Why Incompetence Spreads Through Big Organisations • Hacker News Board

Re: Why Incompetence Spreads Through Big Organisations

#21

Post by sethg » Mon, Jul 06, 2009, 6:05 PM UTC

Retric wrote: post #13

gjm11 wrote: post #7

This is about a (rather bad, I think) paper investigating a really extreme version of the Peter principle: what happens if competence at one level is totally uncorrelated with competence at the level below? In that situation, duh, it turns out that you want to promote the worst-performing people. (Because they'll do just as well after promotion as the best-performing people, by hypothesis, but promoting them out of t…

I think the basic problem has more to do with the inability to verify how competent someone is. In the real world I suspect the ability to fake competence is often what is promoted to the highest levels. But, let's assume each level is independent on every other level and higher levels are also more important. Now how do you promote people? Well find a general measure of how good people are at anything let's call it…

...and after a few iterations of reward, the next level of incentive has to come in the form of "promote to higher level in the hierarchy" rather than simply "increase salary", because of the bureaucratic norm that a manager can't be paid less than any of his or her underlings.

Re: Why Incompetence Spreads Through Big Organisations

#22

Post by gjm11 » Mon, Jul 06, 2009, 6:28 PM UTC

troels wrote: post #19

gjm11 wrote: post #7

This is about a (rather bad, I think) paper investigating a really extreme version of the Peter principle: what happens if competence at one level is totally uncorrelated with competence at the level below? In that situation, duh, it turns out that you want to promote the worst-performing people. (Because they'll do just as well after promotion as the best-performing people, by hypothesis, but promoting them out of t…

That was more or less my thoughts as well, after reading this post. Did you read the actual paper? I would be interested in knowing a little bit more about which method they used to simulate the situation.

Yes, I read the paper. (Not, I confess, every word, but enough to be pretty sure I'm not misrepresenting it.) It's linked right at the end of the article. http://arxiv.org/abs/0907.0455

Re: Why Incompetence Spreads Through Big Organisations

#23

Post by rbanffy » Mon, Jul 06, 2009, 6:55 PM UTC

gjm11 wrote: post #7

This is about a (rather bad, I think) paper investigating a really extreme version of the Peter principle: what happens if competence at one level is totally uncorrelated with competence at the level below? In that situation, duh, it turns out that you want to promote the worst-performing people. (Because they'll do just as well after promotion as the best-performing people, by hypothesis, but promoting them out of t…

"I don't see much value in what they've done."

I think what they did is absolutely hilarious. And, since they admit their two extreme scenarios are just that, extreme, it's even good science.

I also think their model can be easily tweaked for middle-ground scenarios and vertical correlation gradients (where abilities correlate differently between layers).

I guess I have a new weekend project.

Re: Why Incompetence Spreads Through Big Organisations

#24

Post by rbanffy » Mon, Jul 06, 2009, 6:56 PM UTC

thisduck wrote: post #3

How about training people in the skills they will require for their new position?

Training can change the wiring of someone's brain as much as swimming classes can turn people into fish.

Re: Why Incompetence Spreads Through Big Organisations

#25

Post by eru » Mon, Jul 06, 2009, 7:01 PM UTC

Retric wrote: post #13

gjm11 wrote: post #7

This is about a (rather bad, I think) paper investigating a really extreme version of the Peter principle: what happens if competence at one level is totally uncorrelated with competence at the level below? In that situation, duh, it turns out that you want to promote the worst-performing people. (Because they'll do just as well after promotion as the best-performing people, by hypothesis, but promoting them out of t…

I think the basic problem has more to do with the inability to verify how competent someone is. In the real world I suspect the ability to fake competence is often what is promoted to the highest levels. But, let's assume each level is independent on every other level and higher levels are also more important. Now how do you promote people? Well find a general measure of how good people are at anything let's call it…

And paying top managers top wages, let's you get away with lower wages at the lower levels. As long as you can keep the dream of getting to the top alive.

Re: Why Incompetence Spreads Through Big Organisations

#26

Post by udekaf » Tue, Jul 07, 2009, 1:30 PM UTC

kareemm wrote: post #20

udekaf wrote: post #14

Earlier quoted context omitted.

Most of organizations in our society have hierarchical structure. The goverment, military, academic, enterprise show you the hierarchical traits in position titles. The pyramid is necessary for managing responsibilities and decision making. Social stratification is normal to all societies to some extend.

> The pyramid is necessary for managing responsibilities and decision making. It's actually one way of making decisions and managing responsibilities. It's likely the most common because it evolved from the manufacturing-era model (think Henry Ford making cars). Ford's model evolved from looking at the military. Ford has a famous line - “Why is it every time I ask for a pair of hands, they come with a brain attached?…

No doubts that open source model creates vast value to software industry, but that model is distributed only in terms of the big picture. If you look at a specific open source project, it's probably still organized in a hierarchical way. It still needs a leader to do the high level design and decide direction. New commers still need to climb the ladder - by contributing to the project - to a point before they can significantly affect the project.