The Shell Game
How to Increase the Reliability of Your IT Infrastructure Using Predictive Analytics REGISTER >
The Shell Game
Among the first to expand on Kahn's technique and apply it to business planning was the planning unit at Royal Dutch/Shell Group. Since the end of World War II, both the price of oil and the growth in demand had been remarkably stable, and few oil executives had the foresight to imagine that things would ever change. By the late 1960s, Shell had developed a complex forecasting tool called the Unified Planning Machinery, used to predict growth in energy demand and upcoming oil prices. With that information, the company could generate strategies for investment in tankers and refineries as well as trading strategies. For years, growth in demand had cooperated nicely, edging up just ahead of the planners' expectations.
But beginning in the late 1960s, Shell's London-based planning group, led by Ted Newland and Pierre Wack, began generating scenarios that ventured beyond the expected, into political territory that no forecasting mechanisms at the time had accounted for. "Among the futures they imagined was one in which the rising power of Arab governments led to a catastrophic drop in the supply of oil and huge hikes in its price," says Dr. Paul Schoemaker, chairman and CEO of Decision Strategies International, Inc. and research director at the Wharton School's Emerging Technology Program. "Wack and Newland recognized that political logic in addition to economic logic would dictate price." While Shell's upper management found it compelling, the scenario proved difficult to communicate to the company's loosely connected line managers around the globe. At the time, Schoemaker says, supply and demand, in addition to long-term contracts, dictated price in the energy markets, with no exceptions, "so to suggest otherwise at the time was seen as being really out there."
Shell managers, according to Schoemaker, "listened politely to Newland and Wack and then shelved their warnings." But the Arab oil embargo of 1973 forced Shell's managers to revisit themand to incorporate scenario planning into their regular planning process. As a result, the company reduced the size of its oil tanker fleetforecasters had accurately predicted an overcapacity in the tanker industryand entered into longer-term contracts for oil, thereby locking in prices against what scenario planners felt would inevitably be politically-driven spikes. The effort paid off: Shell was better prepared for the oil shocks of 1979 and the subsequent mid-1980s collapse in price.
Typically, scenario planning uses a high-level team to determine the underlying drivers of a business and the fundamental uncertaintieseconomic, technological, political, social, demographic, environmentalthat surround a company and its markets. At best, the process is rigorous, systematic and ongoing. The planning team, enlisted from a variety of departmentsplanning, IT, finance, human resourcesand sometimes from outside the company, begins by identifying the key drivers of the business, along with trends in the industry, the market and the environment as a whole. The team then determines how the various forces interact, and works to understand the sources of uncertainty that will affect the company's future. For example, a key uncertainty might be unreliable energy supplies, which could have large implications for hardware design, centralization of servers or design of networks. By combining the forces and uncertainties in different ways, the team produces several divergent and internally consistent views of how the world might look in two or five or 10 years. Those scenarios are used to aid current strategy and decision making.
The tool is particularly well suited for dealing with major long-term decisions involving a high degree of uncertainty about the future. Will the growth in penetration of the Internet continue at the current pace? Will increased privacy concerns prove a drag on the growth of e-commerce? Will continued tight capital markets significantly slow down the build-out of high-bandwidth networks?