Anyone who follows baseball knows that it would be silly to evaluate every player based on batting statistics. Shortstops, playing the toughest position, get a lot of slack when it comes to batting. First basemen, playing the easiest position, better hit a ton.

Yet businesses do the equivalent of basing everything on batting. Just about all of the hundreds of billions of dollars of annual capital investment decisions rely on projected return on investment (ROI).

ROI is useful in many cases, but it's too limited. If the Securities and Exchange Commission tells you that you have to file more information, more frequently, you don't calculate an ROI on installing a new reporting system. There is none. But you'd better install the system-using tools other than ROI to evaluate what we call a Staying in Business investment in your infrastructure.

If you think a technological development such as the Internet might change your business radically, how do you figure the ROI on an electronic commerce experiment? There is one, but you don't know what it is. Not yet. You need to make experimental investments to protect yourself from violent changes in your markets and to see if you can find any surprising opportunities, so you need new tools here, too. We think of this class of financing decision as Option-Creating Investments partly because the tools for evaluating their prospects draw heavily from sophisticated Wall Street option theory, and from the related portfolio-management techniques that have made so much money for so many venture capitalists.

Let's look at how to evaluate what merits your limited investment dollars:

STAYING IN BUSINESS: How do you assign an ROI to a project that ensures your critical information systems can deal with the Year 2000 problem? You don't. What is the precise value of improved communication and collaboration? Who knows? But would you be willing to shut down your phone system if you couldn't show that the upkeep of it had a sufficient ROI? We doubt it.

The overreliance on ROI means that many infrastructure projects are not even attempted unless they can somehow be hidden beneath a "killer app" project, such as one to fix the Year 2000 problem. Instead, companies need to learn that the benefits of Staying in Business projects are too diffuse to quantify and should be evaluated against a list of business requirements and cost. They should also be evaluated based on the risk that they'll fail or come in late and over budget.

In the years to come, many more projects will fall into the Staying in Business (or SIB) bucket because your computers, networks, etc. are becoming more like your phone system. They aren't just supportive. They're fundamental to your ability to do business.

RETURN ON INVESTMENT:ROI is well-enough understood that we'll offer just two observations.

First, there's still room for improvement. The chief information officer of a Fortune 100 company told us that his staff still can't do basic business analysis, a problem that he thinks is the biggest facing companies investing in information technology.

Second, ROIs are seldom what they seem. We recently were called by a chief executive who wanted to know why $15 million of cost-cutting that we did for him hadn't dropped straight to the bottom line. We spent two weeks frantically confirming the savings. The problem was that, in the face of competition, most were being passed on to customers. We're not saying he shouldn't have cut his costs, just that operational excellence is only the starting point for competition these days and will seldom deliver a sustainable advantage.

OPTION-CREATING INVESTMENTS: This is where the real payoff can come, by creating a competitive advantage that could last years.

You start by identifying an idea, such as electronic commerce, that could transform your business. You think of several ways to explore its possibilities through pilots or test markets. You determine the minimum that you could invest in each option while still learning quickly which will work. You judge each based on standard option theory. This lets you look at potential return against: the price of the initial option; the likely price to turn that option into a real business; the estimated risk; and so on. (As a general rule, you should look for a potential return at least 100 times the initial investment.) Then you invest, probably in five or six options. You assess them at set intervals, quickly killing off losers. When one starts to pan out, you decide whether to turn it into a real business based, finally, on ROI.

As an example, we work with a large life insurance company that is spending $2 million a year exploring how the Internet could transform its business. By aggressively learning about the Internet, the company can defer deciding whether to spend $100 million or so on upgrading information systems and changing business processes until the timing, implementation plan, and expected benefits become clear. Of course, the insurer can always walk away from any Internet plans, for only the cost of the option.

This is much better than the traditional approach, where, despite great uncertainty, a senior vice president takes his best guess at the results and asks you to approve a $100 million project right away.

Ultimately, you evaluate your success or failure based on the total portfolio of options you create. The model here is venture capitalists, who assume that at least four of five ventures will fail but that the fifth one will be Netscape and more than make up for the losers.

Option-Creating Investments (or OCIs) are the most underexploited type of investment because their high risk and uncertainty obscures their ability to generate high returns. But if you don't invest in Option-Creating Investments, you'll wind up with a portfolio overweighted to propping up the past and underweighted to exploring the future. It costs surprisingly little to achieve a healthy balance.

Messrs. Kingley and Andrews are partners with Diamond Technology Partners. They can be reached at kingley@diamtech.com and andrewst@diamtech.com.