Maximum Earnings

March 16, 2007

Angels on a Pin

In my reply to John Dawson's insightful comments on "Eleven Cents Ain't Much" I said that:

In 1991 just 1/100th of IBM's next market share point had an incremental cost of $15.6 million. The earnings from that market share basis point after SG&A expenses were $4.9 million. In other words, acquiring just 1/100th of next revenue share point theoretically would have cost IBM $10.7 million in lost earnings after SG&A expenses.

How do I know this? "The Rule of Maximum Earnings." This rule is based on the most widely known (and seldom applied) principle of microeconomics.

Unless you studied in a seminary you almost certainly took a microeconomics course in college. You, like the great majority of "micro" students probably took the course only because you had to. If so you may have thought marginal cost and revenue curves were useful only in determining –- theoretically -- how many angels could dance on the head of a pin ... before one falls off! You may even have been among those who found micro mildly interesting. I confess I was one of the very few students in my class who loved it. This is all the more surprising because I was a fine art major.


Microeconomics was the first course I took in my masters program. The way my prof made the marginal cost and revenue graphs dance across the blackboard was positively mesmerizing. Maybe it was just the steamy August heat in that LSU classroom without an AC. Or maybe it's because I was an art major that these graphs appeared to me to be more than just abstractions. They were abstractions with meaning! The intersection of these schedules identified the sweet spot in any measurable combination of costs and benefits. If I spent more at the margin then the benefits returned I was throwing money down the drain. If I spent less then marginal returns I was leaving money on the table. What an incredibly simple and powerful idea. One that even an art major could understand!


So why are marginal cost and revenue not part of the CMO's tool kit? There were three bumps in the road. The first bump was that marginal cost and revenue schedules are pure theory. Only in the last twenty years were methods developed to apply the theory to the real world of marketing.

The second bump? The methods were complicated. You couldn't apply the theory unless you were an econometrician working for a company in a highly regulated market -- like alcoholic beverages, pharmaceuticals, and tobacco -- or in closely measured markets like groceries, travel and entertainment -- because you didn't have the detailed data on competitors.

The third bump was the biggest. Even when you cleared the first two you ran up against a brick wall. You had no way to factor in the impact of competitive spending on your company's earnings. In other words, you couldn't calculate IBM's incremental cost and earnings per market share point.

If you're interested in finding out exactly how I ran the numbers for IBM (and Dell) check out my narrated Breeze presentation on Chapter 6 "The Battle for Your Desktop." And if you have questions, remember I'm just a comment away.