I am an information systems
strategy professor, so it would be expected that I would disagree
with Nicholas Carr's provocative assertion that IT doesn't matter.
Indeed, I do. While I agree with much of Carr's excellent-but
incomplete-analysis, I disagree with his conclusion.
Certain areas in IT have
become commoditized and continue to be commoditized. Just like
the phone system: A business user does not have to be a network
engineer to use it; the phone is a plug-and-play utility available
to anyone. The same is basically true for office-productivity
software and computer networks-although many would argue that
it is still much easier to plug in a new phone or fax machine
than it is to hook up a PC to the Internet at home or to share
The analogy with the phone
system breaks down at the point where Carr's analysis stops. Information
systems, and software applications in particular, differ in versatility
and adaptability. To exaggerate somewhat-but only a little-anything
is possible with software, if not today then tomorrow.
In fact, the phone system
will soon be an information system, with much of its value-added
provided by voice-over-IP software. Value-added is being shifted
from mechanical systems and their operations into software. This
is yet another example of the evolution of business automation.
The history of modern production is intimately tied to the automation
of business processes. First, companies used steam engines, then
conveyor belts, and today we use information systems, and especially
software, to automate business activities. We might call it "softwarization."
Companies in many industries now use ERP and CRM software to automate
business activities. And this softwarization is not a one-step
affair, like flipping a switch, but an ongoing process. Value-added
is constantly being shifted into or embedded in software, with
mature areas obviously becoming commoditized. Examples include
computerized antilock brakes, credit cards and calling cards,
airline ticketing, and yield-management systems.
Why would this process
stop? Why would there suddenly only be mature areas? Are there
not enough business activities left to be automated? Would it
be too difficult or expensive to automate the remaining ones?
The very commoditization of mature infrastructure technology reduces
unit cost, which in turn frees up funding for continued softwarization
without necessarily increasing total IT budgets.
Two trends ensure that
the sky is the limit for softwarization. One of them is well known,
while the other appears to be hidden. Carr mentioned the popular
one-Moore's Law, which establishes that hardware will become more
powerful and cheaper over time. Yet, even more important are advances
in how increased processing power can be used. This leads us into
the world of systems and software architecture design,
with its fast-growing jungle of acronyms and ideas. One key advance
in this field has been the recent breakthrough of object-oriented
programming. The concept, as well as some tools such as the Smalltalk
programming language, have been around for decades, but only very
recently have these concepts turned into commercially viable implementations.
The bottom line is that powerful hardware together with more flexible
software tools continues to fuel an ongoing process in which value-added
is increasingly achieved with information systems.
While mature areas do indeed
get commoditized and probably outsourced, new softwarization should
receive more, not less, top management attention. Why? As Michael
Porter argues, "[Business] activities are the basic unit
of competitive advantage." As these activities get automated
using software, top management's attention should shift to information
systems architecture design.
Chris Schlueter Langdon