(FORTUNE Small Business) – Ed Catmull is president of Pixar Animation Studios, the outfit behind such cartoon classics as Toy Story and Monsters, Inc. When he was growing up, it was Catmull's dream to be a Disney animator, but he was soon confronted with the limits of his drawing abilities. Instead he enrolled in the University of Utah's computer science program. It proved a fortunate choice: During the early 1970s the school was on the forefront of a revolution in computer graphics. Catmull's classmates included Silicon Graphics and Netscape founder Jim Clark as well as John Warnock, who founded Adobe Systems. After earning a Ph.D., Catmull ran a research laboratory before landing at Lucasfilm. That was in 1979, when Star Wars was a recent phenomenon, and George Lucas recognized the value of computer-generated special effects. In 1986, Steve Jobs bought Lucasfilm's computer division for $10 million. Catmull became one of the co-founders of a new, stand-alone company, christened Pixar. The first decade was tough--so tough that Pixar had to earn its keep selling software rather than movies. Then came Toy Story in 1995. That watershed film and three subsequent full-length efforts have grossed $1.5 billion worldwide. The studio has won 15 Academy Awards, primarily for its technical innovations. Meanwhile an unusual combination of technical expertise and management savvy has kept Catmull atop Pixar, where he oversees a staff of 670. FSB's Justin Martin met with Catmull at the company's studio in Emeryville, Calif., for an animated discussion about innovation: how to continually push the envelope; rewarding people; and the joys of happy accidents.

What sets Pixar apart from other movie studios?

We are the only major studio working exclusively in computer-generated animation. We have created a variety of original software programs. Our programs can model surface textures, such as metal or cloth--or hair, in the case of Sulley, the lead monster in Monsters, Inc. We also have programs that create lighting effects--bright sunlight or flickering candlelight. All the data are captured in a program called RenderMan, which, as the name suggests, helps artists create the final images. RenderMan won an Academy Award for technical achievement in 2001. Pixar has a wealth of technical wizardry at its disposal. But I think we're also distinguished by our movie scripts. You have to tell a good story. That's the foundation of our business.

It sounds as if you're suggesting that in the race to innovate, merely possessing gee-whiz technology is not enough.

I believe that this holds for a lot of fields. Cutting-edge technology can be exciting, and you can use it to get both employees and audiences motivated. But you have to strike some kind of balance. Clearly there are a lot of movies where the special effects are excessive. Twister is a prime example. Why does this happen? I believe it's because directors come in and sign up a good effects studio, of which there are plenty. It's like a kid in a candy store. It winds up taking all their energy. The film comes out, and the effects look great, but the story sucks. We have some amazing technical abilities. But we still have to rely on good old-fashioned storytelling.

You've worked with two master innovators, George Lucas and Steve Jobs. What did you learn from them?

From George, risk taking. It's important to be able to think you can take risks. As for Steve, his talent is identifying good people and giving them free rein. His style can best be summed up as "don't micromanage." He's the CEO of Pixar. But you have to understand that he's also running Apple. We have all these incredible people here, and Steve doesn't tell them what to do or what stories to pick. Part of his genius is recognizing their genius and leaving them to it.

How do you institute a culture of innovation?

You have to stay ahead of the curve. As soon as you're successful, competitors jump on the bandwagon. They say, "We now have proof that this works. Therefore we'll try it." And you have to be careful because sometimes the imitators will do a great job. You can't get complacent just because they're copying you. It's conceivable that they may outdo you.

I encourage our people to think in terms of steps. Every new movie is a stepping stone where one can learn new things about what works and what doesn't. The question is, Are we a stepping stone for ourselves or for somebody else? How do we become the ones who benefit from what we learned?

How do you ensure that you're the beneficiaries?

Intense self-scrutiny. We constantly worry whether we're taking the right path. I remember when we were nearing completion of Toy Story 2. It wasn't taking shape, and we were running out of time. We were very concerned. Given the tight production deadline, it would be virtually impossible to revamp the existing story. Our solution: We threw the script away and started over. We redid the entire movie. We made Toy Story 2 in nine months, when it usually takes four years to create an animated feature. I'm very proud of that. At the end of each feature we do a brain dump. It's not a matter of blame for things that went wrong; it's a candid self-assessment. Out of this self-assessment we come up with an operating theory for the next film. During the next film we'll fix those problems, and we'll discover new ones. You can never perfect the process.

Is there such a thing as a happy accident?

