The article concludes with an anecdote about someone who found an individual contributor track to be more amenable to his talents/interests. My suspicion is that most people funneled into management positions would have much rather continued to be an individual contributor, and only moved into management positions due to the negative perception of not having direct reports. Perhaps this augurs a structural and corporate change in America, though I'm cynical enough to assume that managerial bloat will reassert itself over time in most companies.
I can imagine more companies pushing individual contributors more, and that would be good for companies. Including hiring manager/admin folk at lower salaries than most individual contributors.
It’s profit watching that keeps down manager/ parasite bloat. Or not.
In general I've always preferred firms that don't employ many entry level employees. The value add is often little, and good ones just move on to advance their careers faster. You end up wasting a lot of mid career peoples time managing them.
Musk promotes the idea of cutting too much to find out which jobs actually matter, then refilling them. So that's his answer to your Type 1/Type 2 question. It's an interesting idea - that you can't tell if a position matters or not without trying to do without it. His plan seems to work at X, where he cut the headcount dramatically and - despite all the dire predictions that the platform would collapse because he let go too many essential people - the technology seems to still function. It would be interesting to try this approach via DOGE at some federal agencies, (although probably not with, say, air traffic controllers.) But civil service rules and budgetary inertia are likely to prevent too much deployment of this strategy. However, in Argentina Milei has announced a rule that to hire 1 new bureaucrat, the agencies must cut 3. That seems like a good approach to finding the fat.
Two or so years ago I left my old company. Partly this is because I felt I was underpaid relative to people above me that seemed to contribute less. Partly I did not have confidence in senior management to manage the business well.
The senior managers they had in place were smart...twenty years ago when a different sort of paradigm made sense. They mostly just coasted along since then "take last year and add/subtract X%". When the IRA came along, all of a sudden you actually needed to understand how the product works for real, and they couldn't do it. You could explain it to them, but they constantly just didn't get it.
So I followed a long stream of younger and more talented people to the new employer, who paid us in many cases more than our old bosses. In the two years since my company made huge errors I told them not to do and lost a ton of business (which new company has picked up). I'm talking amounts that could have paid all our salaries for 1,000 years.
Finally, the incompetent senior management layer finally got put into retirement when the business losses piled up, but by now it's basically too late to fix. They also are still being cheap with the one smart manager they have left, despite freeing up lots of money by getting rid of the dinosaurs.
So what kind of story can we tell here?
Where the managers at my old firm overpaid aging dinosaurs that needed to get cleared out?
Or was my firm too cheap with labor, given that they lost all their best young talent to another company will to open their wallet big time?
And how should executive management tell? The dinosaurs were really good at telling them what they wanted to hear and producing quarterly KPIs they wanted to see, and bad decisions taken today often take awhile to show up in KPIs in the future when it may be too late.
The KPIs taking awhile, like 2-5 years, is a problem, but they do give good feedback, when they are real Key Performance Indicators.
Top line sales revenue is really hard to game. With incentives to make it larger than real.
Mid line costs is hard to game, incentives to make it smaller than real.
Profit & taxes are hard to game for more than a year or so. Not firing top managers who haven't changed when change is needed seems like a problem IBM had .... but is trying to fix. Yet getting rid of the old is something that can get them sued, if it's really just the age rather than the performance.
My product space is counter intuitive. Without the ability to abstract and project, success it basically random.
Twenty years ago the dinosaurs figured out something counter intuitive and future looking. The next time that was necessary they could not. There were more then two variables and more uncertainty.
@Arnold, one way to think about this is that senior execs typically gravitate to the same playbook they've seen others, both inside and outside their firms, execute on historically. i.e.: Making cuts vertically is a way to cut costs when things are no longer booming, without having to make strategic decisions (i.e. horizontal cuts). So it almost becomes a self-fulfilling prophecy, where ex ante you know you're going to be able to do this, so you (semi-unconsciously) let bureaucratic "fat" build up, knowing you have an expense buffer you can subsequently dip into, write off as a restructuring, and quickly boost your earnings/share in the quarters following. This is not an endorsement of the practice, merely an observation that tracks with the many times I've been involved in it and seen it from afar.
