> I became a manager because the organizational problems can be so frustrating and I wanted to fix them.
The biggest "carrot" held out to idealistic people wanting to become a manager is exactly this: the (fake) opportunity to "fix things". Only after you've become a manager does it become apparent that long-lasting problems are long-lasting for a reason and that none of the choices low-level managers are allowed to make can meaningfully move the needle. Sometimes they fall for it a second time, but being a manager of managers is often even worse since now you don't even have direct reports to effect change.
I've long wondered why we insist on modelling company structure as strict trees / hierarchies. Perhaps there are other classes of graphs which would be better suited in order to avoid such situations?
In any large company there is the official HR hierarchy and then the informal collaboration graph of how work actually gets done.
The former is necessary to coordinate coarse-grained decision making, policy and vision that needs to be unified across thousands of people, most of whom will never meet each other, but who should ideally be rowing in the same direction.
The latter is necessary for the armies of individual contributors to get their respective jobs done. Trying to document and formalize this ad-hoc network holistically is impossible because it's too complex to be understood by any one person. Attempting to do so would require a non-trivial time commitment from all the workers, which would actually take away from them, you know, getting work done.
It's tempting to look at an org chart and assume this represents how things work on the ground, but Conway's Law is more ironclad than it may appear at first glance. Don't confuse legibility for operational capability. If lower level folks did not understand their goals and improvise, then large corporations would be even more rigid and brittle than they already are. They would be utterly incapable of responding to changes in the marketplace and smaller firms would dominate.
The official hierarchy still represents decision power. I'm not suggesting trying to document all work relationships but wondering whether a non-hierarchical model of decision power would be better than a classic hierarchy.
The official hierarchy does represent decision power, but it is not the only source of decisions in a company. The vast majority of decisions in a company are taken on lower levels and either not passed through any hierarchy at all or only presented through the classical technique of presenting a list of options, all but one of which are unpalatable. In practice specialists prepare most decisions in advance (cloaked in boss-pleasing terms like "advice" or "RFC"), so the non-hierarchical model you speculate about is already there.
Is your take then that the classic official hierarchy is purposeless? That it is simply an obstacle? Or that it cannot be improved upon because it is already optimal for its purpose (whatever that may be)?
I don't think I said any of those. The official hierarchy is a neat compact description of formal lines in an organization, not less but also not more. At most I think any formal org chart represents a vast oversimplification of the intricate graph of relationships that exist between people in any company. There are things about a company it can describe well (like who is in charge of performance reviews for who), and things it cannot describe well (like the nebulous role of individual popularity on company decision-making, or in distinguishing between productivity differences between individuals on the same "level"). It is also almost always limited to the company itself and excludes any factors outside it like competitors or suppliers, despite those factors sometimes being more important to company decisions than its internal organization.
In short, my take is that the "classic" hierarchy is a useful but limited tool. It is not sufficient in the slightest to describe how a company makes decisions, yet too many people treat it as if it is all you need to know.
You didn't, and you didn't leave the impression of having done so; it's just me poking to understand.
> like who is in charge of performance reviews for who
These are the kinds of things I'm trying to challenge: Are performance reviews useful? Would performance reviews actually be more useful in some other structure than a classic hierarchy?
In other words, I'm not convinced beyond all reasonable doubt that the classic hierarchy is the optimal structure for the limited, but useful purpose you describe it having. Obviously the standard thinking is that it is.
> It is not sufficient in the slightest to describe how a company makes decisions, yet too many people treat it as if it is all you need to know.
I've lead a engineering department where we had something more akin to a matrix org. If a team wasn't doing well, "debugging" it was a disaster. I had to talk to several managers, PM, ICs and disentangle a mess of he said/she said feedback.
When tasked with data viz on these lines, one of the early things I ask is what the edge relationship is supposed to represent. Specifically!
Organization / chain of command? Parts information? Charge Codes? Messages/Sentiment? Business Information Systems capture way more data than mahogany row might realize, and you can get some "split the atom" visualizations when you combine the right parameters, like RnD funding + messaging. "Huh. Looks like new tech needs a LOT of communications with the field technicians. Like, a LOT a lot - totally wiping out bandwidth in remote locations"
The data is maybe there but if you model everything at once it's going to be a minimally-significant graphviz blob.
Stupid punchline? Execs hold up the blob as scientific proof that their job is hard. "Don't change it! It's so complex and pretty! I can show this to the VP-Manager of Goofball Systems Inc to validate my existence!"
No, it's not hard or pretty, you just can't tell the difference, conceptually, between a hex driver and a lathe.
“Execs”, a reference to the old fashioned setup of a corridor filled with executive offices that all have impressive mahogany desks. Feels like a British phrase to me but I’m not sure if it actually is.
Organizations do implement matrix structures to greater or lesser degrees which have their own advantages and disadvantages. Typically someone formally reports to one manager but will be "dotted line" into one or more additional people.
It's because there's ultimately one person responsible for company's performance (the CEO), so everybody has to report to him/her through a tree-like structure. A company is a very centralized structure, no different than the army.
"The CEO is ultimately responsible" is just the highest level of the carrot from my post a few levels up. A CEO may seem quite powerful from the inside of a company, but from their perspective they have to deal with competitors, shareholders, suppliers, regulators, and a host of other actors, all of whom have different objectives than the CEO and all of whom can constrain the possible actions of the CEO to varying degrees. Not to mention that many of the managers at or just below CXO-level are highly ambitious people who more likely than not have aspirations to become CEO themselves, so it may be in their interest to do some tactical backstabbing to make the current CEO look bad to the shareholders. All of this adds up to conditions where a CEO definitely cannot do whatever they want, because resources are limited even at this level. Just look at all the failed projects various CEOs at (say) Apple and Google have tried that didn't work even with all the money in the world.
(As a former military officer, this is the same for generals btw. They may have a lot of "power" in the organization itself, but they're heavily constrained by outside factors. They have to make do with the budget they're given, and have very little control over hiring targets etc. Not to mention that during wartime the enemy will not be under control either)
Yeah, of course. The CEO is judged based on how the company under his management is doing in its overall environment (vs what would be the baseline expectations), not just on the absolute numbers.
Perhaps there shouldn't ultimately be one single person responsible for the company's performance. I know the standard way things work, but that's not necessarily the best way.
It's even worse. Even if you made a difference you won't know for certain and you'll mostly hear from people who didn't like what you did. One of the clearest success you can get is avoiding worse shit that would have come to your org. That alone can have a massive impact, many times that of a average IC, but "avoiding worse shit" isn't entirely satisfying. It's just frustrating that it had to be avoided in the first place.
The biggest "carrot" held out to idealistic people wanting to become a manager is exactly this: the (fake) opportunity to "fix things". Only after you've become a manager does it become apparent that long-lasting problems are long-lasting for a reason and that none of the choices low-level managers are allowed to make can meaningfully move the needle. Sometimes they fall for it a second time, but being a manager of managers is often even worse since now you don't even have direct reports to effect change.