Monday, June 29, 2009

How fast can we go?

Economists view organizations something like giant computers that solve problems by executing "programs" that take the form of step-by-step methods. These methods are broadly known as technologies and include "hard" technologies, like steelmaking, but also include the "soft" technologies of the service industries, including healthcare. Different types of organizational structures are graded against one another based primarily upon how well they can process information to solve problems for customers by providing appropriate products and services.

It is well known among economists who study the history of business that the decentralization of decision making within the organization improves an organization's ability to process information, and thus solve the customer's problem. This is the explanation offered for General Motor's victory over the Ford Motor Company in the 1920s, when GM's Alfred Sloan invented the multidivisional corporation by decentralizing decision making into GM's divisions: Chevrolet, Buick, Pontiack, Oldsmobile, Cadillack, and Fischer Body. Compared to the Ford Motor Company, which relied upon Henry Ford's personal problem solving capabilities, GM tapped into the information and ability of its divisional presidents as well as Alfred Sloan and his staff of management accountants. To put it simply, GM could simply "outhink" Ford. While Ford gave its customers "any color they wanted, so long as it was black," GM gave its customers "a car for every purse and purpose." Ford didn't recover until quite recently, when GM slipped into virtual (perhaps actual) bankruptcy.

Historians will argue about why, after 80 years, Ford regained its ascendancy over GM. But I like to think it's because Ford, under CEO Alan Mullally, has finally gotten serious about adopting the new organizational structure created by the real champ: Toyota. To understand why, let's take the computer analogy a step further.

In the language of computer science, GM's "decentralization" can be likened to "parallelization," the submission of a problem in multiple pieces to multiple processors. Henry Ford's decision making model can be likened to the use of a central processing unit. Successful parallelization of problem solving in computing leads to faster problem solving (and improved energy efficiency). In fact, using a result known as Amdahl's law, we can predict roughly how much faster our problem solving should be as a function of the extent to which we can parallelize it. Amdahl's law, named after computer architect Gene Amdahl, predicts that S = 1/(1 - P), where:

S = the rate of speedup; and

P = the extent of parallelization.

So, if P = 50%, S = 1/(1 - .5) = 1/.5 = 2. In other words, if we can parallelize (or decentralize) problem solving by 50%, we should expect problem solving to be twice as fast. By the same token, if we parallelize problem solving by 90%, we should expect it to be ten times as fast!

Back to Toyota. Experts maintain that Toyota has created a new organizational form that takes GM's original decentralization of decision making to the limit. Toyota has "radically decentralized" decision making by empowering front line workers to make decisions when problems are discovered. In fact, they can stop an entire factory if the problem is serious enough. Roughly speaking, GM's reorganization of business in the 1920s may have amounted to to a 50% "parallelization" of problem solving. It is not a stretch to say that GM was at least twice as fast as Ford. In 1927, in the face of rising sales of bigger, more powerful, more option-loaded Chevrolets, Ford shut down its giant new plant on the River Rouge for 10 months (!) to decommision the Model T and tool up for the more competitive Model A. It was clear that Ford simply could not adapt to the changing conditions of the Roaring Twenties (and later the Great Depression) as quickly as GM.

Toyota's reorgnaization may amount to a 90% "parallelization" of problem solving. This, then, explains the devasting effect of Toyota's accomplishment and why, despite years of attempting to adopt Toyota's production methods, almost all Western automakers have failed to compete against Toyota, except in particular geographical and product design-related niches. Western companies have been unwilling to transform their relatively centralized organizational structures and the now outdated management methods (notably management accounting) that once inventive GM created to run them.

Moreover, because Toyota's accomplishment is structural, it completely transcends industry boundaries. It isn't about automobilies at all. It's about how managers and employees and suppliers and customers share and process information. Thus, like GM's innovation four score years ago, Toyota's innovation in organizational structure can be--and already is being--generalized across manufacturing and service industries. The latest industry to catch fire, of course, is healthcare.