Learning in Action

5

Experimentation

Most approaches to learning accept the world as given. They begin with data that already exist—in the field, in the minds of customers, in accumulated experience—and then draw inferences and conclusions. The resulting lessons are often invaluable, as the previous two chapters have shown. But they are limited in an important respect. Because critical variables are taken at face value, with their ranges defined by natural variation, managers seldom consider the full array of alternatives or possible explanations. This is rarely a problem when challenges are conventional or a large knowledge base exists. But when unfamiliar concepts or unproven theories are involved, the desired data may first have to be produced. For real innovation ...