My view is that the term ROI is often used very loosely as a synonym for "positive benefit". In other words, a significant percentage of organizations lack a formal ROI calculation process, and it largely comes down to, "if I spend $xxxx on this investment, can I get this other deal for $yyyy or save $zzzz over some not-clearly-defined period of time?"

Cloud computing can provide cost savings under certain circumstances, depending on what your alternatives are for the specific functionality or capabilities in question. More often, however, what it provides is better flexibility, the ability to manage costs more effectively by starting out small, or the ability to more easily reduce services and costs should demand go down. This is very true of IaaS.

SaaS, on the other hand, can often afford an organization capabilities and tools that they would otherwise have to spend a significant amount of money for up-front. Getting 2-4 people from your firm to use a SaaS-based CRM or ERP or Financial Reporting product will take less time and cost less money than building it out yourself -- especially if you intend to grow the number of users significantly over time.

Likewise, the cost of spinning up 5 servers to play with for a few months, and these ramping that up to 30 or 40 servers for a year and then ramping that back down to 10 is much more feasible in the cloud vs a traditional hosting arrangement (or on-premise arrangement).

The key for deriving cloud benefit is to evaluate the costs of your computing needs in one or two traditional approaches, and then contrast that with what the cloud option would afford you in terms of reduced time for provisioning, lower startup costs, ease of matching costs to variable volume, etc.

The metrics you use will be determined by the type of project and the needs of your particular business.

-ASB: http://XeeMe.com/AndrewBaker