In these tough economic times, green initiatives can be a hard sell. Companies don’t want to take a gamble on pricey projects that lie outside their core mission. Yet lots of eco-friendly ideas promise to pay for themselves—and then some—by slashing costs and boosting efficiency.

How should companies approach the problem? To find out, we looked at green initiatives in one critical section of businesses, the corporate data center, and placed potential projects into four categories. At one end of the spectrum are obviously useful ideas that are simple and inexpensive. At the other end are expensive distractions that should be avoided at all costs. By figuring out which category an idea fits into, companies can better weigh the risk and potential return.

One caveat. This system—based on an earlier model developed in collaboration with Prof. Leslie Willcocks from the London School of Economics—relies heavily on the judgment of a company’s chief information officer. We assume the CIO is closely monitoring promising technologies and can evaluate their possible impact on the business.

Here are the four categories.

No-Brainers. In these cases, the green technology is a commodity. It not only cuts power use and emissions—thereby fulfilling its green mission—it’s easy and cheap to obtain and implement. The bottom line: Companies should pursue these projects as soon as possible.

Some examples include virtualization, which lets one computer do the job of many; advanced cooling techniques that save power in data centers; and software that identifies when a machine is inactive and powers it down efficiently and securely.

Promising but Pricey. Here, the green technology is clearly useful but isn’t yet popular enough to be a commodity. So, CIOs face a gamble. If they adopt the technology before it becomes the Next Big Thing, they might cut their costs and boost their efficiency before everyone else does—which could give them a big edge on rivals. But they’ll pay more upfront than if they waited, and there’s always the risk that the technology proves to be a flop in the long run. And that could leave them stuck with systems that aren’t supported by vendors and may need to be replaced eventually.