As I See It | by Tom Temin: Complexity curve rises sharply

Austin Yerks, with his more than three decades of experience in selling services to the government, is just such a guy. Yerks is president of the Defense Integrated Solutions and Services Division at Computer Sciences Corp. He spoke earlier this month at a small conference, organized by ICG Government of Richmond, Va., that took place during the Consumer Electronics Show in Las Vegas.

(Why CES? Several reasons, but perhaps most significantly: It's getting more and more difficult to distinguish between consumer and business-industrial technologies.)

Ranging over the history of the ceaseless interplay between government and industry, Yerks made several incisive observations.

First: Enterprise systems developments should be done via fixed-price contracts to get agencies and vendors out of the dreary cost-overrun game. But, Yerks said, that would require any change in requirements offered by the government to be answered with a cost and scheduled impact statement by the prime contractor.

Every big project goes over cost and over schedule. Yerks said this is not only understandable, it is even to be expected. Why? "Because we never really understand the complexity" before diving in, he said.

Yerks' conclusion: "I don't think cost should be a variable in large acquisitions, except those for commodity products." For integration or modernization efforts, contractors, like divers, should get a degree-of-difficulty consideration, he said.

That's not as radical as it sounds. Doesn't the government already end up paying for complexity and shifting requirements?

Enterprise efforts also would be enhanced, Yerks said, by having much more formal governance setups in place from the outset. An important element of that governance would include much clearer expectations agreed to by government and contractor.

As an aside, Yerks cited consulting companies as being better at clarifying, and therefore limiting, government expectations. The mind-set of most aerospace and integration contractors leads them to say, "Sure, we can do that," to every government question, he said.

Yerks questioned whether agencies could, in any cogent way, meet their enterprisewide integration and transformational goals using multiple task orders under multiple-award contracts. Given the quantity and scope of these contracts expected this year, it's a sharply relevant question. (See story, page 14)

It's almost as though, having given up on large-scale requirements contracts, the government is careening in the opposite direction and into multiple multiple-award contracts. The theory behind that move seems to be that because anything smacking of grand design is dismissed as impossible, the only viable alternative is a thousand small bites. Yerks wondered how to keep the big picture in mind in an environment of year-by-year appropriations and constant chief information officer and program manager turnovers.

Yerks' last piece of advice was for contractors. Hooking a big fish is only the beginning; reeling it in requires executive attention throughout the life of the program until that fish is landed and in your creel. Too often, he said, once the reeling has begun, top management's attention wanders to the next new thing, and the program suffers.

Much as the environment has changed, Yerks said, some ethos endures.

"Today, I see the same truths as 20 years ago," Yerks said. "We need to understand the government customer's mission and unique requirements."

Tom Temin is executive editor and senior vice president of PostNewsweek Tech Media, publisher of Washington Technology. He can be reached at ttemin@postnewsweektech.com.