, we could have more great leaders (or at least on average better), but we collectively use that word / title (i.e., leader) so loosely that in too many cases people get the luxury of wearing the crown without out actually doing the work.
If it doesn't walk like a duck...and it doesn't talk like a duck...yet it's still a duck?? That is creating a false sense of accomplishment.
This is an incredible post with a lot of insight. Can you please (please!) blog some stories about what you have seen? You are lucky to "work[ing] with hundreds of them as a consultant".
Do you think it can be trained into people? I do, but most companies are too lazy. Plus, it is not a one time exercise of X hours of training. It needs to be constantly monitored. In my career, I frequently see two-faced managers who say one thing when in front of the big boss, but do another when they are not around. (TL;DR: I call it: "The mouth moves, but the body doesn't.") It's always about the interpersonal clashes that drives away great ICs. The high performers get tired of the b-s from their manager and get hired away. Terrible loss for the company, but a (political) gain for the manager. (One less fly in the potato salad!)
Example: Below, someone mentioned "servant leadership". It is my first time to learn about term. It sounds brilliant. This idea seems possible to train into managers. Plus, the it is an interesting idea to strictly tie manager's comp to reports' comp. It would much better align the manager's incentives instead of today's world of "empire building".
> Below, someone mentioned "servant leadership". It is my first time to learn about term. It sounds brilliant.
It's a term from Agile/Scrum.
> Plus, the it is an interesting idea to strictly tie manager's comp to reports' comp.
Then the incentive exclusively becomes raising the reports' comp. That might be detrimental to organizational goals. Many salaries are structured as base + short-term incentive + long-term incentive, but clearly that obviously doesn't seem to work as well as it should, either. For companies, it's well-known that aligning objectives is complicated. After decades of 'Management Science', it doesn't seem to me that we are any closer to managing alignment across orgs ideally than we've been in the past. There may not be a silver bullet to this.
That said, thinking about OP's comments about natural-born managers, and how there are natural-born artists, and how generative AI like Stable Diffusion can deliver good results for those that are not natural artists, we may actually need to turn to generative AI to solve for managerial tasks in an optimal way across large orgs. That might even resolve issues with misaligned incentives. Training such an AI could be done on a curated corpus of HBR-type case studies. And you could actually test performance of different corpuses by measuring real-world performance of the corresponding business units in which these AIs operate.
I had cautiously gotten into management because I wanted to learn the skills of delegatin and really building through others and growing/coaching others. I thoroughly enjoyed this. But over time what I had realized was the role had been changing (not sure if it was just for pandering) towards being a mouth piece for leadership and hr - having to relay decisions I didn't believe in (perf seasons, promo rationales, layoffs anybody?) As if they were on purely my own. Given Ive always enjoyed building this may actually be a good forcing function to go back to IC and just ... Build!
I've recently held a senior management position then left and obtained a technical position at a different organisation.
Not kidding, I nearly cried with happiness moving away from SM and into a tech position again. I was GIVEN work to do. I had NO responsibilities for others. Moreover, I was DISCOURAGED from attending unnecessary meetings. I arranged no meetings! Not one!
Went from 7 direct reports to zero. No more approving holiday requests. No more performance reviews. No more management town-halls. No more arguing strategy with anyone. It was brilliant. 'Please write a procedure that does X. Return it by Thursday.' 'Please optimise this statement that hangs during the overnight run'. Yes, absolutely, more of that please.
I too went from management back to coding (freelance) since the start of the pandemic and I also love it. Today I happened to notice that Google Calendar was reporting an anomalous week in terms of meetings: I had 2.5 hours of meetings in my calendar this week, while my average right now is 1.2.
I average less than an hour and a half of meetings per week! I have so much more time in my life for the things I love, it’s crazy.
It because a theory of mine that a large part of my job function was to simultaneously water down accountability from above, while adding enough complexity to the layers below me to discourage them from changing the status quo.
The double edged sword being, without that authority, that same power structure is forced on me, and now I’m more aware of it.
I have a feeling being in middle management leads to that squeeze.
If you're closer to the doing, you are more isolated from the companyspeak nonsense. If you are in the upper part of the food chain, you get to enjoy more influence, freedom, a better risk/reward tradeoff, with costs you mention.
