The term ROI is the most misled and misused term when evaluating the true potential of a given technology. Vendors use ROI as the "silver" bullet to prove how good the investment is if made in their offering, CIOs use it to look smart in front of the vendors, Corporate Finance uses yet another method to calculate the "true financial impact" of technology investments. The main issue is ROI is a point in time metric i.e. it only tells us the Net Gains from an Investment. Let us take an example to illustrate how this works.
Example: We need to invest $500K in some new piece of software that automates the billing department. By investing in this we can free up 2 billing clerks who are annually paid $50K each. The cost of training the other people in the department is $50K and is a one time expense. The annual maintenance on the software (charged from year 2) is 10% of the price
ROI (Year 1) = ( ( 2*100) - (500+50) ) / (500+50) = -64%
This is misleading as from the second year, there is a Net Savings of (2 * 100)- (10% of 500K), which $150K on an ongoing basis. Hence, by just looking at the ROI of the project, it does not look like something that we should invest in. However the NPV for a 5-year period (Net Present Value) of the project is $76K (assuming a 14% Cost of Capital), and an IRR (Internal Rate of Return) of 26%.
Hence, from a pure Corporate Finance 101 perspective, this is a project that we should embark if we look at the NPV and IRR, however this project is a "no-go" if we just look at the ROI.
To summarize, we need to look at stronger long-term metrics such as NPV and IRR rather than just the ROI of the project. It is important for IT organizations to see the impact of a Capital Investment on their long-term operating structure and to explore "value-creation" opportunities rather than look at a "point-of-time" measurement - Return on Investment (ROI)