Insights from the people behind the technology at Hollywood's hottest start-up

Barely two years ago, three of the most powerful men in Hollywood came together to do something that hadn't been done for 50 years: build a movie studio. It's called DreamWorks SKG, as in Steven Spielberg, Jeffrey Katzenberg, and David Geffen, and it's one of the best-financed start-ups in history. The three founders dug deep into their silk-lined pockets to produce $33.3 million each, but their combined offerings account for only about 5% of the company's equity. The remaining $2 billion or so is spread out among a who's who of wealthy investors and institutions. Most notable is Paul Allen, cofounder of Microsoft, who purchased 23.4% of the studio for $492 million.
So what will S, K, and G do with all that money? The plan, as it stands now, is to produce about eight motion pictures a year, television shows, interactive games, and music CDs. There's even a video arcade in the works.
Of course, DreamWorks doesn't just have cash; it's also got talent. So far, it's had little trouble luring the most gifted writers, actors, musicians, producers, directors, and animators in the business into its fold. But for DreamWorks to sustain its growth from start-up to entertainment giant, it'll take more than just a star-spangled cast. Like any new company, it will have to be faster, smarter, and more efficient than the competition. That's why the studio was as selective in hiring its chief information officer as it was in recruiting its top animators. Soon after launching the company, the founders plucked Lynn Jacobs from MCA to head up their information-systems (IS) team. Jacobs is a star in her own right, with credits that include 10 years as an aerospace engineer.

Jacobs was charged with assembling information systems that could keep pace with a company expanding at supersonic speed. When she first arrived, in January 1995, DreamWorks employed a scant 20 people. Today the DreamWorks machine has more than 900 workers. In many fast-growing companies, technological systems are built ad hoc -- applications are added one by one as needed. Jacobs and her team are taking a more global approach: first determining -- even anticipating -- the automation and business needs of each of the company's nine units and then designing applications that are both specific to each unit and capable of being combined into one centralized system. The goal is to have applications that can one day fit together without a hitch.

Although the new studio shuns titles (and dress codes), each top member of Jacobs's technology department has a distinct role: Petrus Williams heads up operations; Ann Delaney, Afshin Cangarlu, and John Sasali design and build applications and systems; Tom Kotlarek is responsible for the overall systems architecture; and Hugh Mendelsohn acts as a bridge between IS and the business end. We brought together Jacobs and her crew in Jacobs's office to discuss the advantages and disadvantages of starting a company from scratch.

Besides the star power of the founders, what attracted you to DreamWorks?

Jacobs: DreamWorks offers a chance to start with a blank page and do things the way you think they ought to be done -- to use your creative energies to reinvent business practices. I've always had different ideas about how to build technological systems. This is an opportunity to prove to myself and others that some of those ideas actually make sense. I've always believed that the entertainment industry could do things better and more efficiently. But it's hard to make the changes necessary for those improvements when you're working in a big company with an established bureaucracy.

* * *

What's it like to have infinite resources at your fingertips and such deep pockets?
Jacobs: That's probably the biggest myth about us. Everyone thinks that we must have a sky-high budget and unlimited resources. It's really not true for the technology group -- we operate pretty close to the bone. The company is well-funded, but the plan is to spend as little cash as possible on back-office and technological infrastructure. The real money gets dumped into developing product.

* * *

What are the advantages of working for a start-up?
We are trying to build applications for business units that still haven't been entirely defined. That's an incredible opportunity for us, because we can work with the businesspeople to determine the best business practices. It's also an incredible challenge. But if we really do our jobs well, we should add value to the process.

Delaney: One way we differ from the competition is in how we're setting up the accounts-receivable infrastructure. We've had to analyze how cost receivables work for every individual business unit, from animation to interactive computer games. We've then used that intelligence to think about how we could design one centralized accounts-receivable system. Most studios have decentralized systems to accommodate the specific needs of all the different business units, and at this point it would be too costly for them to change. But since we are a start-up we can build those systems just the way we want them. By keeping the receivables in one central place, we can very quickly figure out the needs of our customers, even if they work with us across many business units. For example, we'll be able to quickly view the receivables for Blockbuster regardless of whether the company is buying music, video, or interactive products. And that capability will help us make accurate cash projections for the entire company.

Can you give other examples of the applications you're designing?

Cangarlu: We purchased a system from MGM to help us track the distribution of our films and then customized it to better accommodate the needs of our sales force. Typically, the sales force begins scheduling the release of a film before we even have the final print, which could be three or four months before the actual release date. Then the salespeople head out and begin signing up theaters. The MGM system was originally designed so that the sales force would have to fax or mail back potential booking dates, which bookers would then enter into the system. But we will never hire bookers because our salespeople can access the system from the road and enter the information themselves.

In addition, we've attached some historical data to the system. Salespeople can pick out a group of theaters in a certain zone and find out their grosses for a particular genre of film for the past two-and-a-half years. They can download that information to their laptops and have it in the field. So if we have an animation release, for example, we can target the theaters that have had a better track record with animated releases.

What special challenges do you face?

Jacobs: Many companies have a fast rate of growth, but they're growing very fast in one manufacturing capacity, and from a technological-systems point of view that may require only one set of applications. But here we have nine business units that are vastly different -- they have different products, different distribution, and different methods for calculating revenues. We are growing more like the solar system. It's what I call the big-bang theory of business growth. So we have to be able to learn different business needs quickly. Also, many of our business ideas are unproven -- they're essentially experiments. What's interesting from an information-technology point of view is that we have to build systems that are nimble enough to support our new visions of business.

* *

You're building systems in a modular way, much as General Motors builds one type of engine for, say, Chevys and Pontiacs. How does that approach work for you?
Consider our tracking system for talent contracts. One problem with talent contracts is that there are many variables that may change based on information that you don't know. For example, you might specify that a certain producer will begin working six days after a certain actor has been signed on. So we'll build a module that automatically triggers changes in the contract as certain pieces of information get entered. It's what we call a cascade effect. That same module will then get reused in our availability system, which helps track entertainment product as it moves from one market to the next, based on a traditional life cycle. For example, a film might migrate from domestic movie theaters to international movie theaters and then to airlines and videos. The time period in which the product can move from market to market is driven by availability windows. The close of the international-movie-theater window, for instance, governs the opening of the video window. Since we've done our enterprisewide analysis up front, we don't have to duplicate our programming efforts.

* * *

Is DreamWorks' growth overwhelming?
Do I get worried when I hear that music is no longer on the back burner? Yes. But I see 1997 and 1998 as the real rubber-meets-the-road years for DreamWorks. By that time we'll have major product out in the marketplace, and the first animated film will have hit -- and then our systems will really begin to get tested. We'll see which units take off and which don't, and we'll start to see what this company is and what it isn't. Right now we have a lot of vision. And visions sometimes change. We see what we're doing as a marathon rather than a sprint, and until 1998 it's uphill all the way. We can wait until the turn of the century for our vacations.

* * *

Joshua Macht ( is a staff writer for Inc. Technology.