In the middle you're kind of screwed from both ends.
I had a couple of accidental ah-ha moments about this. Before I got into management all managers (fully up the stack) seemed like pointy hairs bosses. But what is easy to miss in Dilbert is that the PHB is really part of exec suite and rarely your front line manager. A hated front line manager might be usually incompetent or just inexperienced than evil!
The other thing was demonization of front line managers. There was always a terrible (there are definitely terrible traits) managers but never terrible "leaders". Even worse was you'd see articles by hr "influencers" expounding this. Where are the same hr loud speakers now that "leadership" is the one inflicting these layoffs and ridiculous policies?
The best thing you can do in these roles is protect your team. It's your primary job to ensure they can build and with minimal wasted efforts. If you view it as building for the company, your team is a victim to all the dysfunctions of the organization and nobody wants to be on that team. The mouthpiece part is tricky, I always try to give the official message with a dose of my personal candid opinion on it even if it's in opposition. You just can't lay it on too thick and try to find a way to spin it to a positive as to not be demotivating for the team, "it might be challenging, but at least we'll learn X and they're aware of the risks that are at top of your minds."
> towards being a mouth piece for leadership and hr - having to relay decisions I didn't believe in (perf seasons, promo rationales, layoffs anybody?) As if they were on purely my own.
The role doesn't change though. It has been like that forever.
This is another part of Management (whether you like it or not). Management should be seen as a cohesive "Leadership" unit because if the leaders aren't on the same page, what good are them?
The other thing about Management/Leadership is that they should be able to relay the message in the right format and at the right level for their direct reports. C-level typically have OKR/Objective in 3-5 bullet points (the bigger the company, the more bullet points) and it's the VP and Dir jobs to break these individual points to high-level goals for their Organizations. It is the Manager's job to distill it even further to their level + their direct reports.
It's one way to keep ICs on the same frequency with leadership and to execute for a common goal.
Most ICs don't want to attend meetings, don't want to hear "high-level goals", and more importantly, don't have the skill to consume those high-level goals at their level. They need help from various layers of leaderships. Sure, there are those exceptional ICs that wouldn't mind to be involved at that level and understood the social/human aspect part of the work but let's be frank here: majority of ICs just want to bang keyboard, produce clean code, and marvel at the end results (ignoring everything in between).
Imagine this from a military perspective. The general/field marshal etc makes the call that you're going over the top and as a smart Major or whatever you know this means most of your men are going to die. Are you going to say to your men:
"Hello gents, we have our new orders, we're going to go over the top tomorrow at 8am and try and take the village one mile down the road, however, I think this is a horrible idea and you're all probably going to die."
Even if you really believe this is the case, it's a bad idea to tell your men as it means the plan (which is going to happen anyway) is even LESS likely to succeed. That is, if you can make your men believe the plan is genius and they're going to destroy the enemy easily, it may give them more confidence and leave more of them alive at the end of it.
In general, if you want your reports to accept company decisions, it is in your interest as a manager to align your stated opinions with company policy. Openly disagreeing, or just trying to take a neutral position will only reduce faith in the company and sow unrest for which you will be held accountable.
> Is this actually the expectation at the places you work?
Can you imagine a manager saying, "I don't agree with these decisions, but you need to do A, B, C." This might be being frank and open with your reports, but at the same time you reduce their motivation. And if top management learns about this, they won't be happy.
So yes, there is such unspoken pressure to at least not to distance oneself from the things you ask others to do, whether they come from you or from your boss.
I'm not a manager, but at the very least it seems like an implicit expectation associated with the role. I wouldn't be surprised if it was explicit at some companies.
Depending on the company saying I don't believe X gets relayed back up your chain with eventual unpleasant consequences (hey so and so doesn't agree with this management decision so they must be a misfit and can't be trusted to put the company first)
I think this is a great move. My understanding is that some reporting chains in facebook have 12 links between IC and Zuckerberg, which is crazy inefficient. Enormous organizations are able to run effectively with 5-6 links from IC to CEO. Managers managing managers managing turtles all the way down is a horrible form of "administrative bloat". You see bloat like this in enormous organizations like colleges, governments, and now monopolistic tech companies. This is besides the point that I don't think that big tech has successfully groomed many managers into people are capable of actually managing businesses; they don't drive value, they preside over decisions and make a case for higher headcount in the next planning cycle.