Several decades into my career in management/leadership - two observations: 1. I’ve never seen a cut that went too deep. I may have disagreed with individual decisions, but never thought “no fat left” after a cut. 2. There is a ‘dead wood’ dynamic that occurs in organizations due to a combination of cowardice and HR. It is hard to fire non-performers. An occasional cost cutting exercise gives air cover to putting a few or bunch of people on the list who should have been gone long before. This works the same for managers & employees. Do your cut, get a little leaner, and a couple years later you are back to bloated. Even if everyone is great - what are the chances that everyone hired 2 years ago fits what you need next year?
I think the underlying causes of overstaffing are fairly straightforward, yet hard to avoid (innate human biases and all). Growing the org when costs are coverable and things are going well is seen as a positive signal in almost every way regardless of specific composition, and of course much hiring and complexifying with management layers is necessary if you’re expanding and doing more. But for a variety of reasons exec teams tend to wait too long to make corrections, and often when there’s a need to correct, they don’t go deep enough — layoffs and major restructurings are hard, no one seeks them out.
As for the “trendiness” of companies downsizing and restructuring, I think this is as simple as finding the “top cover” so you’re not the only company doing it. There’s a mimetic aspect to companies noticing it’s okay to do because others are, too. Rather than being perceived, either by third parties or internal employees, as a negative wind isolated to the company, it can now be mentally grouped in with a “macro headwind” — which to a degree it is.
History shows that it takes a looooong time for the full effects of a new technology to be felt. And for a long time it was said that computers were everywhere but the productivity statistics. Perhaps this is computers getting closer to their full potential.
I'm inclined to believe that this is largely the answer. If automation adds value it's by allowing any given individual to do more work in less time, which to me translates to fewer people overall -- and thus fewer people to be managed overall.
I think GenAI also simplifies some of the more onerous managerial tasks (composing employee reviews, for instance), which either allows a manager to do more of them in less time, or spend less time on them and more time on other tasks... like managing. It may also make 'managing at scale' easier, though I'm less sure of that.
Finally, I would point out that at both the manager and individual contributor level, labor is not really fungible -- some people are better than others and some people will adapt faster / better to the new technology than others. That says to me that the wheat will, at the margin, differentiate itself from the chaff more distinctly. Thus (if you are running the company well), you can not merely cut lower-quality employees, but have tested, higher value employees take on that work, which should improve quality and add to overall company profitability, all things being equal. Those folks who remain are then increasingly compensated better and the gap between winners and losers expands. As this trends continues, it should be possible to manage larger and larger enterprises with fewer and fewer people.... maybe allowing a sub-Dunbar number of people to run a super-Dunbar organization. The effects of that could be... interesting. On the one hand, it should allow a large organization to be as nimble as a small organization while reacting to changes in the marketplace. On the other hand, such an organization would also lose some of the bureaucratic back-stops that catch errors and scuttle poorly thought out initiatives before they create unintended consequences. I think it stands to reason that those organizations could experience much higher highs, but the lows could be spectacular.
Anyway, that's where my thinking is at the moment.
"spend less time on them and more time on other tasks... like managing."
My experience is that managers spend most of their time on problem employees who mostly shouldn't be there. And for everyone else, more managing made things worse.
I agree with the first part, but (although I have certainly seen my fair share of it), the second part does not ring true. The best managers I have had / seen have spent their political capital keeping the organization from distracting / deterring the people they are managing so they are able to do the job they were hired to do. Good managers hire the right people and leave them alone to work. Great managers are also able to coach those people, improve them and elevate them in the organization. Truly remarkable managers -- leaders -- can do that at scale in a large organization. I have been fortunate enough to know a few who fit that last category. It was like watching athletes performing in the Olympics, but the sport was management.
Another example of this is lasers. I remember an uncle complaining that there was so much hype about lasers, but they weren't actually being used for anything. This was probably in the 70s, possibly 80s. Now lasers are practically ubiquitous.
I think this is the unwinding of DEI initiatives, staving off the "go woke, go broke" inevitability. Companies can coast for awhile on incompetency but of course the sum of all the bad decisions along the way fortify the axiom, "reality bats last." This seems to me to be an acknowledgement of that reality.
One more thing that might be related is the impact of requiring people to be in the office and those who refuse and are terminated. I expect that's a bigger issue for non-supervisory workers but maybe managers too.
