Kellogg World Alumni Magazine, Spring 2002Kellogg School of Management
In DepthIn BriefFaculty NewsClass NotesClub NewsArchivesContactKellogg Homepage
From the Dean
Development Update
Research: Ronald Dye, Accounting
Research: Angela Lee, Marketing
Alumni Profile: Molly Battin '98
Alumni Profile: Bradley "Buzz" Calkins '00
Address Update
Alumni Home
Submit News
Internal Site
Northwestern University
Kellogg Search
  Professor Ron Dye
Operating in uncertain environments
Professor Ron Dye explains why business problems are ‘bandit’ problems

By Rebecca Lindell

A gambler stands poised between two slot machines. He knows one of them has better payoff odds than the other, but he doesn’t know which machine that is.

He’s got enough cash for 20 pulls of the handle. He chooses randomly between the two machines for his first play. But what should he do after that first choice?

Suppose his first play yields a payoff. Should he “stick with the winner?” Or should he move to the other machine?

“Staying with the machine that produced the payoff may seem to be the better strategy,” says Ronald Dye, the Leonard Spacek Professor and Chairman of Kellogg’s Accounting Information and Management Department. But that choice, he adds, has the defect of not providing any information about the potential payoff of the other machine. “It may make sense to experiment and try the other machine, if the number of times you intend to play is large enough.”

Does this gambler’s dilemma provide any insight into how business decisions should be made? Dye says yes.

“Every choice between slot machines has an effect on the gambler’s immediate payoff and on what the gambler learns about the two machines,” Dye says. “Similarly, every important business decision a manager makes has two effects. It influences the firm’s profits, and it influences what the firm learns about the environment that it operates in.”

A manager wishing to improve the marketing of a product, for example, could expand the firm’s sales force or do more advertising. Though the manager will never be certain which was the better strategy, he or she can reduce that uncertainty by experimenting with the different tactics.

Dye says this approach can be applied to almost any important business decision, from deciding what supply chains to use to what prices to set to what products to market.

“Your best decision today may not be the one that maximizes today’s expected profits, based on what you know about the environment you face today,” he says. “Instead, it may be the one that gives you the most information about your operating environment, so that your expected profits increase tomorrow.

“For example, you could pick a price for your firm’s product today based on your best understanding of the market demand for it today. Or, you can purposely pick a price that you don’t believe will be the most profit-maximizing price. The information you gain about the product’s demand curve from the latter choice may help you make better decisions about how to price your product in the future.”

The feedback from such experimentation can come from many places. For experimentation in product pricing, product markets are the sensible place to look. But for major business decisions, Dye says the capital market’s reaction may be a useful source of information.

“Conventional wisdom generally asserts that managers have better information about their strategies than the capital markets do,” Dye says. “But the fact that senior managers are often surprised by the market response to their strategic proposals suggests that, at least sometimes, the capital market has information the managers don’t have. This is information managers could exploit — if they are willing to listen.”

Dye offers as one example the fall in computer maker Hewlett-Packard’s stock price after it announced its plans last year to acquire rival Compaq.

“This should have been a wake-up call to the firms’ senior managers that their view of the desirability of the take-over was in doubt,” Dye says. “They may have been sensible to propose the takeover — as an experiment — and equally sensible not to have tried to consummate the merger, given the market’s reaction.”

©2002 Kellogg School of Management, Northwestern University