We’ve all heard many times the idea that “people will do what they’re measured on.” Is it really true? If I told you that the defect rate for this company’s premier product dropped from 12% to 1% in only three weeks and then further declined to 0.5% in just three more weeks, you’d probably think I was making it up. Well, that is exactly the result that occurred—and it occurred solely from beginning to measure quality in parallel with measuring productivity (in this case, speed). Perhaps what is even more revealing is that there was no incentive linked to quality improvement. All the while, workers were being paid a bonus for speed—which, incidentally, did not suffer when quality improved. That they were not being equally rewarded for quality made no difference. The metrics alone changed their behavior.