Wow. I'll say that even in the Army, famous for its bloat, there are arguably fewer than 12 links (following the legal chain of command) between the lowest ranking private and the President of the United States...
Of course we also pad that with a lot of senior enlisted advisors, executive officers, chiefs-of-staff, deputy commanders, etc. who can effectively be additional buffers between each echelon.
If you want a flatter org structure, it seems like it would be easier to achieve this if manager/IC roles were distinct. If a manager is also writing code, then they won't have the time to manage as many other employees, and the total number of managers/layers would need to increase.
They might be pushing hybrid manager/coders to move back to full IC / tech lead type roles, while keeping the dedicated managers with bigger spans of control. That would have the effect they want, anyway.
yea, I don't think the manager/IC thing is a point of leverage TBH. I think the real issue here is that you have senior managers that manage 4 ICS and 1 manager of a 6 person team, and that manager should just report to a director and the senior manager should be retitled "manager" and manage 8 ICs instead of 4.
IC reports to tech lead manager (link 1). TLM reports to senior manager (link 2) senior manager reports to other senior manager (yes, this happens, link 3), senior manager reports to director (link 4) director reports to senior director (link 5), senior director reports to VP (link 6) VP reports to SVP (link 7), SVP reports to C level (link 8) C level reports to CEO (link 9). I saw these reporting chains when I worked at FAANG, and I can imagine its pretty easy to stuff 3 more links in here to hit 12.
I have seen such chains at Amazon. IC -> Manager -> Sr Manager -> Director -> VP -> SVP -> CEO of AWS -> CEO of Amazon.
That's only 7, but some links repeat. For example, I have seen Sr Manager report to another Sr Manager. Likewise, a VP report to another VP, and a Director report to another one.
At Meta many engineers tried management and then went back to being an individual contributor. I'm not sure of the exact numbers but it was a very common career path. It did build up empathy for management on the senior IC side and meant that teams didn't get stuck with managers that didn't actually want to be a manager. If the company is not growing headcount this move makes perfect sense.
That's a pretty common thing once ICs reach a certain age. I've seen it everywhere I've worked. Sometimes it's because they fear age discrimination but more commonly I've found it caused by the organization not providing a great career path beyond a certain level. With wisdom comes the ability to accept that it's OK to stick at an engineering career level for years at a time.
Chamath talked about manager bloat on a recent All-in podcast in a way that seemed to make sense. Basically there is pressure to grow your career by moving from being an IC to a manager and once a manager there are incentives to grow your number of reports. The organizational efficiency suffers when you have top tier ICs start managing teams of average people. You might get a team of 6 engineers that’s only doing 2x as much work as the former IC that’s managing them used to do, which means the cost per unit of work has increased tremendously.
If the reward for competence is promotion, in the steady-state equilibrium everybody's incompetent at their current job: If they were competent, they would have been promoted into a different job.
It's called the Peter Principle [1] and it's pretty well-known.
I suppose Facebook's realized it's in a Peter Principle situation, and the remedy for it would, in theory, be to send some of the people back to previous jobs where they were competent.
The problem is that, in practice, it seems a lot like a demotion, especially if the original move was touted as a promotion, and the pay and prestige of your new old job is less.
It's a real puzzle how this can be done without demoralizing the affected employees into quitting -- especially the best ones, who presumably have lots of good career options elsewhere, and whose value is the entire reason for doing this in the first place.
The premise of this always puzzled me. I’ve seen many people “promoted” to a “higher” level because they displayed characteristics that they would be more successful at that role than anyone else despite lacking competence in their current role.
I’ve also seen people who are great at their current role but lack the attributes to rise into higher roles.
So the premise of competence being rewarded by promotion doesn’t really seem useful in practice.
It is called 'Up or Out' culture. It is practiced everywhere, even in the military. You don't see 50+ year soldiers, do you? Except may be in 'Top Gun' sort of movies, which just only proves the point.
You are supposed to retire or give way to top performers to take the top places.
