I think I'm understanding what you are saying. The problem with rapid business expansion is the assumption of unlimited demand. The dilema of the modern 20th century is that we now know that our supply solutions are so effective that they can easily surpass demand.
In the 18 and early 19th century hand artisans could not exceed demand. Factory automation in the early 20 century taught us that it was now possible to exceed demand. The concept of market share became a new 20th century reality.
With the advent of computers it is possible to look at customers in a new way. As you mention, when the economy slackens, focus on the most profitable customers. Don't be afraid to let go of the weak customers. The problem is how to predict who will be the best customers two years from now?
Pareto's rule tells us that 80% of proffits are driven by 20% of customers. If this is true, if we can identify those 20% we can follow through on this idea you've presented of letting go of all but our best customers.
It seems counter intuitive that letting go of 80% of our customers could actually help the business. Wouldn't that mean losing 80% of business? The reality though is that we only lose 20% of business, but reduce our infrastructure expenses by almost 80%.
Do the numbers and ROI goes up by a factor of 7. Imagine life working with only your best and happiest customer. instead of expanding the factory or building a second factory, reducing the least productive customers is the key to building a stable business that functions in any economic situation.
Until the computer, it was impossible to accurately identfy the best customers. Now this task is automated and possible.
So Shigeo I appreciated your article that seems on the surface counter intuitive, but is actually the key to success in a limited market environment.