In my experience, firms have one or two simple rules for staffing (e.g., assets under management per full-time-equivalent) and may have an associated productivity growth goal (e.g., 5% improvement in the measure). If growth doesn't materialize then cuts must be made to hit the EBITDA goal. Staffing AI investment (in heads and dollars) to remain competitive likely requires that incremental headcount be released from other units.
These last few years, the main steady hirers have been government and healthcare, negatively productive and who knows, respectively. Tech overshot following the boost that Covid gave the industry (see Peloton). The rest of the economy has been pretty darn stagnant as far as I can tell, while DEI and other compliance had to be staffed up until the election.
It was some 6 years ago that IBM offered me a package (extra months of severance) to leave, when I was 62. They were sued for age discrimination, but I didn't participate in it. They are still at it, cutting old, usually highly paid folk. Plus moving jobs to India.
IBM, with Watson (old style ai) and far too many Vice-President titles, should continue being a good target for management restructuring. I can even imagine a Jeopardy winning level of fact-checking Watson could do a job on the first draft of a chatGPT article with possible hallucinations, tho I have not heard of them doing it. Watson was supposed to help doctors, but IBM sold that Watson med business.
I sincerely hope that more middle managers get laid off -- and start trying to start their own companies. The US needs more small & medium companies, and fewer mid-managers.
Another aspect I saw unremarked upon but have seen take off over the last ten or so years in Silicon Valley - Lots of promotions to "management" to give someone an upward career trajectory but with only minimal managerial responsibilities and little or no pay increase. Basically the cousin to the title inflation that took off hugely over the last decade.
Every org I've come in to has had dozens of these instead of senior ICs. It creates an absolute mess in the org chart and is difficult to undo because it's viewed as a demotion. It's also a larger symptom of the absolute collapse of managerial ability in tech companies since the late nineties
They say a great programmer in Silicon Valley can be 10x productive as a good programmer, or a great manager can add 10x value as a good manager. In that environment, and when you are cash heavy, it makes sense to cast a wide net and hoard talent. Then you try and coach them up for a few years and see which ones worked out.
Arnold poses some great questions that industrial organization ("IO") economists have grappled with for decades. Ronald Coase (Nobel Prize 1991) and Oliver Williamson (2009 prize) attempted to explain why firms (corporations) exist at all and why and where the boundaries lie between companies on the one hand and external contracting on the other.
Arnold's blog earlier this week about "producers and parasites" also informs the answers to these questions. Large organizations in both the public and private sectors employ both producers and parasites. The latter tend to be very good at "office politics" and making themselves appear to be useful without actually being useful. But they have power because of their political skills. Organizations that have surplus to burn can afford to carry "parasites" for a while, until economic or competitive pressures force upper management to pay attention. I'd suggest that they pay more attention during economic downturns or when a disruptive actor appears - think Tesla versus the big car companies. If due to widespread economic shocks (recessions), then there would be a pattern of trimming the fat concurrently; but otherwise, as Arnold suggests, the phenomenon should be more randomly distributed across the economy.
The assumption of most observers of corporate restructuring is that cuts are made in a logical fashion. My experience working with executives is very different. One problem is that executives often make the mistake of promoting people who are excellent at domain skills (accounting for accountants, sales for sales people), but who do not have strengths to be leaders. The reason for the promotion is that company compensation plans assume that leadership and decision making, even at the lowest level, are more valuable than critical domain skills, which is often untrue.
A second observation: executives are managing a complex array of outcomes: corporate strategy, technology investment, succession planning, organizational structure constraints. Some executives are simply not adept of negotiating among these competing concepts and they layoff people periodically as a substitute for weaknesses in one or more.
I’m short on time today since I commented on Dan Williams’ post. I will just add the following regarding this: “What is the relationship between Type I errors (cutting too much and regretting it) and Type II errors (letting costs rise too much and regretting it)?” I have trouble remembering what Type I and Type II are partly because their names convey little meaning. Can we take the words in parentheses and come up with better names for them? Suggestions?
The article concludes with an anecdote about someone who found an individual contributor track to be more amenable to his talents/interests. My suspicion is that most people funneled into management positions would have much rather continued to be an individual contributor, and only moved into management positions due to the negative perception of not having direct reports. Perhaps this augurs a structural and corporate change in America, though I'm cynical enough to assume that managerial bloat will reassert itself over time in most companies.
