Joe's Jottings

Jottings Number 50, by Joe Podolsky:

From: uunet!HP-PaloAlto-om4.om.hp.com!JOE_PODOLSKY

Date: Tue, 27 Feb 96 12:22:00 -0800

Subject: Middlepeople R Us

One of the better thinkers and advisors on information systems
and technology subjects is University of Texas professor and
former Ernst & Young consultant  Tom Davenport.  And I'm not
saying this just because Tom is a frequent visitor at HP and is
also kind enough to remain on the distribution list of these
jottings.  The real reason I like Tom's work is that he has
interesting viewpoints.  He often jogs my loose thoughts into
new patterns.

The current case in point is his column titled "Think Tank" in
the February 1, 1996 issue of _CIO_ magazine.  In it, he
introduces us to "Penzias' Law" named after Arno Penzias, one of
the "more visible smart guys" at Bell Labs.

Penzias' Law says "that anyone who stands between a customer and
a computer that can fully meet the customer's need, will
eventually not have a job."  People and organizations like that
are called "intermediaries" and Davenport says that Penzias Law
predicts "disintermediation" by IT.

Davenport explains that we are surrounded by intermediaries:
travel agents, brokers of all kinds, wholesalers, and
distributors.  In the computer industry, we have a whole
alphabet of intermediaries, VABs and VARs and ISVs, all of whom
in some way take hardware and add combinations of software and
services to give customers exactly what they want.

Disintermediation by IT is not a new phenomenon.  Wal-Mart
changed the rules of the retail game by using IT to work
directly with manufacturers.  E-mail allows us to by-pass the
post office.  Layers of hierarchy have melted away as strategic
planners and operational managers learned to use IT to
communicate without the interpretation services of middle
managers.

Davenport predicts, however, that the rate of disintermediation
is about to skyrocket thanks to the killer app of the '90s, the
Worldwide Web.  The Web is the closest thing we have now to a
completely open system.  We can be male or female or (as the New
Yorker cartoon says) even canine on the Web.  More amazing than
shape-shifting, we can be Wintel or Apple or even UNIX with
satisfying indifference.  And, as the explosion of home pages
demonstrates, Web technology is relatively cheap and easy to
use, both as providers and users.

I certainly agree with Davenport's observations, but I think
that he's just describing the tip of a management of change
iceberg.  In IT, I think "middlepeople R us."

Let's look at a typical personal computer distribution value
chain where there is a producer of computer hardware, an VAR
that adds custom installation and support services, and an
end-user. Penzias' Law correctly predicts that, as a hardware
family evolves, the intermediaries' services become less and
less needed.  The hardware and systems software become easier to
use, and the users become better educated and/or more tolerant,
squeezing the user support intermediary from both sides. 
Computer-Based-Training and context-sensitive help features
enable software producers to squeeze the traditional class-based
training intermediary.

A major driver of Penzias' Law, of course, is Moore's Law.  As
we get lots more MIPS for relatively constant costs, we can use
those instruction cycles to do work that once required people. 
Thus, all of us working in IT, no matter what our jobs, are
affected by Penzias' Law.

I suggest, also, that Penzias' Law is rooted as much in
economics as in technology.  When a technology is young, the
producer is happy to have various intermediaries to help provide
what Geoff Moore calls "the whole product."  And the value of
the new technology is sufficiently high to the users that they
are willing to pay prices that give adequate profits to all
players.  As a given branch on a technology tree matures,
however, competition quickly cuts the profit margins available
in the value chain, and the first limbs (but not the last) to be
pruned are the intermediaries. 

What I infer from this Penzias' Law lifecycle, however, is not
that intermediaries are going to disappear; rather, in order to
survive, intermediaries need to adapt and evolve themselves even
faster than are the producers and users.

The evolution can be linear; e.g., user support people need to
let go of DOS-based Windows and become experts in NT.  C
programmers learn how to (re)use objects.

This is the easiest case, because skill sets are relatively
constant even as specific content varies.  Experienced classroom
trainers, however, are not likely to have the software skills
needed to write computer-based-training packages.  Even the 
best coders may not be good at cobbling together business
solutions from purchased applications.  To survive these kinds
of shifts requires a whole new set of aptitudes.  Some trainers
will evolve to course planners, some programmers to systems
designers, but, even then,  to drive down costs, fewer of these
jobs will be needed for a given business set.

We in IT have foisted the effects Penzias' Law on others for
years, but we've been able to keep our own intermediary jobs by
linear evolution. 

No more.  Dr. Frankenstein's fate might be ours.  We have built
the Web and other user friendly systems, and our users don't
need us as much any more.

Do you resemble this?  If so, what are you doing to survive? 
Or, do you think you're in a job that's immune from Penzias'
Law?  In either case, please share what you've learned.

Penzias' Law alive and well and living in our IT career paths. 
Thanks, Tom, for giving us a label and a wake-up whack.  The
first step in dealing with a problem is to recognize it.



Joe 



Back to Joe's Jottings