Some organizations have tried to limit manager-bloat by offering an alternative track for ICs (senior-> staff -> principle -> whatever). But the expectations at those levels have been hard to define and individual output seems to plateau or even degrade back to 'senior' level or even less when these high ranking ICs are asked to join a plethora of meetings on behalf of the team.
Years ago playing Civ 4 I distinctly remember hearing Leonard Nimoy's voice say "The bureaucracy is expanding to meet the needs of the expanding bureaucracy". That's what management in these companies is like. It became a meme that every 6 months you'd get an email saying "your manager's manager's manager's manager now reports to a new manager". You'd never seen, heard of or let alone met any of these people.
A large company like this should ahve a structure that's really no deper than:
1. CEO
2. SVP
3. VP
4. Director and really high level ICs
5. Manager and high-level ICs
6. ICs
Any deeper than that is organizational bloat. The above will allow ~5000 ICs per SVP without any layer having two many branches (eg a manager should really have no more than 10-15 reports).
This is better than layoffs because mid-level management is more responsible for bad company strategy than any IC is. Even though this is a step in the right direction, Zuck still faces two way bigger problems:
1. There is no vision for the company's key assets (ie FB, IG, WA). All of these seem to be relatively stagnant (if not in outright decline) and facing pressure from new platforms, most notably Tiktok as well as the behemoth that Youtube remains; and
2. There is absolutely no product market fit for the metaverse, certainly none that justifies sinking tens of billions of dollars into it.
By that structure, the maximum number of people that such a company should have is 7×7×7×7×7 = 16807 Given the best practices of number of direct reports [1].
How would you deal with the rest 30,000+ Full time employees that Facebook has?
> By that structure, the maximum number of people that such a company should have is 7×7×7×7×7 = 16807
That actually sounds like pretty reasonable limit. There are very, very few organizations that should exceed this size in my view. In fact, even 16807 sounds far larger than ideal in most cases.
I would say it's about as popular among the gamers at Meta as it is among gamers outside the company. I play it myself periodically, but I wasn't even aware that Mark was a fan of the game, and I've been here a long time.
The vast majority of the managers I've worked with wouldn't be particularly good ICs, not so sure of the wisdom of this move. Maybe it's a way to slow roll a bunch of layoffs.
Depends on the company, of course (and I've no idea how it works at Facebook), but some companies make their EMs out of ICs (much like Pokemon, I assume touching a special rock is required to effect the evolution) and it's somewhat common for EMs to re-IC-ify on their own anyway.
When the EM becomes an IC, how does this help the team grow their own skills and not get distracted by problems the EM should be moving out of their way?
When the EM is an IC, what stops them from either interfering with small stuff (I had an EM -- back then we called them "managers" -- stop us for a week trying to determine whether we should use List or Vector in Scala. Nobody was happy) or taking the big interesting problems for themselves?
"Here's an option you'll hate—take it, or leave" is a tried-and-true way of making people quit. "You've been re-assigned" is a common form for it to take.
It can't be that difficult to contribute to the achievement of absolutely tanking the company's value by investing in a dead-end VR-for-work platform that was always just a distraction to avoid responding in any way to Congress pressing him about election interference.
I love management and, after over 17 years in an IC capacity, find that it fulfills me in a way that writing code stopped doing. I was also a manager at Meta in a past life, and really didn't love the experience due to the nature of the company and roles of manager. I do think I'd really like being an IC there, due to the high level of autonomy and focus on doing impactful work.
Readit News
I used to despise managers, until I met two really good ones. This is after working with hundreds of them as a consultant. Actual, natural-born managers are such an incredibly rare species that you can go a long time without seeing any in the wild. It's basically the equivalent of the legendary 10x programmer — you hear about them, but it's rare to actually meet one, and there are sure a lot more people claiming to be one than who actually are.
My theory is that there are a set number of great managers in the world, and that number doesn't scale with the number of people who take on the role. They're just out there, being awesome, while a bunch of pretty lousy ones share their job title.
I have been a senior/principal engineer, as well as a director/senior director. The fact is that being a manager or director is just fundamentally a much harder job than being an IC. It's not that it's inherently more difficult, it's just that the day-to-day is much more of a grind than being an IC. For people wondering why engineering interviews can be so obscure/difficult, it's often because the cost of a bad hire can be catastrophic to a manager. I had a great team of about 30 people, except for 1 person who just couldn't get along with others. I spent about 80% of my energy on that person, and it sucked.