I can imagine more companies pushing individual contributors more, and that would be good for companies. Including hiring manager/admin folk at lower salaries than most individual contributors.
It’s profit watching that keeps down manager/ parasite bloat. Or not.
In general I've always preferred firms that don't employ many entry level employees. The value add is often little, and good ones just move on to advance their careers faster. You end up wasting a lot of mid career peoples time managing them.
Musk promotes the idea of cutting too much to find out which jobs actually matter, then refilling them. So that's his answer to your Type 1/Type 2 question. It's an interesting idea - that you can't tell if a position matters or not without trying to do without it. His plan seems to work at X, where he cut the headcount dramatically and - despite all the dire predictions that the platform would collapse because he let go too many essential people - the technology seems to still function. It would be interesting to try this approach via DOGE at some federal agencies, (although probably not with, say, air traffic controllers.) But civil service rules and budgetary inertia are likely to prevent too much deployment of this strategy. However, in Argentina Milei has announced a rule that to hire 1 new bureaucrat, the agencies must cut 3. That seems like a good approach to finding the fat.
Two or so years ago I left my old company. Partly this is because I felt I was underpaid relative to people above me that seemed to contribute less. Partly I did not have confidence in senior management to manage the business well.
The senior managers they had in place were smart...twenty years ago when a different sort of paradigm made sense. They mostly just coasted along since then "take last year and add/subtract X%". When the IRA came along, all of a sudden you actually needed to understand how the product works for real, and they couldn't do it. You could explain it to them, but they constantly just didn't get it.
So I followed a long stream of younger and more talented people to the new employer, who paid us in many cases more than our old bosses. In the two years since my company made huge errors I told them not to do and lost a ton of business (which new company has picked up). I'm talking amounts that could have paid all our salaries for 1,000 years.
Finally, the incompetent senior management layer finally got put into retirement when the business losses piled up, but by now it's basically too late to fix. They also are still being cheap with the one smart manager they have left, despite freeing up lots of money by getting rid of the dinosaurs.
So what kind of story can we tell here?
Where the managers at my old firm overpaid aging dinosaurs that needed to get cleared out?
Or was my firm too cheap with labor, given that they lost all their best young talent to another company will to open their wallet big time?
And how should executive management tell? The dinosaurs were really good at telling them what they wanted to hear and producing quarterly KPIs they wanted to see, and bad decisions taken today often take awhile to show up in KPIs in the future when it may be too late.
It seems all of these are true.
FYI your posts are weird in that they don't word wrap on my phone and ONLY your posts. You somehow post them WIERD?
Also, when he replies to me the email notices go to junk. Rarely if ever anyone else.
The KPIs taking awhile, like 2-5 years, is a problem, but they do give good feedback, when they are real Key Performance Indicators.
Top line sales revenue is really hard to game. With incentives to make it larger than real.
Mid line costs is hard to game, incentives to make it smaller than real.
Profit & taxes are hard to game for more than a year or so. Not firing top managers who haven't changed when change is needed seems like a problem IBM had .... but is trying to fix. Yet getting rid of the old is something that can get them sued, if it's really just the age rather than the performance.
My product space is counter intuitive. Without the ability to abstract and project, success it basically random.
Twenty years ago the dinosaurs figured out something counter intuitive and future looking. The next time that was necessary they could not. There were more then two variables and more uncertainty.
@Arnold, one way to think about this is that senior execs typically gravitate to the same playbook they've seen others, both inside and outside their firms, execute on historically. i.e.: Making cuts vertically is a way to cut costs when things are no longer booming, without having to make strategic decisions (i.e. horizontal cuts). So it almost becomes a self-fulfilling prophecy, where ex ante you know you're going to be able to do this, so you (semi-unconsciously) let bureaucratic "fat" build up, knowing you have an expense buffer you can subsequently dip into, write off as a restructuring, and quickly boost your earnings/share in the quarters following. This is not an endorsement of the practice, merely an observation that tracks with the many times I've been involved in it and seen it from afar.
