The best discipline, maybe the only discipline that really works, is self- discipline.

One practical point many experts will attest to is that if you want to develop someone as a leader, give them lots of responsibility early in their lives and careers. The military does that. I can remember being officer of the deck on a destroyer, on watch and in charge at two in the morning as we plowed through the Mediterranean while 300 shipmates slept below decks. I was 25 at the time. I don't know how much of a leader I ever became, but the experience certainly brought home to me a sense of responsibility for others.

The experience curve says that your costs should probably decline by 15% or 20% with every doubling in your experience making a product, approximately how many of them you turn out. It also says that if you have the biggest market share, meaning the most experience of anybody in your competitive set, you should have the lowest costs, and the resultant capability to underprice your competitors, maybe forever. The abiding lesson of the experience curve is that companies need to discipline themselves to keep reducing their costs, year in, year out, if they are to remain competitive.

People believe that management consultants are mostly useless parasites. Up until about 1980 it was consultants more than anyone else who came up with the critical concepts behind strategy. The history of strategy suggests there are lots of things consultants can do for a company that the company can't typically do for itself.

I argue that once it became clear that the most important function of the CEO was to develop and enact the corporate strategy, that often had the effect of distancing him from people below him in the organization. It also encouraged the idea that if a CEO were a great strategist for a company in one industry, he would probably be a great strategist in another industry. And that usually hasn't proved to be the case.

The best discipline, maybe the only discipline that really works, is self- discipline.

The business schools could do a better job teaching face-to-face management, the actual work of organizing and helping along the efforts of others in the organization. The more quantitative disciplines have gotten more attention, often more research dollars. Areas like organizational science or, even mushier, leadership have had more trouble settling on what it's important to teach, and how. It's rather like strategy itself, which as I argue in the book, has had trouble through most of its history figuring out how to incorporate people, their motivation and ability, into its calculations.