So for people wondering why managers get paid more, it's just that it's a shittier job that fewer people want to do than program.
If you switch to becoming a manager, your salary increases a lot and you have many more opportunities for promotion. So a lot of engineers hit senior and say "screw that, I can just be a manager" which is a much easier path to high salary.
It's very common and we need to incentivize people to stay as ICs.
I really think we need more “reluctant leaders”: people who are good at leadership but don’t want to lead, and yet can be relied upon to lead because of their sense of duty, especially their duty to the team. People who say, shit, I really don’t want to do this but damnit if I don’t step up, this will be a failure.
But ICs making more than managers is not what should be normalized; building teams with structures that optimize for performance rather than "management by spreadsheet" will get you further than just one dimensional rules about comp, especially in larger orgs where the comp tiers are deep.
Smart companies have been doing things like this for decades. The first company I worked for 20 years ago had distinct tracks for management and engineering staff once people got to a certain level.
1. All people who really want to be managers are good managers
Or even
2. Best managers really want to be managers
To put it another way, a lot of people would rather be IC than managers. Not all of those people would be bad managers. In fact some of those might be the best managers.
Or to put it yet another way. Everybody wants to tinker. Managing is a pain. And the higher up, the fewer of the sane people would want it :-)
The broad version, that managers just don't contribute as much as the IC's they manage, doesn't hold much water in general - but may in a dysfunctional organization.
I've worked in a few environments where managers were replaced. Only then can you really make the distinction between great managers and bad ones. With one you think "wow this is such a tough business, the manager is doing crazy effort to keep things going", and then another manager jumps in and "wow, this is an easy project. Manager is taking a walk through the park. All seems like a piece of cake."
I am a massive sufferer of imposter syndrome, I do my best to help everyone and smooth over things and set direction in a clear way with as much freedom in execution as possible.
but i am always second guessing myself. I know I am almost certainly a middling manager (if not a bad one), but I would like to improve.
Not because I love doing it, but because I don't trust that anyone else in my position (or who would take my position) actually wants to improve.
1. Psychological security.
2. Crystal Clear targets, planning, goals, written down. To paraphrase this, a top manager will write their ICs an 8-12 week JIRA backlog. Obviously, these goals will iterate. However, the ironclad clarity is essential.
3. Positive, effective, coaching and feedback at the principle skill the IC performs. Aka, coding.
4. Lead by example. Conspicuous, hard, effort, devoting majority of time and energy to managing down and shielding.
The issue is simple. All of the above requires sweat from the manager.
Manager's are lazy, and nobody is watching them manage their IC's.
They believe that supporting and developing their people is their main job. Execution is just a prerequisite to being able to do the rest.
[1] https://en.wikipedia.org/wiki/Servant_leadership
Reminds me of baseball. Many of the best managers weren't the best when they were players. Some never even made it to the major leagues. But as a manager/coach they're incredibly well-regarded.
https://fivethirtyeight.com/features/most-managers-are-heade...
Those well-regarded managers are often so because they create a whole greater than the parts. Average managers break even. Bad managers decrease value.
On the other hand, I've had maybe 2 great managers my entire career. The majority were either bean counters or empire builders. The latter is particularly toxic - they don't manage down at all beyond the minimum required. They focus almost exclusively on external optics and climbing the ladder.
So given that, I can see an appeal to trim down on managers. Of course, choosing the right managers to let go or convert is incredibly challenging.
It's not that there's a "set number of great managers", but we fail to train people, and many who could've been simply wash out because they were thrown into the deep end without any support. ("Hey, you're a good IC, you must be a manager")
And before somebody suggests it, no, MBAs are not the answer[2]. And I don't think we're fixing that if we don't manage to move away from a world where the ends justify the means. As a manager, your fundamental job is to care. Not to goose numbers, but to care enough about people that they feel safe enough to take risks and grow. In an environment that lives and dies by "what did you do for the metrics/the quarterly numbers", developing that care without losing the job or becoming a deep cynic is really, really hard.
