A
yy
SEPTEMBER-OCTOBER 1994
^ "I The Theory of the
Business
orei
y Peter R Drucke
N
ot in a very long time-not, perhaps, since the late 1940s
or early 1950s-have there heen as many new major man-
agement techniques as there are today: downsizing, out-
sourcing, total quality management, economic value
analysis, henchmarking, reengineering. Each is a powerful tool.
But, with the exceptions of outsourcing and reengineering, these
tools are designed primarily to do differently what is already heing
done. They are "how to do" tools.
Yet "what to do" is increasingly becoming the central challenge
facing managements, especially those of hig companies that have
enjoyed long-term success. The story is a familiar one: a company
that was a superstar only yesterday finds itself stagnating and frus-
trated, in trouble and, often, in a seemingly unmanageable crisis.
This phenomenon is hy no means confined to the United States. It
has hecome common in Japan and Germany, the Netherlands and
France, Italy and Sweden. And it occurs just as often outside husi-
ness-in labor unions, government agencies, hospitals, museums,
and churches. In fact, it seems even less tractahle in those areas.
The root cause of nearly every one of these crises is not that
things are being done poorly. It is not even that the wrong things
are being done. Indeed, in most cases, the right things are being
done-but fruitlessly. What accounts for this apparent paradox?
The assumptions on which the organization has been huilt and is
heing run no longer fit reality. These are the assumptions that
Peter F. Drucker is the Clarke Professor of Social Science and Manage-
ment at the Claremont Graduate School in Claremont, California,
where the Drucker Management Center was named in his honor. This is
Drucker's thirty-first article for HBR.
icies
HARVARD BUSINESS REVIEW September-October XXXXXXXXXX
THEORY OF THE BUSINESS
What underlies the
malaise of so manv
large and successful
organizations
worldwide is that thei
theory of the business no
longer ivorks.
shape any organization's hehavior, dictate its decisions about
what to do and what not to do, and define what the organization
considers meaningful results. These assumptions are about mar-
kets. They are about identifying customers and competitors, thei
values and hehavior. They are about technology and its dynam-
ics, about a company's strengths and weaknesses. These assump-
tions are about what a company gets paid for. They are what I call
a company's theory of the business.
Every organization, whether a husiness or not, has a theory of
the husiness. Indeed, a valid theory that is clear, consistent, and
focused is extraordinarily powerful. In 1809, for instance, German
statesman and scholar Wilhelm von Humboldt founded the Uni-
versity of Berlin on a radically new theory of the university. And
for more than 100 years, until the rise of Hitler, his theory defined
the German university, especially in scholarship and scientific
esearch. In 1870, Georg Siemens, the architect and first CEO of
Deutsche Bank, the first universal hank, had an equally clear the-
ory of the husiness: to use entrepreneurial finance to unify a still
ural and splintered Germany through industrial development.
Within 20 years of its founding, Deutsche Bank had hecome Eu-
ope's premier financial institution, which it has remained to this
day in spite of two world wars, inflation, and Hitler. And, in the
1870s, Mitsubishi was founded on a clear and completely new the-
ory of the business, which within 10 years made it the leader in an
emerging Japan and within another 20 years made it one of the
first truly multinational businesses.
Similarly, the theory of the husiness explains both the success
of companies like General Motors and IBM, which have dominat-
ed the U.S. economy for the latter half of the twentieth century,
and the challenges they have faced. In fact, what underlies the cur-
ent malaise of so many large and successful organizations world-
wide is that their theory of the husiness no longer works.
Whenever a hig organization gets into trouhle-and es-pecially if it has been successful for many years -people hlame sluggishness, complacency, a
ogance,mammoth bureaucracies. A plausihle explanation?
Yes. But rarely the relevant or co
ect one. Consider the two most
visible and widely reviled "a
ogant hureaueracies" among large
U.S. companies that have recently been in trouble.
Since the earliest days of the computer, it had been an article of
faith at IBM tbat the computer would go the way of electricity.
The future, IBM knew, and eould prove with scientific rigor, lay
with the central station, tbe ever-more-powerful mainframe into
which a buge numher of users could plug. Everything-economies,
the logic of information, technology-led to that conclusion. But
then, suddenly, when it seemed as if such a central-station, main-
frame-hased information system was actually coming into exis-
tence, two young men came up with the first personal computer.
Every computer maker knew that the PC was absurd. It did not
have tbe memory, the database, the speed, or the computing ahili-
ty necessary to succeed. Indeed, every computer maker knew that
the PC had to fail - the conclusion reached by Xerox only a few
years earlier, when its research team had actually built the first
PC. But when that misbegotten monstrosity-first the Apple, then
HARVARD BUSINESS REVIEW Sept ember-October 1994
the Macintosh-came on the market, people not only loved it, they
ought it.