Several decades into my career in management/leadership - two observations: 1. I’ve never seen a cut that went too deep. I may have disagreed with individual decisions, but never thought “no fat left” after a cut. 2. There is a ‘dead wood’ dynamic that occurs in organizations due to a combination of cowardice and HR. It is hard to fire non-performers. An occasional cost cutting exercise gives air cover to putting a few or bunch of people on the list who should have been gone long before. This works the same for managers & employees. Do your cut, get a little leaner, and a couple years later you are back to bloated. Even if everyone is great - what are the chances that everyone hired 2 years ago fits what you need next year?
I think the underlying causes of overstaffing are fairly straightforward, yet hard to avoid (innate human biases and all). Growing the org when costs are coverable and things are going well is seen as a positive signal in almost every way regardless of specific composition, and of course much hiring and complexifying with management layers is necessary if you’re expanding and doing more. But for a variety of reasons exec teams tend to wait too long to make corrections, and often when there’s a need to correct, they don’t go deep enough — layoffs and major restructurings are hard, no one seeks them out.
As for the “trendiness” of companies downsizing and restructuring, I think this is as simple as finding the “top cover” so you’re not the only company doing it. There’s a mimetic aspect to companies noticing it’s okay to do because others are, too. Rather than being perceived, either by third parties or internal employees, as a negative wind isolated to the company, it can now be mentally grouped in with a “macro headwind” — which to a degree it is.
History shows that it takes a looooong time for the full effects of a new technology to be felt. And for a long time it was said that computers were everywhere but the productivity statistics. Perhaps this is computers getting closer to their full potential.
I'm inclined to believe that this is largely the answer. If automation adds value it's by allowing any given individual to do more work in less time, which to me translates to fewer people overall -- and thus fewer people to be managed overall.
I think GenAI also simplifies some of the more onerous managerial tasks (composing employee reviews, for instance), which either allows a manager to do more of them in less time, or spend less time on them and more time on other tasks... like managing. It may also make 'managing at scale' easier, though I'm less sure of that.
Finally, I would point out that at both the manager and individual contributor level, labor is not really fungible -- some people are better than others and some people will adapt faster / better to the new technology than others. That says to me that the wheat will, at the margin, differentiate itself from the chaff more distinctly. Thus (if you are running the company well), you can not merely cut lower-quality employees, but have tested, higher value employees take on that work, which should improve quality and add to overall company profitability, all things being equal. Those folks who remain are then increasingly compensated better and the gap between winners and losers expands. As this trends continues, it should be possible to manage larger and larger enterprises with fewer and fewer people.... maybe allowing a sub-Dunbar number of people to run a super-Dunbar organization. The effects of that could be... interesting. On the one hand, it should allow a large organization to be as nimble as a small organization while reacting to changes in the marketplace. On the other hand, such an organization would also lose some of the bureaucratic back-stops that catch errors and scuttle poorly thought out initiatives before they create unintended consequences. I think it stands to reason that those organizations could experience much higher highs, but the lows could be spectacular.
Anyway, that's where my thinking is at the moment.
"spend less time on them and more time on other tasks... like managing."
My experience is that managers spend most of their time on problem employees who mostly shouldn't be there. And for everyone else, more managing made things worse.
I agree with the first part, but (although I have certainly seen my fair share of it), the second part does not ring true. The best managers I have had / seen have spent their political capital keeping the organization from distracting / deterring the people they are managing so they are able to do the job they were hired to do. Good managers hire the right people and leave them alone to work. Great managers are also able to coach those people, improve them and elevate them in the organization. Truly remarkable managers -- leaders -- can do that at scale in a large organization. I have been fortunate enough to know a few who fit that last category. It was like watching athletes performing in the Olympics, but the sport was management.
Agreed on protecting.
While managers may coach more, overall more coaching happens peer to peer.
Another example of this is lasers. I remember an uncle complaining that there was so much hype about lasers, but they weren't actually being used for anything. This was probably in the 70s, possibly 80s. Now lasers are practically ubiquitous.
I think this is the unwinding of DEI initiatives, staving off the "go woke, go broke" inevitability. Companies can coast for awhile on incompetency but of course the sum of all the bad decisions along the way fortify the axiom, "reality bats last." This seems to me to be an acknowledgement of that reality.
Interesting topic and great questions.