If you have a good manager, hang on to them. If you're reporting to one, think hard about following them if they leave, because they are rare, and you might not get the experience again. And if you have the opportunity to become a manager, think carefully if the environment will allow you to care enough.
[1] https://fortune.com/2023/02/06/managers-impact-worker-mental...
[2] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2962506
When I was younger I thought maybe that's what business schools did, but they produce a completely different kind of manager who's probably worse at managing people than they would've been had they not gone to business school in the first place.
1. Psychological security.
2. Crystal Clear targets, planning, goals, written down. To paraphrase this, a top manager will write their ICs an 8-12 week JIRA backlog. Obviously, these goals will iterate. However, the ironclad clarity is essential.
3. Positive, effective, coaching and feedback at the principle skill the IC performs. Aka, coding.
4. Lead by example. Conspicuous, hard, effort, devoting majority of time and energy to managing down and shielding.
The issue is simple. All of the above requires sweat from the manager.
Managers are lazy, and nobody is watching them manage their ICs.
that sounds nice (meaning, "I know what you are saying") but it's not really true. As organizations grow, they become less efficient. If a manager can keep a team of N working at N*1 productivity, that's amazing, but it's not a multiple. Now, people can be dragged down by poor communication, poor leadership, changing targets, stepping on each other's feet, bad attitudes, etc. and a good manager can get that 0.1 efficiency multiplier back up to something approaching 1, so it might feel like a multiple.
So how do we replicate this ability in people? For sports, there are academies to surface and nurture sporting talent. As a society, at least to me, it seems we have not been able to do something similar for "management" at societal scale. Maybe because the whole domain requires a broader set of competencies in many fields, including the specific technical domain, the humanities, business skills, and political savvy, amongst others.
Maybe we no longer need to replicate this with people. Instead, repeating what I said elsewhere in the thread: consider how there are natural-born managers, and how there are natural-born artists, and how generative AI like Stable Diffusion can deliver good results for those that are not natural artists, we may actually need to turn to generative AI to solve for managerial tasks in an optimal way across large orgs. That might even resolve issues with misaligned incentives.
Training such an AI could be done on a curated corpus of HBR-type case studies. And you could actually test performance of different corpuses (reflecting different management styles, for example) by measuring real-world performance of the corresponding business units in which these AIs operate.
There is no such thing as a natural born manager. Like there isn't a natural born electrician, politician or nurse.
There are combinations of skills that make one person far far better than everybody else, for that job. Some adapt existing skills, some learn new ones. But no one is natural born great at anything useful.
Those skills are acquired, either with willed direction from internal intuition or very rarely forced via external environment. But still acquired, not handed down, not born with it.
Can anyone acquire skills and be great? No! No one knows the winning combination of skills. It takes a great factor of luck.
Evidence of this is found in abundance. Ask someone great "How can I make my son be as great as you" and see how tangential an answer pops out. You'd think its because they want to keep it a secret. Its because they don't know.
The claim is that people are born with different aptitudes for management that allow them to progress faster along the management skill curve per unit of effort.
That aside, managers are a function of leadership. Got a bad manager? That means some leader is not doing their job.
Yes, there are few great managers, just as their are few great leaders.
My theory is, we could have more great leaders (or at least on average better), but we collectively use that word / title (i.e., leader) so loosely that in too many cases people get the luxury of wearing the crown without out actually doing the work.
If it doesn't walk like a duck...and it doesn't talk like a duck...yet it's still a duck?? That is creating a false sense of accomplishment.
https://www.indeed.com/career-advice/finding-a-job/what-is-a...
Do you think it can be trained into people? I do, but most companies are too lazy. Plus, it is not a one time exercise of X hours of training. It needs to be constantly monitored. In my career, I frequently see two-faced managers who say one thing when in front of the big boss, but do another when they are not around. (TL;DR: I call it: "The mouth moves, but the body doesn't.") It's always about the interpersonal clashes that drives away great ICs. The high performers get tired of the b-s from their manager and get hired away. Terrible loss for the company, but a (political) gain for the manager. (One less fly in the potato salad!)
Example: Below, someone mentioned "servant leadership". It is my first time to learn about term. It sounds brilliant. This idea seems possible to train into managers. Plus, the it is an interesting idea to strictly tie manager's comp to reports' comp. It would much better align the manager's incentives instead of today's world of "empire building".