Sure. When you're experimenting a lot and taking risks, sometimes things work out in ways you weren't expecting. Early in production, for example, we use our employees to do sample voice-overs for the various animated characters. Later we go out and select actors such as Billy Crystal to do voices. But once in a while an internal voice ends up being just right. That's how artist Joe Ranft became the voice of Wheezy in Toy Story 2. Happy accidents do happen. The key is to look for them, encourage them, let them happen.

Tell us something counterintuitive about the process of innovation.

You can't do it full-time. What we have at Pixar are a lot of overachievers working for overachieving managers. People need periodic breaks to recharge their batteries.

We try to set limits here. On-site we have a soccer field, volley ball courts, a gym--and we offer classes in yoga, tai chi, and Pilates. There's a full-time ergonomics expert at Pixar to make sure people have the proper desks and their computer keyboards in the optimal position. We have masseuses and a doctor who comes in once or twice a week so that employees can catch problems early on. We want people to enjoy long, productive careers and not burn out. They need replenishment.

Pixar seems to place a large emphasis on people. So which would you rather have: a really great idea or a really great team of people?

That's easy. If you have a great idea, and you give it to the wrong people, they'll screw it up. If you have the wrong idea, and you give it to the right people, they'll fix it. That was the lesson of Toy Story 2. So what's our priority going forward? Is it finding good ideas? No. It's finding the right people.

How do you hire the right employees, people who will contribute to a culture of innovation?

This is an area where we used to have some wrongheaded assumptions. It drove me nuts for a while. When we hire people, we're generally under the incredible time crunch of a production. The natural inclination was to hire people who had reached a certain skill level, people who assuredly could do the job.

That's a mistake. It took a lot of effort to break down that mindset. What you really want to do is select people who are on an upward trajectory in the arc of their careers. You are better off picking people based on where they are going than on where they are now. We've learned to hire based on potential rather than on position.

You also need to hire people who are better than you are. As managers, we have to be able to hire people who can do things that we can't. On the technical side, there are people here with skills I can't possibly match. John Lasseter [Pixar's executive vice president--creative] hired people such as Andrew Stanton, who does things that he can't. Andrew wrote the script for Toy Story. Often managers hire down because they don't want to be threatened. If you do that, you're dead.

How do you reward people?

We have our star performers here, like anywhere else. And we know who they are. They command higher salaries, receive larger blocks of options. But moviemaking is a collaborative process, and when a project succeeds, it isn't usually due to any one person. When one of our films does well, everybody gets a bonus.

One thing that's unique about our culture, I think, is that we recognize that the artistic side and technical side are equal. We've set it up so that each has the potential to earn the same compensation. We have had threats to that structure in the past. We've hired people, for example, who have tried to bring in the class structure of other places. But we've kept the culture at parity, and that's healthy, I think. I look at other places and see that one group is considered first-class citizens, another second-class. It can be sales forces vs. R&D, or whatever. But in any endeavor, it's not as if you have the geniuses on one side and the drones over there. I think there's a lot of unhealthiness in cultures that let one side predominate.

Pixar has now made four commercially successful movies. But you continue to make shorts that don't earn a dime. Can you explain why?

We do them for training and to try out new technology. Short films provide a great opportunity to take risks. And sometimes we fail utterly. Then we say to ourselves, "I'm glad we had that train wreck with miniature trains instead of with real trains." But often they succeed in ways that benefit our full-length features. An example is a short called Geri's Game [about an old man playing chess against himself]. That was where we pioneered software that helped skin appear realistic and expressive. Meanwhile, we had a huge team working on the feature A Bug's Life. They were looking over the shoulder of the Geri's Game guys and saying, "Whoa, that's cool." The innovations from the short wound up being spun off into A Bug's Life.

What's next for Pixar?

You'll see some remarkable underwater effects in Finding Nemo [summer 2003 release]. Then The Incredibles [winter 2004] will be our first film where all the animated characters are humans. We have to be able to make them inexpensively and control them and make them look good in this world we create. We've never worked with that many human characters before. It requires a great deal of precision. You can do things approximately correctly with toys, and it's okay, because how are toys supposed to look anyway? But moviegoers know every nuance of human movement and facial expressions. The difficulty with humans is that we are fussier with how they appear than with anything else. We'll encounter all kinds of challenges, no question, but I'm confident we'll find innovative solutions.

Justin Martin is a New York-based writer and the author of biographies of Alan Greenspan and Ralph Nader.