I've presented to Bezos once. It went about the same as Yegge's experience, except I didn't have his balls to laugh in my CEO's face ;)
The whole thing felt like I was meeting the President - layer after layer of aides and executive assistants, tons of security, and briefings on where to sit, where to not sit, where to stand, where to not stand, look him in the eye when you shake his hand, etc etc. It was kind of surreal.
I haven't presented to any other big-corp CEO other than him, so I can't really draw any comparisons.
One thing Yegge brought up that I also felt the same about walking out of that room: Jeff is really smart, like really, really smart. I was warned about this beforehand and spent two weeks poring over every single detail, potential feature, everything about the idea I was presenting. He still managed to come out of left field with relevant questions that none of us had even thought of. So I can corroborate that side of the story.
The comparison to Liszt also seems apt - before we were even done presenting, he'd already grokked it to a surprising degree, and was already expanding upon the idea in out loud. You could hear my manager scribbling like a madman trying to get all of this down. This might not seem especially amazing, but the Director and VP level people I presented the very same idea to had not the same depth of insight. I remember being very impressed on the spot.
I'm no longer with Amazon, and I don't particularly want to go back - but that's not due to bad top management, IMO Bezos is one of the keenest tech CEOs around. He might be the closest thing we have left to Jobs.
To balance it out a bit and not make this seems like a complete Amazon love-fest: Jeff loved our idea, he gave us the green light right away. We left that room ecstatic thinking we were about to change the online retail experience forever.
Then middle management showed up. There was political infighting about who owns the project - it straddled multiple disciplines (hence why it was so groundbreaking), and there was a mixture of both hot-potato-oh-god-you-take-it and this-is-amazing-we-need-the-credit-on-this.
The teams that were interested were unwilling to yield to other teams that (rightfully, given their expertise and domain) wanted in, and some teams we needed support from kept punting it since it wasn't in their yearly plan (put together, well, a year or more ago). The project would have provided very powerfully tangible, very high-profile benefits to the customer, but said benefits weren't part of the metrics on which our department was getting judged, so at the VP level the willingness to devote resources was almost non-existent. There was lots of lip service given - especially about how a few people were able to hack together this thing and make it all the way to a Jeff Presentation.
But ultimately the project froze. It was, actually, probably the main reason I decided to leave Amazon for a smaller, more agile startup, where if the CEO wants something done, by golly, it's gettin' done.
That's a good illustration of no matter how brilliant the leader is, a bad team under them and/or poor execution can lead to failure.
Even with this infighting and whatnot going on Amazon is still doing amazing things. I'd be interesting to see what it'd look like if they were able to execute everything well.
It seems that's also a crucial role of the exec - if they see something that needs to be happening and it's not, they need to get in there make the necessary changes.
There are probably a million different ways for it to go bad.
There is a theory that gets posted here from time to time that As hire As and a few Bs that get through the filter, Bs hire Bs and a few Cs get through the filter, and so on down the line until you have hired people you really shouldn't have.
The other theory is promote until incompetent.
My guess is much more mundane, when you hire middle management they act like middle management. Which works as long as Their sphere of responsibility and control is well defined. When you need to cross cut concerns middle management actively fights against it. All most like using inheritance for code reuse and having deep inheritance hierarchies.
Since classifying people appears to be so trivial it can be boiled down to one character, does being bothered by the use of apostrophes in pluralization make me an A, a B, or a C?
Even the best managers are led by the incentives. If you want a project to succeed, make sure incentives are lining up. If you have a project outside the incentive structure, plan on it failing.
Indeed. The whole job of a manager is to keep whatever metrics are used looking good for his department and/or team. If your metrics don't correlate well to a given project, it will be almost impossible for it to work unless you have a command structure that's aware of this potential dissonance and can send down the edicts necessary to fix it for a given situation.
Middle management is really dangerous for this reason. Managers effectively are stuck with blinders; their one and only job, and often the only thing they know how to do (frequently through surreptitious or less-than-stellar means), is to keep their team and/or department sitting pretty in terms of normal company metrics. This correlates both to the resource allocation given a team and the rewards allocated specifically to the individuals that comprise that teams (including the manager, of course), which is why people appreciate "good" managers, until something unanticipated comes up and is killed by this rigid process.
Because at some point, the team gets so big, it's harder to weed out bad managers.
Additionally, managers are human, and will make mistakes. Often it's like the frog-in-the-pot story--you don't notice the little mistakes, only the sum of their results.
Honest question: can a company manage $34.2 billion a year in sales with a small management team?
I suppose the answer is "yes," given that MS seemed to manage with a minimal number of middle managers, at least based on the Joel Spolsky story shared elsewhere in this thread.
I think a good skill for a big company CEO would be doing debugging to see what is exactly happening here. If Bezos did that, they would have noticed what was happening and fixed the problem.
The whole thing felt like I was meeting the President - layer after layer of aides and executive assistants, tons of security, and briefings on where to sit, where to not sit, where to stand, where to not stand, look him in the eye when you shake his hand, etc etc. It was kind of surreal.
I haven't presented to any other big-corp CEO other than him, so I can't really draw any comparisons.
One thing Yegge brought up that I also felt the same about walking out of that room: Jeff is really smart, like really, really smart. I was warned about this beforehand and spent two weeks poring over every single detail, potential feature, everything about the idea I was presenting. He still managed to come out of left field with relevant questions that none of us had even thought of. So I can corroborate that side of the story.
The comparison to Liszt also seems apt - before we were even done presenting, he'd already grokked it to a surprising degree, and was already expanding upon the idea in out loud. You could hear my manager scribbling like a madman trying to get all of this down. This might not seem especially amazing, but the Director and VP level people I presented the very same idea to had not the same depth of insight. I remember being very impressed on the spot.
I'm no longer with Amazon, and I don't particularly want to go back - but that's not due to bad top management, IMO Bezos is one of the keenest tech CEOs around. He might be the closest thing we have left to Jobs.
To balance it out a bit and not make this seems like a complete Amazon love-fest: Jeff loved our idea, he gave us the green light right away. We left that room ecstatic thinking we were about to change the online retail experience forever.
Then middle management showed up. There was political infighting about who owns the project - it straddled multiple disciplines (hence why it was so groundbreaking), and there was a mixture of both hot-potato-oh-god-you-take-it and this-is-amazing-we-need-the-credit-on-this.
The teams that were interested were unwilling to yield to other teams that (rightfully, given their expertise and domain) wanted in, and some teams we needed support from kept punting it since it wasn't in their yearly plan (put together, well, a year or more ago). The project would have provided very powerfully tangible, very high-profile benefits to the customer, but said benefits weren't part of the metrics on which our department was getting judged, so at the VP level the willingness to devote resources was almost non-existent. There was lots of lip service given - especially about how a few people were able to hack together this thing and make it all the way to a Jeff Presentation.
But ultimately the project froze. It was, actually, probably the main reason I decided to leave Amazon for a smaller, more agile startup, where if the CEO wants something done, by golly, it's gettin' done.