Wednesday, May 23, 2007

Does IT matter?

There are two ways of interpreting the question: "Does IT matter?" These could roughly be labelled macroeconomics and microeconomics.

The macroeconomic question is about the global aggregate impact of IT on business or society.
  • Does IT (as a whole) make business (as a whole) more profitable?
  • Is there any evidence that businesses with more IT are more profitable than businesses with less IT? (If wealthier companies spend more on IT, does that just show they have more money to spend?)
  • If IT has any demonstrable impact at all, is this merely a short-term sugar rush (aka turbulence), or are there enduring long-term consequences?
See for example Andrew McAfee: The Shocking Processes of IT Impact.

The microeconomic question is about the extent to which differentiation within IT matters.
  • Does one kind of IT produce significantly better results than another kind of IT?
  • To what extent do the technical issues to which IT people pay most attention have any real economic significance?
For example, if two alternative technical solutions, competently executed and managed, both produce around 20% ROI, then the difference between them doesn't matter.

As some aspects of IT are increasingly turned into utility computing, then maybe some of the big questions of IT become small questions, and some of the small questions become irrelevant. At least from an external business perspective.
  • Does Open Source matter?
  • Does XML matter?
  • Does architecture matter?
  • ... and so on ...
From an IT perspective, these kinds of questions are of course very interesting and important. But this is sometimes hard to demonstrate to an external audience - including business people and politicians. They want IT, or at least they want the benefits of IT, but they don't want to think about IT. Should IT matter in this sense?

To whom does IT matter? Does the vast amount of technical knowledge accumulated within the IT industry matter to anyone outside IT? Does IT matter?

No comments: