Revenues, operating profits, earnings per share--companies are always searching for just the right performance measure to drive compensation and motivate behavior that boosts the stock. In increasing numbers they are turning to economic value added (EVA), a measurement tool that determines whether a business is earning more than its true cost of capital. This gives managers a clearer idea of whether they are creating (or destroying) shareholder wealth. Eli Lilly, maker of Prozac, is among a pioneering group of companies that are tying pay to EVA goals. FORTUNE's Justin Martin spoke with Randall Tobias, the pharmaceutical giant's CEO, to learn how such a system is paying off for the company.

Why did Eli Lilly decide to adopt EVA as a performance measure in the first place?

It really makes you align your thinking with shareholder value and causes you to focus on capital expenditures--and the pharmaceutical business is a very, very capital-intensive business. We started rolling out EVA around the middle of 1994. As a measurement, it really highlights the fact that once you've invested in an asset like a lab or a new drug, you're required on an ongoing basis to command a shareholder return on that asset. This year we're targeting our capital expenditures to be somewhere around $500 million.

Why did you link compensation to EVA?

We had a cross-functional team, which for some time had been examining what we saw as a shortcoming of our incentive system-- that executive pay was linked to sales and net income. There just wasn't a very good correlation at all with shareholder value. So with EVA consultant Stern Stewart, we developed a pay plan that fit our needs.

Basically, Lilly's bonus plan now requires managers to achieve continuous, year-to-year improvements in EVA. First we set EVA targets based on competitive factors. As a hypothetical, let's say we have a 10% cost of capital. Each year we have to beat that target--we have to earn more than that 10% cost of capital--and each year the bar is raised. It's a small percentage increase, but it keeps pounding at you. Whatever you did yesterday, you need to do better tomorrow to keep raising shareholder value.

If the targets are met or exceeded, there's a bonus, based on a proprietary formula. Last year, for example, the company did very well. Many managers got bonuses worth close to 50% of their total compensation.

The bottom line is: As you're making decisions, you've got to think about aligning them with EVA. It's easy to see EVA as a very sophisticated financial tool, and indeed it is, but I think it's important to understand that it is really a tool to change behavior too. Linking bonuses to EVA is meant to change the whole culture.

Is the whole company on an EVA bonus plan?

For the initial go-round, beginning in 1995, we tied EVA to bonuses for the 90 most senior executives at Lilly. This year we extended the plan to a broader group, which now totals about 350. Our intention is to go forward and to expand the size of the group to about 1,400 in 1997.

Down the line, do you intend to extend the plan even further into the organization?

Perhaps. But my impression is that many of the most successful EVA firms, Coca-Cola, for example, have not made massive shifts all at once. They have started at the top. You have to work out the bugs; you have to educate people. We want to see how this goes and not stir up the entire organization at once.

Still, as you would probably expect, the introduction of this plan has had an impact on behavior throughout the whole company because all the employees see what's coming. They know what's driving senior officers' bonuses, and there's really a pretty rapid change in attitudes toward capital throughout the business.

EVA is now something that has crept into the everyday language here at Lilly. There are stories about EVA in our employee publications. Managers talk with their employees about EVA and how it works and what it means.

Does Lilly's plan call for paying bonuses based on corporatewide EVA performance or that of the individual business units?

That's something we debated for quite some time. We're aware of both pros and cons in different firms that have approached it in each fashion. In our own case, we ultimately chose to base bonuses on corporatewide EVA performance. We are able to carry the calculation down to the various business units, but we do that more from a strategic standpoint.

I think one major factor that played into our thinking was that at about the same time we were implementing EVA, we were also implementing global business units for the first time. We thought that it would create too much complexity if we tried to make a very significant organizational change and put in a brand-new measurement plan on top of it.

But there's another important consideration. If we tied compensation to EVA performance at the business unit level, we might get each individual business unit behaving in a way that optimized its own performance. When we added up all the business units, I might find myself sitting here with an aggregation of EVA performances that might not optimize the interest of Lilly's shareholders.

Isn't there a danger that managers will make shortsighted investment decisions to boost their bonuses?

Yes. One of the things you never want to set up is a system where there's a great incentive to do something stupid in the short term. Theoretically you could lower your asset base and improve your EVA by depleting all your inventory. But then, come the first of next year, you'd have nothing to sell. That would be foolish.

To make sure that kind of shortsighted decision-making doesn't happen, we've set up something called a bonus bank that encourages managers to take a longer-term perspective. The bank, in effect, acts as a long-term scorekeeper. If a decision a manager makes today boosts EVA three years from now, the bank will reward him for it. But if the decision hurts EVA, the bank will ultimately cut his reward. That way, anytime you make a significant investment--build a new plant, say--you have to think about how it's going to pay off in the years to come.

Here's how it works. If you meet your EVA performance target in year one, you receive a bonus of which a portion pays out immediately, and the rest goes in the bank. The next year, if EVA improves, the manager not only gets his regular bonus (part of which gets deposited into his account) but also gets to withdraw, if you will, a portion of the bonus money that had been saved in his account from the year before. The more EVA improves, the bigger his total bonus--his regular bonus increases as does the amount he's allowed to withdraw from his account. And so on year after year.

What this does is make our managers accountable for the long-term effect of their decisions. If they make investment decisions, say, in year one that end up reducing EVA in year three, we will actually deduct money from their accounts and give it back to the company. But if the company has an extraordinary year? The EVA compensation system has an interesting characteristic in that, unlike our prior systems, which had limits, there is theoretically no ceiling to the EVA bonus. I think most of our shareholders would say, "That's a good a idea. Drive the value of my stock as high as you can."

What about linking Lilly's stock incentive plan to EVA performance?

We don't intend to do that as of now. Currently, we determine how many options employees will get based on a number of factors. Once employees have those options, the benefit they derive from them depends on how well the stock is doing in the marketplace. So to the degree that they understand that EVA really has the best correlation with stock price, EVA, in a way, is helping to drive the value of the stock options.

Is there any benefit, then--as is often claimed--in lining up all of a company's incentives with EVA?

I think there's actually a potential disadvantage in having everything solely and directly tied to EVA. But I think it's very important to understand that all of our incentive plans are completely consistent with EVA and that the rewards that pay out in our other plans will pay out more attractively as we produce increasingly positive EVA.

What advice would you offer other companies considering linking compensation to EVA?

There are probably a lot of very sophisticated theoretical approaches to doing it. But one of the principles that's very important to bear in mind is: Keep it as simple as you possibly can. Once you begin to make these plans overly complicated, employees don't understand them, and they don't have the desired effect.

I would also not underestimate the challenge of educating a global work force on EVA. We have a training course, and we've actually gone around to folks in each country to lay out the concepts of EVA. We generally spend a couple of rather thorough days talking about how it works, why it works, how to use it. It's one of those investments that go with rolling out EVA. But we've found it very, very beneficial.

Would you deem Lilly's EVA-linked compensation plan a success at this point?

I think we're off to a good start. One of the things we obviously looked at was how people fared in the EVA plan in 1995, relative to how they would have done had Lilly produced the same results under the old incentive plan. The fact is, they actually did slightly better.

I can also tell you that we exceeded our targeted EVA in 1995. That meant a positive balance went into everyone's EVA bonus bank.

Shareholders are making out well too. Since the end of June 1994, when we first adopted EVA, the share price has gone up 105%. That's what really matters.