Strategic Innovation and the Science of Learning

The conventional planning process does not work for strategic experiments that are truly bleeding edge. Nevertheless, many companies cling to what they know — planning that holds managers responsible for numbers. But that is not practical for entering completely new territory, when numbers are essentially pulled out of a hat and their underlying assumptions rarely revisited. A better approach to planning comes from researchers at Dartmouth College’s Tuck School of Business. It emphasizes learning instead of numbers, and it draws on in-depth studies of such companies as New York Times Digital, Thomson Corp., Corning and Analog Devices.

Their approach, theory-focused planning, diverges from conventional planning in six critical ways. Companies that use it concentrate on a few critical unknowns instead of the usual horde of details in conventional plans; they focus on the theory underlining the predictions rather than the predictions themselves; they look for trends rather than numerical benchmarks; they review the plan often, in response to important new data, instead of annually; in that review, they consider the experiment over time instead of just for the current period; and they emphasize leading indicators rather than financials. Companies still hold managers of strategic experiments responsible for performance, but performance is gauged according to how quickly managers learn from new data. To be successful in uncharted waters, the ability to learn from experience is paramount.