It's a term from Agile/Scrum.
> Plus, the it is an interesting idea to strictly tie manager's comp to reports' comp.
Then the incentive exclusively becomes raising the reports' comp. That might be detrimental to organizational goals. Many salaries are structured as base + short-term incentive + long-term incentive, but clearly that obviously doesn't seem to work as well as it should, either. For companies, it's well-known that aligning objectives is complicated. After decades of 'Management Science', it doesn't seem to me that we are any closer to managing alignment across orgs ideally than we've been in the past. There may not be a silver bullet to this.
That said, thinking about OP's comments about natural-born managers, and how there are natural-born artists, and how generative AI like Stable Diffusion can deliver good results for those that are not natural artists, we may actually need to turn to generative AI to solve for managerial tasks in an optimal way across large orgs. That might even resolve issues with misaligned incentives. Training such an AI could be done on a curated corpus of HBR-type case studies. And you could actually test performance of different corpuses by measuring real-world performance of the corresponding business units in which these AIs operate.
What did they do that made them so special?
Not kidding, I nearly cried with happiness moving away from SM and into a tech position again. I was GIVEN work to do. I had NO responsibilities for others. Moreover, I was DISCOURAGED from attending unnecessary meetings. I arranged no meetings! Not one!
Went from 7 direct reports to zero. No more approving holiday requests. No more performance reviews. No more management town-halls. No more arguing strategy with anyone. It was brilliant. 'Please write a procedure that does X. Return it by Thursday.' 'Please optimise this statement that hangs during the overnight run'. Yes, absolutely, more of that please.
More money too, oddly.
I average less than an hour and a half of meetings per week! I have so much more time in my life for the things I love, it’s crazy.
It because a theory of mine that a large part of my job function was to simultaneously water down accountability from above, while adding enough complexity to the layers below me to discourage them from changing the status quo.
The double edged sword being, without that authority, that same power structure is forced on me, and now I’m more aware of it.
If you're closer to the doing, you are more isolated from the companyspeak nonsense. If you are in the upper part of the food chain, you get to enjoy more influence, freedom, a better risk/reward tradeoff, with costs you mention.
In the middle you're kind of screwed from both ends.
The other thing was demonization of front line managers. There was always a terrible (there are definitely terrible traits) managers but never terrible "leaders". Even worse was you'd see articles by hr "influencers" expounding this. Where are the same hr loud speakers now that "leadership" is the one inflicting these layoffs and ridiculous policies?
Not that I expect them to do anything with it…
The role doesn't change though. It has been like that forever.
This is another part of Management (whether you like it or not). Management should be seen as a cohesive "Leadership" unit because if the leaders aren't on the same page, what good are them?
The other thing about Management/Leadership is that they should be able to relay the message in the right format and at the right level for their direct reports. C-level typically have OKR/Objective in 3-5 bullet points (the bigger the company, the more bullet points) and it's the VP and Dir jobs to break these individual points to high-level goals for their Organizations. It is the Manager's job to distill it even further to their level + their direct reports.
It's one way to keep ICs on the same frequency with leadership and to execute for a common goal.
Most ICs don't want to attend meetings, don't want to hear "high-level goals", and more importantly, don't have the skill to consume those high-level goals at their level. They need help from various layers of leaderships. Sure, there are those exceptional ICs that wouldn't mind to be involved at that level and understood the social/human aspect part of the work but let's be frank here: majority of ICs just want to bang keyboard, produce clean code, and marvel at the end results (ignoring everything in between).
Is this actually the expectation at the places you work?
"Hello gents, we have our new orders, we're going to go over the top tomorrow at 8am and try and take the village one mile down the road, however, I think this is a horrible idea and you're all probably going to die."
Even if you really believe this is the case, it's a bad idea to tell your men as it means the plan (which is going to happen anyway) is even LESS likely to succeed. That is, if you can make your men believe the plan is genius and they're going to destroy the enemy easily, it may give them more confidence and leave more of them alive at the end of it.
Can you imagine a manager saying, "I don't agree with these decisions, but you need to do A, B, C." This might be being frank and open with your reports, but at the same time you reduce their motivation. And if top management learns about this, they won't be happy.