One more thing that might be related is the impact of requiring people to be in the office and those who refuse and are terminated. I expect that's a bigger issue for non-supervisory workers but maybe managers too.
In my experience, firms have one or two simple rules for staffing (e.g., assets under management per full-time-equivalent) and may have an associated productivity growth goal (e.g., 5% improvement in the measure). If growth doesn't materialize then cuts must be made to hit the EBITDA goal. Staffing AI investment (in heads and dollars) to remain competitive likely requires that incremental headcount be released from other units.
These last few years, the main steady hirers have been government and healthcare, negatively productive and who knows, respectively. Tech overshot following the boost that Covid gave the industry (see Peloton). The rest of the economy has been pretty darn stagnant as far as I can tell, while DEI and other compliance had to be staffed up until the election.
It was some 6 years ago that IBM offered me a package (extra months of severance) to leave, when I was 62. They were sued for age discrimination, but I didn't participate in it. They are still at it, cutting old, usually highly paid folk. Plus moving jobs to India.
https://www.theregister.com/2024/09/18/ibm_job_cuts/
IBM, with Watson (old style ai) and far too many Vice-President titles, should continue being a good target for management restructuring. I can even imagine a Jeopardy winning level of fact-checking Watson could do a job on the first draft of a chatGPT article with possible hallucinations, tho I have not heard of them doing it. Watson was supposed to help doctors, but IBM sold that Watson med business.
I sincerely hope that more middle managers get laid off -- and start trying to start their own companies. The US needs more small & medium companies, and fewer mid-managers.
Another aspect I saw unremarked upon but have seen take off over the last ten or so years in Silicon Valley - Lots of promotions to "management" to give someone an upward career trajectory but with only minimal managerial responsibilities and little or no pay increase. Basically the cousin to the title inflation that took off hugely over the last decade.
Every org I've come in to has had dozens of these instead of senior ICs. It creates an absolute mess in the org chart and is difficult to undo because it's viewed as a demotion. It's also a larger symptom of the absolute collapse of managerial ability in tech companies since the late nineties
They say a great programmer in Silicon Valley can be 10x productive as a good programmer, or a great manager can add 10x value as a good manager. In that environment, and when you are cash heavy, it makes sense to cast a wide net and hoard talent. Then you try and coach them up for a few years and see which ones worked out.
Arnold poses some great questions that industrial organization ("IO") economists have grappled with for decades. Ronald Coase (Nobel Prize 1991) and Oliver Williamson (2009 prize) attempted to explain why firms (corporations) exist at all and why and where the boundaries lie between companies on the one hand and external contracting on the other.
Arnold's blog earlier this week about "producers and parasites" also informs the answers to these questions. Large organizations in both the public and private sectors employ both producers and parasites. The latter tend to be very good at "office politics" and making themselves appear to be useful without actually being useful. But they have power because of their political skills. Organizations that have surplus to burn can afford to carry "parasites" for a while, until economic or competitive pressures force upper management to pay attention. I'd suggest that they pay more attention during economic downturns or when a disruptive actor appears - think Tesla versus the big car companies. If due to widespread economic shocks (recessions), then there would be a pattern of trimming the fat concurrently; but otherwise, as Arnold suggests, the phenomenon should be more randomly distributed across the economy.
The assumption of most observers of corporate restructuring is that cuts are made in a logical fashion. My experience working with executives is very different. One problem is that executives often make the mistake of promoting people who are excellent at domain skills (accounting for accountants, sales for sales people), but who do not have strengths to be leaders. The reason for the promotion is that company compensation plans assume that leadership and decision making, even at the lowest level, are more valuable than critical domain skills, which is often untrue.
A second observation: executives are managing a complex array of outcomes: corporate strategy, technology investment, succession planning, organizational structure constraints. Some executives are simply not adept of negotiating among these competing concepts and they layoff people periodically as a substitute for weaknesses in one or more.
I’m short on time today since I commented on Dan Williams’ post. I will just add the following regarding this: “What is the relationship between Type I errors (cutting too much and regretting it) and Type II errors (letting costs rise too much and regretting it)?” I have trouble remembering what Type I and Type II are partly because their names convey little meaning. Can we take the words in parentheses and come up with better names for them? Suggestions?