Newsgroups: sci.space.history
From: henry@spsystems.net (Henry Spencer)
Subject: Re: technology development cycles
Date: Fri, 20 Nov 1998 02:48:05 GMT

In article <nAW42.1642$4%.23702718@fozzy.nit.gwu.edu>,
Dwayne Allen Day  <wayneday@gwis2.circ.gwu.edu> wrote:
>To the best of my knowledge (and I've looked fairly hard) there are few
>studies of "skunk works" organizations in general.  It is an untested
>argument that they are amazingly efficient.  But one might note that
>recent attempts by other firms to emulate them have not been very
>successful.

I found one bit of Ben Rich's autobiography ("Skunk Works") noteworthy in
this regard.  Rich has just been offered the chance to head a Skunk Works
clone for Northrop, he's awfully tempted, and he's talking to Kelly
Johnson about it.  Johnson's response is that he flatly doesn't believe
that Northrop will *really* do it, and he identifies several reasons (some
of this, mind you, is my own analysis based on what's in the book):

(a) Most upper managements aren't brave enough to keep their hands off,
to accept that a significant part of their business will be done by a
semi-independent operation that they aren't allowed to know much about
and can't interfere with.  (Despite the Lockheed Skunk Works's success,
some of Lockheed's own management deeply resented its existence.)

(b) Low manpower and efficient operations mean low profits on traditional
government contracts, which reward effort rather than results.

(c) Small groups with high personal responsibility mean little opportunity
for raises and promotions, especially in a traditional organization that
wants to reward managers based on number of people supervised rather than
on results achieved and responsibility taken.

(d) Starting a Skunk Works out in a blaze of publicity, as the official
leading edge of the company, guarantees lots of internal enemies and a
complete lack of forgiveness for any mistakes made... and there will be
some, especially since such a startoff guarantees that the Skunk Works
will be given only the projects which are too difficult and too crucial to
the company's future to entrust them to the orthodox organization.

(e) The Skunk Works sort of model, where the Skunk Works has considerable
freedom to innovate and come up with its own solution to the customer's
problems, doesn't work well with customers who want total control... as
the USAF increasingly did, which is why the Skunk Works never built
another fighter after the F-104 [the F-117, despite being called a
fighter, is a light bomber].
--
Mass-market software technology has |  Henry Spencer   henry@spsystems.net
been deteriorating, not improving.  |      (aka henry@zoo.toronto.edu)