I study organizations partly because they scare me. They have the power to make smart people do stupid things and be proud of it.
Most of us believe that we are too moral, smart or stubborn to follow a superior's misguided orders or to give in to the temptations of a pecking order or a reward system that presses us to take destructive actions. Yet a huge body of psychological and sociological research shows that most people dramatically overestimate their wisdom and strength of character. We humans usually do what we are pressed to do, have always done and are rewarded for, even when it harms others and undermines the effectiveness of our organizations.
I was reminded of these facts of organizational life by a Ph.D. dissertation that I recently supervised: Sally Baron's research on why the U.S. Department of Defense wastes so much money developing custom technologies even when "off- the-shelf" products work at least as well, and for a lot less money.
Calls for commercial off-the-shelf, or COTS, products are not new. In 1949, President Truman's Scientific Advisory Board urged the Air Force to make greater use of commercial products, chastising Air Force leaders because "existing organizations, personnel policies and budgetary practices do not allow the Air Force to secure the full and effective use of the scientific and technical resources of the nation."
William Perry, the U.S. Secretary of Defense from 1994 to 1997, has been the most adamant supporter of COTS in recent years. In a landmark 1994 memo, Perry emphasized that the technical superiority of U.S. forces depended on taking advantage of innovative information technologies. He argued that developing custom hardware and software made sense during the early days of the computer revolution, when off-the-shelf solutions were rare and typically inferior. But he said it didn't make sense anymoreand, in fact, was wasting billions of dollars.
Unfortunately, the Defense Department is still wasting huge sums on custom technologies. Consider the fates of two DoD satellite ground control facilities that acquired new computer systems. Baron, in her dissertation, called one Sword and the other Stream, to protect the confidentiality of the people involved. Sword and Stream had identical technical requirements: Both were responsible for the command and control of orbiting satellites and related tasks, such as analyzing orbits. The main difference was that Stream's new system (which became operational first) used only COTS hardware and software, while Sword's new system used a mix of custom and COTS hardware along with custom software. Sword program managers knew of Stream's success, and Sword's initial request-for-proposal specified that COTS products would be used when possible. Yet Sword still used custom software.
The resulting differences were staggering. For example, Stream's system cost $1.25 million (mostly Dell computers and COTS software); Sword's cost $40 million. Stream required annual maintenance of $200,000 per year; Sword required $2 million. Stream took six months from the time the acquisition was approved until it became operational in late 1999; Sword took 32 months from approval until it became operational in mid-2000.
The response? The Air Force brass ignored Stream's lower costs and other virtues, including ease of training and reduced downtime, and shut the program down. Unfortunately, this is not an isolated case. Baron uncovered numerous cases in which COTS solutions were available but procurement officials still selected more expensive custom solutions.
How can such stupidity persist? Why does the DoD keep wasting taxpayers' money? First, Baron found in interviews with dozens of top DoD officials that the DoD's budget system provides no rewards for a program that saves money; leaders are simply asked to give it back. Yet if a program overspends, the budget is often increased, so there is little incentive for contractors or DoD personnel to save money. Not only that, but military officers who oversee many people and are responsible for large budgets are promoted over those with "less" responsibility, so officers who cut costs by reducing staff size and budgets are undermining their career prospects. Next, many older and larger defense contractors have strong ties to the government. It is difficult for new, lesser-known suppliersoften those with new technologyto elbow their way in. Third, historical precedent also is an inhibitor: The U.S. has the strongest and most advanced military in the world, which has bred an "if it ain't broke, don't fix it" attitude. One consequence of this "success trap" is that if an officer uses a superior new technology, the scrutiny and penalty for failure is greater than if he or she uses an inferior but tried-and-true method. This isn't unique to the military; much psychological research shows that humans strive to be consistent with past behavior because it makes them and others feel more comfortable. Beyond that, abandoning old ways is an implicit acknowledgement of their inferioritya sign of disrespect in a rigid culture.
I get angry when I read about these hurdles. After all, they are wasting my money. But then I realize that the Pentagon is much like other large organizations that I admire. Great companies such as IBM, Gillette and Intel have, as a result of similar forces, endured years of people doing the wrong things for the wrong reasons. When Lou Gerstner took over at IBM, there was so much dysfunctional internal competition that it undermined cooperation and knowledge sharing. Gillette spent much of the 1980s in commodity hell, trying to eke out minuscule profits from disposable razors. Intel is now famous for the quality and consistency of its manufacturing, but during its early years, the company had serious quality problems in new plants.
Changing how workplace behavior is rewarded, though, can turn things around. Gerstner drastically reduced internal competition at IBM by paying team players more and promoting them more often. Gillette broke out of commodity hell because insiders fought to reduce the influence of an entrenched small network of executives, and brought Gillette back to its roots as the maker of quality razors that could fetch higher prices. At Intel, quality had become a problem, in part, because each plant was operating in a slightly different fashion. Craig Barrett and other Intel executives responded by aggressively pushing through their "copy exactly" philosophy, which requires every Intel plant to be exactly the same, driving out unwanted variance.
Change is probably harder to achieve in the Defense Department, thanks to high leadership turnover and strategies that shift from one ad-ministration to the next. Fundamental change, though, requires cultural change, and the most sweeping change often starts with a simple message, repeated again and again by senior management. IBM's Gerstner emphasized repeatedly that team players deserve the biggest rewards. Gillette executives like John Darman and John Symons kept reminding people that they needed fine razors to teach customers to focus on quality rather than price, and Intel's Craig Barrett still talks about the virtues of the "copy exactly" philosophy.
I once interviewed Ford's former CEO, Donald Petersen, who is widely credited with saving the company in the 1980s by improving manufacturing through his "Quality Is Job One" program. Petersen joked that he was boring because he said the same things again and again. But the balance sheet got better because his people knew what had to be done, even if it meant removing formidable obstacles.
When everyone in an organization understands what to doand that they must do itreward systems quickly change, people in established networks lose power, and dysfunctional precedents can magically disappear.
Robert I. Sutton is a management science and engineering professor at Stanford University and coauthor of The Knowing-Doing Gap: How Smart Companies Turn Knowledge Into Action. Sutton also co-leads Stanford's Center for Work, Technology and Organization. His next column will appear in December.
This article was originally published on 10-10-2002