Idea 15 - Decentralization
As the
largest maker of automobiles for as long as most of us can remember, General
Motors looms very large in the world of manufacturing, even if its number one
ranking today is less than secure. It also looms large in the world of
management ideas, having been a touchstone for the concept of the modern
corporation for more than half a century. What GM did in the first half of the
20th century still reverberates in management practice, and its most
influential corporate innovation was decentralization.
Management has much to thank GM for - it gave them Peter
Drucker for a start. Under the leadership of Alfred P. Sloan, GM reinvented
both itself and organizational structure; though it was Drucker who analysed
what had been done and reflected upon it for the benefit of others in Concept
of the Corporation. Later, Sloan wrote about it too, in My Years with General
Motors, famously described by Microsoft's Bill Gates as 'probably the best book
to read if you want to read only one book about business'. It has been observed
that Concept of the Corporation established management as a discipline, but
Sloan established it as a profession.
Sloan was a rare creature - a boss who became recognized as
a management authority. Sloan joined the four-man GM executive team in 1920,
when the company was in serious trouble (he became CEO three years later).
Apart from anything else, GM was chaotically organized, having been created out
of 25 car makers and a number of component manufacturers. It was also going
bust.
First, spending was brought under control, with monthly
forecasts and central budgetary control- accounting and finance were never
decentralized at GM. Then came the first serious segmentation of the automobile
market, with each of five OM brands aimed at a different slice of the market.
Finally, each of the brands, along with three components operations, was made
into a separate division, giving them a degree of autonomy over their
operations. 'Structure follows strategy', as they teach in business school, and
this was, literally, the textbook example.
After
multi-functionalism at much the same time, though for different
reasons (growth and increasing complexity), DuPont had also divisionalized and
decentralized. The prevailing big-company structure at the time was what
business historian Alfred Chandler called 'multi-functionalism'. These were
hierarchical structures with management responsibilities specialized by
function. They had been developed in the mid-19th century by US railroad
companies to cope with their kaleidoscopic spread of functions -passenger
management, goods management, rolling stock, the rail network itself. 'Line'
management was, of course, invented on the railroads.
Divisionalization and decentralization pushed both
responsibility and decision-making closer to the coal face, to the people who
really knew what was going on. It motivated managers and eliminated much of the
time-consuming back-and-forth between operations and head office and, more
importantly, separated strategy from operations. Corporate executives had no
direct operating responsibilities, so couldn't interfere with matters they knew
little about. Instead, they made policy, free from the self-serving
participation of divisional executives, who were excluded from overall strategy
formulation.
The most important corporate function, Sloan believed, was
the allocation of resources. The new structure allowed return on investment to
be measured division by division. And that, in Sloan's words, enabled GM 'to
direct the placing of additional capital where it will result in the greatest
benefit to the corporation as a whole'.
This type of decentralized organization gradually became the
norm among large Western companies, who recognized the efficiency benefits of
'multidivisional' or - as it was later known - 'M-form' structure. Following
Sloan's lead, divisions in M-form companies have to compete with each other for
funds, showing that they deserve them with a combination of past performance
and future plans.
Some more
than others Today, some companies remain more centralized than
others. Japanese organizations are good at empowering factory workers, but less
so at corporate decentralizing. Their consensual style of decision-making makes
this less of an issue than it might be in the West. Certain oil and mining
companies keep their all-important technological expertise at the centre and,
with it, many of the decisions.
In spite of a trend towards team-building, many
'decentralized' organizations remain ultimately hierarchical. However, there
are voices calling for change. In his 1998 essay, 'Management's new paradigms',
Drucker insisted that it was a sound structural principle to have the fewest
number of management layers as possible - if only because the first law of
information theory tells us that 'every relay doubles the noise and cuts the
message in half'. He maintained that there was no one right way to organize a business,
however, noting that we had 'come to tout the team as the one right
organization for pretty much everything'. He thought that team members suffered
from split loyalties- to the team and to their own specialty function chief.
But, he said, there has to be one boss because, in a crisis, someone has to
give the orders and be obeyed. People have to learn to work in a team for one
task and in a command-and-control structure for another. Drucker was always
drawn to the middle way.
MIT Sloan School of Management's Tom Malone notes that
business continues to decentralize, but by a different route. Working from home
or telecommuting is growing on the back of new-generation IT. These new
low-cost communications give many people in big organizations enough information
to make sound decisions on their own, instead of taking orders from someone who
supposedly knows more than they do. And when people make their own decisions,
they are more motivated, creative and flexible.
Reference: 50 Management Ideas You Really Need to Know
Book by Edward Russell-Walling
Comments