Subject: job rant 
Tue Aug 11 17:46:43 1998



> Hmmm.... I thought we were supposed to depend on Mike Stone for humor,
> satire, and ranting....
> 
> Suz


whoo.. nice shot.   you may consider me officially zinged.   at
least i know you took me seriously about pushing me to write, tho'.. 
;-)



okay, here's a rant, hot off the press:



when Adam Smith wrote _The Wealth of Nations_, he said that one of the
key features in a healthy economy is specialization of work.  
economists have taken that idea and run with it, making all sorts of
graphs about companies A and B, who compete in the widget market.   a
widget is composed of a dingus and a doohickey, so if A and B compete,
each company has to build both components.   the problem with trying to
generalize, according to the theory, is that A (or B) can only improve
his dingus-making at the expense of his doohickey-making.   if they make
one a bit better, the other gets a bit worse.   in the long run, this
leads to a world where both A and B compete by selling widgets with
mediocre components, because that's the best either can do on their own.


Smith's principle of specialization says that if A concentrates on
making a really good dingus, and B concentrates on making a really good
doohickey, they can collectively produce a widget which is far superior
to what either could make on their own. this leads to better use of
resources, higher profits, and lower costs.   everyone benefits, and the
world is a happy place.


one of the things i've discovered in the last six months is that this
theory is wrong.   i'm not going to challenge the idea of
specialization, because that does work.   it's just that Smith's idea of
*why* it works is straight out of _The Big Book of Fairy Tales_.   the
*real* reason why specialization works has nothing at all to do with
resources or making a better dingus, it's about communication.


based on my observations, the average company has ten people who
couldn't pour sand out of a boot for every one who can.   in the kind of
utopia Smith dreamed up, that small core of competent workers would make
good dinguses, and the company would train the remaining deadwood up to
that same level.   the way things really work, it's easier for the
company to have the one competent worker act as a remote-control brain
for the other ten, and produce mediocre widgets in quantity.   this is
in fact necessary, because the marketing department has already committed
to ship, in bulk, even though the designs aren't complete and the
production line hasn't been built, yet.

in that kind of environment, specialization isn't an economic theory,
it's a survival skill.   being specialized means you only have to deal
with nitwits who want to debate the nature of the universe as it relates
to dingus-making.   if you generalize (heaven forbid), and become good
at making doohickeys too, you've effectively mortgaged off that much
more of your life.   if you're insane enough to pick up a third field as
well, you have to choose between carrying gun and giving up bodily
functions during business hours, weekends, and the occasional 3am when
somebody decides there's some kind of crisis.

the reality of the corporate workplace is that quality levels are
decided by managers who refuse to have anything to do with the details
of production (they specialize in "the big picture").   a higher level
of competence among the production staff just means they can afford to
make more mediocre products.   that in turn means it's time to "grow in
the marketplace", which means hiring more incompetents to drag the
system back down to a point where it's in balance.


the lesson for freelancers in all this is that an operation with ten
competent people can realistically kick the pants off a company ten
times its size.   the two areas where you have to concentrate are
quality and communication.   quality is actually the easy part, because
anyone who's honestly competent will insist on doing work they can be
proud of.   the communication thing is the secret that nobody ever
mentions.   the higher your signal (defined as information relevant to
getting the project done correctly) to noise (defined as everything
else) ratio is, the fewer people you need to kick some serious
corporate butt.

to succeed in that kind of a setting, you need generalists with fields
of emphasis.   tight specialization wastes time, because you can't
cooperate if you don't understand what the other guy is doing.   if you
already have a decent background in the other guy's field, you can get
on with collaborating and not have to waste time on clueless-newbie type
questions.   that difference in time is precisely what you can use to
knock a larger and less intelligent competitor straight out of the
market.