Prev: Boot Block of NOR flash memories ??
Next: Benchmarking a toy example on SH-4
From: larwe on 19 Apr 2010 10:34
On Apr 19, 8:06 am, Frnak McKenney
<fr...(a)far.from.the.madding.crowd.com> wrote:

> Thanks for the reminder, though as I recall you do need a period of
> regular rewards befoe you start going random. In any case, I wonder
> if this applies to a human being making conscious decisions and with
> the full knowledge that the universe is imperfect, that other people

Rationally, I can calculate the odds of winning a lottery or coming
out ahead when I sit down in front of a slot machine. But I still
enjoy going to casinos once in a while :) As far as guiding one's
actions goes, I think people ascribe far too much meaning to
intelligence and nowhere near enough to the "eat, fight, mate, avoid
being decapitated by saber-toothed tiger" aspect of human motivation.
People think it's maybe 75/25. I think it's more like 20/80, and half
of that 20 is just thinking up "logical" rationalizations for why they
went with their instincts.

There's a good body of research to bear this out.

In summary: despite all your rage, you are still just a rat in a cage.
Run, squeak, press bar, get pellet.

> describing here, which do you think would have a stronger
> "deconditioning" effect: a year of bad sales (inconsistent
> environment), or a last-minute cap on the size of a bonus because
> the company didn't want one salesman "making too much money"
> (inconsistent policy)?

Having seen both of these happen, I would say the latter, but
apparently not for the reason you think: the reason seems to be that
the bad sales were predictable (you can see the bucket of money
filling up over the year, and you can know way ahead of time if it's
not going to fill up to the quota mark), whereas a capricious decision
at the last moment is more of a shock.

> > Personal Improvement Plan, i.e. death row. PIP is the term used by
> > most BigCorps for what happens to you between "pissed your boss off
>
> Ah. Thanks. Where would we be without euphemisms? <grin!>

Oh, it gets better. For example, my BigCorp has a field at the end of
the annual performance review where a standard phrase has to be
selected; these are, for example, words to the effect of "promote this
guy ASAP" or "develop in place". The code for "get rid of this one"
used to be "Outplacement Opportunity". It has now been revised even
more humorously as "Expatriate Assignment", which conjures up visions
of my convict ancestors being shipped off to Botany Bay.

> > I'm saying that the organizational overcomplexity that leads to the
> > formalization of such plans is a symptom that the structure is already
> > so complicated that nobody's activities are directly linked (by a
> > simple formula) to meaningful company goals, and so bonus levels
>
> Ah (again). It seems to be the range of the phrase "such plans"
> that I'm getting hung up on. My own experience is mostly limited to
> one small (under 100) company and consulting; while I've seen
> incentive plans with problems, I have to be cautious about

I have more experience with the BigCorp side of these things. At
smaller companies, the effort-to-results links are stronger and more
visible, so the problem might not even exist there (or at least if it
does exist it's easy to see if one looks). At BigCorps, you can take
it as a fairly universal truth that the machine is so complex you can
only measure it by "unit testing", so to speak - you can only measure
the direct output of your snippet of the process, and there is
[intentionally!] no equation that will let you understand how that
affects the whole.

> > The only programmed incentive plans that really make sense in an R&D
> > type organization are project-linked and cross-functional. I.e. the
>
> But isn't even this structure subject to your concern regarding
> "gaming the system"? Doesn't it offer a perverse incentive for a

Absolutely. It's like democracy - it's been proven not to work, but
it's less unworkable than the next most palatable alternative.

> It gets tricky on the "upper management" side, too. If one forces a
> product manager to accept what that manager considers unrealistic
> goals, won't that be a disincentive for the entire team? (Which is

It is - again, personal experience speaking there.

> And then there's the confusion generated by Reality(tm) in the form

Don't get me started. I have worked for companies (making everything
from bags of wood screws to jet aircraft components) where a Golden
Standard Project Length has been set - say 300 days. Doesn't matter if
you're making a moon lander from scratch or redesigning the packaging
of a tub of margarine - you shall take no more than 300 days. And the
number is reduced every year, too.

The way this lunatic sort of system gets gamed is that projects aren't
officially started until they're nearly complete; they're kept as
skunkworks projects until they're close enough to completion that they
MIGHT meet that arbitrary deadline. Of course that means the project
didn't get real resources from other departments, and there was no
official spec for it, so the project just kind of grows.

> out, or if given, they should be based on a manager's opinion)? Or
> just difficult and not likely to be solved in a one-size-fits-all
> fashion?

In general, I am terribly suspicious of one-size-fits-all policies,
because I think every single one I've come across has caused logical
insanities. I'm not a big fan of Kant's philosophies, either - more a
J.S. Mill guy - for the same reason.

> > product, or Engineering does some unusual work to fix a production
> > problem or value-engineer the design, there are ad-hoc incentive
> > programs in place to reward that sort of thing.
>
> "Incentive programs" but not "bonuses"? That is, non-financial
> rewards?

Jargon issue here. "Bonus" has a specific technical meaning, which is
why for instance when I get a cash payout for filing a patent, it's
listed on my paycheck as "technical award", not "bonus". "Incentive
programs" are not "bonus", which is an annual payment considered part
of one's compensation package as opposed to incentives, which are one-
time events. I was in fact referring to cash incentives there.

> D**n. I _knew_ that those silly Calculus and Physics clases were a
> waste of time; I should have been spending my time on Underwater

Yeah, think how I feel working a fulltime week and sitting in five
classes (15 credits) of control systems theory and Fourier transforms,
never sleeping, always in exam mode...

> > The exact same sequence of events leads to stultification of an entire
> > company by means of process. The process becomes the product, and any
>
> But do you see this as inevitable? Or do you think that the
> symptoms can be recognized before the "Medusa effect" becomes

It is somewhat avoidable, but only if guided from the very top of the
organization by someone who understands the danger and is competent to
evaluate (and explain to shareholders) why it doesn't make long-term
financial sense to the company to use process tools in R&D roles.
Google is an example of a company that's doing it right so far,
although it's a bit young to be sure how long that's going to last.

Also, for any large diversified company, it must be recognized that a
single process often cannot sensibly be applied to every branch of the
company. If it was possible to make a Grand Unified Process that works
for any branch of a diversified company, then by the same logic it
would be possible to make a Grand Unified Process that would allow any
bicycle workshop to be an aircraft manufacturer, or vice versa. Part
of the difficulty at some of these ossified BigCorps is that military/
aerospace processes are being applied across the board to (for
example) disposable consumer electronics products. For some spaces, ad
hoc "processes" are in fact the ideal. Google's 20% time perk is a
great example of how to avoid this kind of trap.
First  |  Prev  | 
Pages: 1 2 3 4 5 6 7
Prev: Boot Block of NOR flash memories ??
Next: Benchmarking a toy example on SH-4