So yes, there is such unspoken pressure to at least not to distance oneself from the things you ask others to do, whether they come from you or from your boss.
Why could you not say "I disagree with X, Y, and Z, but I am going to follow through with it because 1, 2, and 3"?
Private -> Team Leader -> Squad Leader -> Platoon Leader -> Company Commander -> Battalion Commander -> Brigade Commander -> Division Commander -> Corps Commander -> Combatant Commander -> SECDEF -> POTUS
Of course we also pad that with a lot of senior enlisted advisors, executive officers, chiefs-of-staff, deputy commanders, etc. who can effectively be additional buffers between each echelon.
That's only 7, but some links repeat. For example, I have seen Sr Manager report to another Sr Manager. Likewise, a VP report to another VP, and a Director report to another one.
This is a normal thing. There's a normal fluidity based on what someone wants to do, IC or manager. There isn't really a pay difference.
Being a Meta manager is a lot of work, especially the formality, structure, and data-drivenness of performance reviews.
Almost anyone can move back and forth if they want and have some interest.
It's called the Peter Principle [1] and it's pretty well-known.
I suppose Facebook's realized it's in a Peter Principle situation, and the remedy for it would, in theory, be to send some of the people back to previous jobs where they were competent.
The problem is that, in practice, it seems a lot like a demotion, especially if the original move was touted as a promotion, and the pay and prestige of your new old job is less.
It's a real puzzle how this can be done without demoralizing the affected employees into quitting -- especially the best ones, who presumably have lots of good career options elsewhere, and whose value is the entire reason for doing this in the first place.
[1] https://en.wikipedia.org/wiki/Peter_principle
I’ve also seen people who are great at their current role but lack the attributes to rise into higher roles.
So the premise of competence being rewarded by promotion doesn’t really seem useful in practice.
You are supposed to retire or give way to top performers to take the top places.
https://www.ribbonfarm.com/2009/10/07/the-gervais-principle-...
... illustrates the fallacy of the Peter Principle.
"The Mythical Man-Month" seems relevant in any case.
Dead Comment
Years ago playing Civ 4 I distinctly remember hearing Leonard Nimoy's voice say "The bureaucracy is expanding to meet the needs of the expanding bureaucracy". That's what management in these companies is like. It became a meme that every 6 months you'd get an email saying "your manager's manager's manager's manager now reports to a new manager". You'd never seen, heard of or let alone met any of these people.
A large company like this should ahve a structure that's really no deper than:
1. CEO
2. SVP
3. VP
4. Director and really high level ICs
5. Manager and high-level ICs
6. ICs
Any deeper than that is organizational bloat. The above will allow ~5000 ICs per SVP without any layer having two many branches (eg a manager should really have no more than 10-15 reports).
This is better than layoffs because mid-level management is more responsible for bad company strategy than any IC is. Even though this is a step in the right direction, Zuck still faces two way bigger problems:
1. There is no vision for the company's key assets (ie FB, IG, WA). All of these seem to be relatively stagnant (if not in outright decline) and facing pressure from new platforms, most notably Tiktok as well as the behemoth that Youtube remains; and
2. There is absolutely no product market fit for the metaverse, certainly none that justifies sinking tens of billions of dollars into it.
Disclaimer: ex-Facebooker.
Disclaimer:
How would you deal with the rest 30,000+ Full time employees that Facebook has?
[1] https://www.inc.com/jim-schleckser/how-many-direct-reports-s...
That actually sounds like pretty reasonable limit. There are very, very few organizations that should exceed this size in my view. In fact, even 16807 sounds far larger than ideal in most cases.
10101077 = 49000
Just curious: Is Civ something very popular at FB or something? I know Zuck himself plays Civ [0].
[0] https://www.facebook.com/civ/posts/thanks-mark-zuckerberg-fo...
When the EM is an IC, what stops them from either interfering with small stuff (I had an EM -- back then we called them "managers" -- stop us for a week trying to determine whether we should use List or Vector in Scala. Nobody was happy) or taking the big interesting problems for themselves?
If he were making SVPs become ICs, I'd credit this move to him. Otherwise, it's just the big-business org doing big-business things.
The point is that the company doesn't need the managers.
Deleted Comment
Dead Comment