First, a company had to demonstrate the good-to-great pattern independent of its industry; if the whole industry showed the same pattern, we dropped the company. Second, we debated whether we should use additional selection criteria beyond cumulative stock returns, such as impact on society and employee welfare. We eventually decided to limit our selection to the good-to-great results pattern, as we could not conceive of any legitimate and consistent method for selecting on these other variables without introducing our own biases.
