Stanford Business magazine

He Wields Statistics to Solve Problems

When he is not thinking abstractly, Professor Michael Harrison likes golf. He plays the Stanford course with friends: Denis Coleman, PhD ’75; Tony Muller, MBA ’71; and Tyler Sosin, BA ’08.

What difference does it make how data packets are routed through the internet? What techniques should be used to figure out staffing levels for telephone call centers? Can transactional data be used to guide booking decisions for hotels and airlines?

Welcome to the world of J. Michael Harrison, an operations researcher and the latest member of the Business School faculty to be elected to the prestigious National Academy of Engineering.

Harrison was cited for his contributions to knowledge about stochastic networks, a term used by scholars to describe complex systems in which statistical variation affects performance. In recent work, the Adams Distinguished Professor of Management focuses on applications to such varied problems as call center management, dynamic pricing, and revenue management.

Seated in his well-organized office and surrounded by books with titles such as The Elements of Statistical Learning and Monte Carlo Methods in Financial Engineering, Harrison observes that, like others in his field, “I live in a pretty abstract world. Most days I walk into this office thinking about things that are hard to see, hear, smell, or touch—like stockout penalties, demand distributions, and predictive relationships among business variables.”

Harrison called mathematical models, which provide frameworks within which to organize computations and data, his “stock in trade.”

“I specialize particularly in statistical models, where computations based on probability theory are used to account for uncertainty,” Harrison explained, adding that such models help corporate leaders decide “how many agents to employ in a call center, how big a plant to build, how much of each product to produce this month, whether and what kinds of insurance to buy.” Harrison said such decisions must be made “so that uncertainties are accounted for or hedged against in an economically reasonable way.”

While many theories developed by university researchers remain obscure, Harrison says that some have markedly altered the way people think about everyday issues. Many good examples can be found in the world of finance, where today’s language on Wall Street is vastly different from what it was three decades ago.

 “Words like ‘hedge’ and ‘option’ are much more commonly used than they were when I started on the GSB faculty, as are abstract phrases like ‘derivative asset’ and ‘correlation structure,’’’ said Harrison. “A good theory changes the words people use to describe what they are doing, and often their whole view of the world.”

Harrison’s theories are broadly framed, enabling them potentially to apply to business situations ranging from how to reduce long lines at McDonald’s to how to streamline production. “I’m usually trying not to make it specific—that’s my whole goal,” he said. “It’s like the parodic sales pitch that Dan Ackroyd did on Saturday Night Live long ago—‘It’s a floor wax, it’s a dessert topping, it’s all kinds of stuff.’ A theorist tries to address all possible cases in one fell swoop, to kill many birds with one stone.”

Despite his emphasis on the general, much of Harrison’s research over the past five years has been devoted to specific issues in call center management, collaborating with Stanford grads Assaf Zeevi, PhD ’01 in electrical engineering, and Achal Bassamboo, PhD ’05, who are now on the business school faculties at Columbia University and Northwestern University, respectively. “Call centers are an irresistible focus of study for people like us,” Harrison said, “with their echelons of differently trained agents, ubiquitous information technology, and highly structured operating procedures to handle contingencies. Still the biggest single focus in my work over the past 25 years has been development of general tools—tools that apply regardless of industry context—to support decision making about business systems.”

In addition to his work in operations management, Harrison is the coauthor of two influential papers in mathematical finance that developed the mathematical foundations of option theory. With colleague David Kreps, the Business School’s Theodore J. Kreps Professor of Economics, Harrison introduced the notion of equivalent martingale measures, which have since become a standard tool in theoretical analysis. 

Harrison’s newest initiatives are in the area of dynamic pricing and revenue management, also known as yield management. He introduced an MBA elective on that subject in 2003 and has launched a research project focused on design and pricing of subscription-based services like web hosting and outsourced logistics. 

Membership in the National Academy of Engineering honors those who have made outstanding contributions to engineering research, practice, or education, including “significant contributions to the engineering literature and to the pioneering of new and developing fields of technology, making major advancements in traditional fields of engineering. 

As he has done in alternate summers over the past 20 years, Harrison attended the weeklong Stochastic Networks Conference, in Paris in June. Scholars from various disciplines gathered to discuss problems in fields that range from biology to electronic commerce to wireless communications. Harrison was the principal organizer of a similar conference held at Stanford in 2002.

Other members of the Academy on the Business School faculty are Andy Grove, lecturer in business and cofounder of chip-maker Intel Corp, and William Miller, Herbert Hoover Professor of Public and Private Management, Emeritus.

 

• Back to Top