The Fuzzier Side of ROI. Good discussion by David Joachim in Computerworld regarding the difficulty in measuring the ROI of some IT investments. In my view, some projects such as those undertaken for productivity improvements need a strong ROI. But other projects, such as infrastructure projects, should be justified in terms of the flexibility they give the organization to undertake future initiatives. Understanding the difference between foundational investments, transactional systems, informational systems, and strategic investments makes the distinction clear. For more on this matter, see "Leveraging the New Infrastructure" by Weill and Broadbent.