Asked about the difficulty of decision-making under uncertainty, a Nobel laureate economist once reached into his pocket and pulled out a coin. "I occasionally use one of these," he explained. "It tends to be right about half the time."
While heads and tails remain popular options, most managers require a sharper set of decision tools. "Decision support systems" are the latest approach developed by management scientists and corporations to improve the quality of managerial decisions.
The concept behind decision support systems, says Peter Keen, a professor at MIT's Sloan School of Management who created the phrase, revolves around the question "What makes a useful system usable?" for a decision support system, Keen asserts, the answer rests in the way the techniques of statistical modeling and computer simulation can be customized to complement an individual manager's decision-making style.
"There is this illusion," says Keen, "that one needs big and complicated models to solve big and complicated problems. That is the wrong approach. Approach it from the question of how do people do their jobs? And notions of clarity, rather than complexity, become the important ones."
"Basically," says Keen, "A DDS contributes to the simple nature of decision-making."
This idea of letting a manager directly dictate to the model, rather than the other way around, has begun to gain acceptance at several major corporations -- including Xerox, Prime Computer, Citibank, Eastman Kodak and Carnation.
"It really is something quite unique," says Doss Struse, director of marketing research for Carnation. "It makes people think differently about the marketplace."
Struse has just introduced a DSS for the marketing managers at Carnation -- he had previously structured one at Oscar Mayer. Struse estimates that Carnation will spend almost half a million dollars to get its new system ready.
Carnation's DSS is built around a computer language called Express. Express, says Struse, permits a manager to learn quickly how to develop crude models of the market that can be refined and enhanced as the more sophisticated functions of the language are introduced.
"Our DSS personalizes the decision process," says Struse. "It lets managers operate at a different level of detail, it's much more geared to the data and the manager's perception of what drives his business -- margin, market share, external factors, etc."
However, Struse concedes that it is difficult to measure the success of a DSS. "I guess the real questions are, 'Did they get value from it? And is it worth spending time with?' I've found that they're saying "Gee, this is a valuable thing for me to do." Ultimately, Struse predicts, a DSS will spawn a tougher and more sophisticated array of questions for the manager to answer. "I think what we're going to see is a different type of question. The questions will shift from 'What happened?' to 'What will happen' if we do X?"
This planning aspect of a DSS particularly intrigues George Nix, manager of market intelligence systems for Xerox. Nix has helped develop Xerox's main system -- market intelligence.
"Call it an information utility," Nix says. "It encompasses the people, the procedures, data, information and computer tools. It's our systematic way to improve the process of collecting information."
Main is currently an in-house service for Xerox's corporate planning department, although Nix does not rule out the possibility that it might eventually become a Xerox product.
"With Main," says Nix, "a data base can simultaneously be a model. People who've worked with it feel they've saved a lot of time in their strategic planning efforts."
Xerox is trying to develop a very sophisticated DSS based on the work of University of Chicago philosopher Stepen Toulmin. Essentially, Toulmin presents a linked logical hierachy for assumptions, claims and evidence that Nix hopes to turn into a data-based structure. If Xerox is successful, says Nix, people using this system will be able to fine tune their data and assumptions in such a way that a chain of decision consequences will be displayed for each adjustment made.
At the University of Pennsylvania's Wharton School, several management scientists are trying to incorporate the techniques of artificial intelligence research into their work in DSS.
One fear both Keen and Struse have in regard to DSS is the way the term is fast becoming a buzz-word amidst the rapid changes in the computer field. "What's happening is quite fascinating for me to see," says Struse. "The data processing people and the operations research people are picking up on the DSS terminology and that's not what it's about."
Keen stresses that a goal of DSS is to take computer power away from the computer room and give it to the individual manager.
"The breakthrough," says Keen, "and it really is a breakthrough -- is that computer technology is no different from anything else -- it's merely a means to an end."