Not long ago, the head of a supermarket chain with $5 billion a year in revenue gave a rousing speech to the M.B.A. students at a leading business school. He spoke of the chain’s innovative use of technology, including a “digital supermarket” pilot program that allowed customers to shop from home and have groceries delivered. Afterward, one student innocently asked, “How do you determine the payoff of your firm’s information-technology investments?”

The chairman smiled ruefully and answered. Borrowing a famous line about advertising, he said: “I’m certain half of our information-technology investment doesn’t pay off—the problem is I don’t know which half.”

His admission is all too familiar. Hardly any reliable information exists about the return on I/T investments. Although some studies indicate that businesses investing heavily in technology perform better than their peers, others detect no statistical correlation between dollars spent and results. The sad fact is that most technology investments are made on faith.

That needn’t be. Eight years of research at more than 100 businesses worldwide has shown that executives can determine the value they derive from their technology investments. Once they spot problems, companies can increase the value of I/T by using techniques similar to those for managing an investment portfolio.

When measuring the effectiveness of I/T, the most important thing isn’t to determine whether projects are developed on time and on budget, which is the focus of most studies. The most important thing is to figure out whether the right types of projects are being pursued—in other words, to determine whether the entire portfolio of I/T projects lines up well with overall corporate strategy. Before spending another dollar on I/T, top executives should step back and assess how their I/T operations fit with their strategy.

In doing this assessment, executives should figure out which of three archetypes they want their business to be: a cost-cutter; an agile innovator; or a blend of the two.

Then they should look at their I/T projects, which can be divided into four different classes of assets:

TRANSACTIONAL systems, which usually are developed to cut costs or to handle higher volumes of business. For instance, brokerage firms such as Merrill Lynch & Co. (www.merrilllynch.com) and Charles Schwab & Co. (www.schwab.com) have trade processing systems that handle huge numbers of transactions.

These systems typically have the most reliable payoff of all four asset classes, generating solid returns year after year. Transactional systems are like the certificates of deposit in your personal investment portfolio.

INFORMATIONAL systems, which provide the data for managing and controlling the organization. These systems help companies to sell to highly specific market segments or to scour data for trends that will lead to new product and service opportunities. For example, Seven-Eleven Japan Co. (www.sej.co.jp), the most profitable retailer in the country, each year changes about 70% of its product line. Achieving this industry-leading responsiveness requires tremendous investment in gathering information from the point of sale and skill in analyzing the often-changing needs of the Japanese convenience-store consumer.

Informational investments are like securities indexed to the Dow Jones Industrial Average. The return will be more volatile than for transactional systems but should be positive overall.

STRATEGIC investments, which are made to gain competitive advantage. For instance, Pacific Pride (www.pacificpride.com), a fuel company selling to fleets of cars and trucks in the U.S. Northwest [See “Low-Tech High-Tech,” Context, July/August 1999] pioneered a system that provides management reports to customers on fuel purchases by vehicle each month, including details on engine efficiency. Pacific Pride was able to charge an extra eight cents per gallon for this service.

Strategic systems are like investments in emerging markets. About 50% fail, but around 10% create spectacular returns.

INFRASTRUCTURE investments, such as a computer network or a companywide customer database, are long-term investments often put in place before the company knows precisely what business initiatives or applications they will be used for. By their very nature, these investments have no real value on their own. These investments, however, do allow quick and economical implementations of new applications of technology across a company—for instance, a system that will allow for cross-selling. Firms with better I/T infrastructures tend to be statistically faster to market with new products and services.

Infrastructure investments are like options. If you make the right bet, you generate a terrific return. If you make the wrong bet, you may get nothing.

Once you’ve divided your I/T projects into the four buckets, you’ll be able to make some strategic decisions about what the right mix is for your business.

First, you can do some benchmarking. You can compare yourself against the average company, which had 12% of its investment in transactional systems, 16% in informational systems, 14% in strategic, and 58% in infrastructure. You can also compare yourself against companies in your industry.

You won’t necessarily want to line up with the overall average, or even the average for your particular industry, but you need to be able to explain, based on your particular business strategy, why you differ from the averages. Companies that focus on agility typically want to overweight infrastructure investments and underweight transactional ones, and they tend to spend more on their overall investment portfolios than is average for their industries. Companies that focus primarily on cost usually want to overweight transactional systems and underweight strategic investments. They often spend less on their investment portfolios than the average. Companies that seek a balance will likely want to stick close to the averages.

Comparing yourself against competitors and, more importantly, against your strategy can be painful but is also enlightening. A few years ago, a company we’ll call InvestCo, a large financial-services firm, found itself unable to compete with a smaller company that offered multiple financial services for a single fee and provided the customer with a consolidated report. InvestCo convened a companywide information-technology council to find out why. The answer was that InvestCo had let business units operate entrepreneurially even when making large I/T investments. The systems were ultimately incompatible. So even though InvestCo’s I/T spending outpaced its peers’ by a wide margin, the company couldn’t offer a combination of products or even a report of all its interactions with each customer.

InvestCo, while thinking of itself as an agile company, had unwittingly underspent on infrastructure investments that would have given it a common platform for culling customer data for efficient cross-selling. Once it identified the problem, InvestCo began to address it.

Having conversations about the mix of I/T investments and about how they fit with corporate strategy provided one additional benefit: Those conversations let technical and nontechnical executives overcome the language barrier that so often makes it hard for them to work out joint strategies and plans.

Weill is the director of the Center for Information Systems Research at the MIT Sloan School of Management. He is the co-author of Leveraging the New Infrastructure. Connolly is a partner with DiamondCluster International Inc. and president of its Global Diamond MarketSpace Solutions.