(FORTUNE Magazine) – The world is your oyster. Do you have the right fork? In my last column (fortune.com archive, Feb. 15) I described how uncertain CEOs are about globalization. They know they need to be global, but they don't know if they are doing it right. It's as if they found themselves a banquet where, at every course, they weren't sure which flatware to use or how to hold it. They want to know globalization's best and worst practices.

It sounds odd to compare sophisticated, worldly companies with rubes at a feast. Yet confusion is widespread. At the recent meeting of the World Economic Forum in Davos, Switzerland, Jacques Manardo, Deloitte Touche Tohmatsu's European chairman, said, "Globalization is a hollow word. There is no generally accepted understanding of what it means." With the forum, his firm has completed the first report of a multiyear study of innovative practices in globalization. (Manardo's colleague Keith Ferrazzi led the task force that prepared the study; I was an adviser.)

As I wrote last time, its first conclusion is that the success of any globalization effort must be judged in terms of shareholder value. It can't be measured by, say, the percentage of a company's sales outside its host country; that would rank parochial outfits in Belgium or Denmark ahead of world spanners like General Electric or Ford, whose home market is necessarily a huge hunk of sales.

Nor is globalization just a checklist of skills needed. Creating value demands turning globalization itself into a point of differentiation and advantage. To help companies measure that, the Deloitte/World Economic Forum group created a framework with which to analyze and manage the process. It tests globalization against six broad organizational capabilities: governance and responsibility, strategy and planning, marketing and service, operations and technology, research and development, and organization and human resources management.

Going global places new demands and creates new opportunities in each area. Hewlett-Packard, for instance, must girdle the earth with offices providing round-the-clock tech support to customers. But the company turns necessity into a mother lode of opportunity by networking the computers of all 35 customer-response centers so that technicians anywhere in the world have real-time access to any customer's records; they can easily pass a knotty problem to the best expert, wherever she is, or move an especially time-consuming one from an office staffed by a night crew to one working at full daylight strength.

Each of the six capabilities can be opened up to give a rich, measurable view of how well a company uses globalization (see table). With governance, for example, value-creating best practices show up in a number of ways. On the board itself the right question is not how many directors are citizens of how many countries but what the sum of their global wisdom and experience is. Global regulatory compliance should help lower a company's cost of capital by allowing it to raise money wherever it's most plentiful and cheap. Wise global finance management should lower taxes. And strong governance should measurably lower labor, environmental, and political risk.

Research and development offers other examples of adding value through globalization. A Du Pont review of R&D operations turned up at least three innovative ways to leverage globalized R&D. First, says Du Pont CEO Chad Holliday, no company can manage R&D from an asset-based frame of mind. Just because you have a plant in Delaware is no reason to conduct research there--not if the intellectual capital is in France. Agrees Microsoft's Bill Gates: "Before, because of the overhead needed to build local offices, the history of globalization was the history of big oil companies and big manufacturing companies." In the age of the Net, knowledge processes like R&D can be global irrespective of assets, and companies selling knowledge products can have a global presence without much of a global footprint.

Second, says Holliday, Du Pont's historians find research is most likely to be productive if it's aimed at a specific need. Nylon, for example, was invented to replace silk in women's hose. However, research is most lucrative when it finds a broad market; today silk replacement accounts for less than 5% of the nylon market. How to combine focus and diffusion? Holliday describes a T-shaped process: "Look at your target, hit it, then look laterally for other opportunities." The ability to look laterally is built into Du Pont's marketing, where each local market is segmented, subsegmented, and subsegmented again--then the sub-subsegments are networked across business units.

A third way to leverage global R&D: "Find new friends." Most autopsies of failed R&D projects show the same thing, says Holliday: "The questions asked by management were answered well; it was the questions we had no idea about that killed us." The main reason: too many old friends; too many people from the same industry, the same country. Now every major Du Pont project gets a visit from three outsiders--academics, brainy types from different industries--who look at the research, ask questions, and deflate balloons.

Of the six capabilities, human resources gets CEOs most animated. Fewer than one in six executives think their companies are doing first-rate work in hiring and developing a global staff, according to a Conference Board survey of executives at 373 corporations. Yet most would agree with Gates: "The key is to attract the best talent and get the best out of it."

Among the issues:

--Compensation and advancement: Lavish pay and stock options at U.S. companies make it hard to get good Americans while not offending equally good Europeans and Asians. The ability to attract people is fatally compromised if some feel cut off from the top. Surveys often ask if employees feel they have a chance to move up--but are the results scrutinized for global variations?

--Incentives: What's a good mix of national, business-unit, and global incentives? At DHL, says CEO Robert Kuijpers, "we pick up in one country and deliver to another. The [second] country is only cost--no revenue." Customer satisfaction, however, depends on the second country.

--Values: ABB Asea Brown Boveri put its values statement through a multicultural rewrite. One reason, explains CEO Goren Lindahl: In some cultures people are uncomfortable talking about profit--though they don't mind making one.

There's no gainsaying the worth of getting the people part right. At HP, for example, the famous "HP way" makes for a global culture so strong that HPers say they can sniff each other out in a roomful of strangers. That makes it easy to change jobs within HP--indeed, according to CEO Lew Platt, each year about 10% of HP's employees take a new job within the company. The result, says Platt: "People become loyal and flexible. Our attrition rate is one-third of the industry average."

Ultimately, the World Economic Forum's aim is to create a knowledge base of global practices that work, together with a handbook of ways to measure success in globalization. Companies are too different, probably, ever to give out globalization Oscars. Besides--who cares? The real prize is creating value, and the real lesson is this: The smaller the world gets, the greater the number of its pearls.