Every big, successful company throughout history, wben con-
fronted with such a surprise, has refused to accept it. "It's a stupid
fad and will be gone in three years," said the CEO of Zeiss upon
seeing the new Kodak Brownie in 1888, when the German compa-
ny was as dominant in the world photographic market as IBM
would be in the computer market a century later. Most mainframe
makers responded in the same way. The list was long: Control
Data, Univac, Bu
oughs, and NCR in the United States; Siemens,
Nixdorf, Machines Bull, and ICL in Europe; Hitachi and Fujitsu in
Japan. IBM, the overlord of mainframes with as mucb in sales as
all the other computer makers put together and with record prof-
its, could have reacted in the same way. In fact, it should have. In-
stead, IBM immediately accepted the PC as the new reality. Al-
most overnight, it
ushed aside all its proven and time-tested
policies, rules, and regulations and set up not one but two compet-
ing teams to design an even simpler PC. A couple of years later,
IBM had become tbe world's largest PC manufacturer and the in-
dustry standard setter.
Tbere is absolutely no precedent for tbis acbievement in all of
usiness bistory; it hardly argues bureaucracy, sluggishness, or ar-
ogance. Yet despite unprecedented flexibility, agility, and humili-
ty, IBM was floundering a few years later in both tbe mainframe
and tbe PC business. It was suddenly unable to move, to take deci-
sive action, to cbange.
Tbe case of GM is equally perplexing. In tbe early 1980s-tbe
very years in wbicb GM's main business, passenger automobiles,
seemed almost paralyzed-the company acquired two large busi-
nesses; Hughes Electronics and Ross Perot's Electronic Data Sys-
tems. Analysts generally considered both companies to be mature
and chided GM for grossly overpaying for them. Yet, witbin a few
sbort years, GM bad more tban tripled tbe revenues and profits of
tbe allegedly mature EDS. And ten years later, in 1994, EDS had a
market value six times the amount that GM had paid for it and ten
times its original revenues and profits.
Similarly, GM bougbt Hughes Electronics - a huge but profitless
company involved exclusively in defense-just before the defense
industry collapsed. Under GM management, Hughes bas actually
Increased its defense profits and bas become tbe only big defense
contractor to move successfully into large-scale nondefense work.
Remarkably, the same bean counters who had been so ineffectual
in tbe automobile business-30-year GM veterans who had neve
worked for any otber company or, for that matter, outside of fi-
nance and accounting departments-were tbe ones wbo achieved
tbose startling results. And in the two acquisitions, they simply
applied policies, practices, and procedures tbat bad already been
used by GM.
Tbis story is a familiar one at GM. Since the company's found-
ing in a flu
y of acquisitions 80 years ago, one of its core compe-
tencies bas been to "overpay" for well-performing but mature
usinesses-as it did for Buick, AC Spark Plug, and Fisher Body in
tbose early years-and then turn them into world-class champi-
ons. Very few companies bave been able to matcb GM's perfor-
mance in making successful acquisitions, and GM surely did not
One oJGMls core
competencies has been
to ''overpay^ for well-
performing but mature
usinesses and then
turn them into world-
class champions.
HARVARD BUSINESS REVIEW Scpi cm her-Octoher XXXXXXXXXX
THEORY OF THE BUSINESS
The assumption that
a computer is a
computer-OK more
prosaically, that
the iiidustryis
hardware driven -
paralyzed IBM.
accomplish those feats hy heing hureaucratic, sluggish, or a
o-
gant. Yet what worked so heautifully in those husinesses that GM
knew nothing about failed miserably in GM itself.
What can explain the fact that at hoth IBM and GM thepolicies, practices, and behaviors that worked for de-cades-and in the case of GM are still working wellwhen applied to something new and different - no
longer work for the organization in which and for which they were
developed? The realities that each organization actually faces
have changed quite dramatically from those that each still as-
sumes it lives with. Put another way, reality has changed, hut the
theory of the business has not changed with it.
Before its agile response to the new reality of the PC, IBM had
once before turned its hasic strategy around overnight. In 1950,
Univac, then the world's leading computer company, showed the
prototype of the first machine designed to be a multipurpose com-
puter. All earlier designs had been for single-purpose machines.
IBM's own two earlier computers, huilt in the late 1930s and 1946,
espectively, performed astronomical calculations only. And the
machine that IBM had on the drawing hoard in 1950, intended fo
the SAGE air defense system in the Canadian Arctic, had only one
purpose: early identification of enemy aircraft. IBM immediately
scrapped its strategy of developing advanced single-purpose ma-
chines; it put its hest engineers to work on perfecting the Univac
architecture and, from it, designing the first multipurpose com-
puter ahle to he manufactured (rather than handcrafted) and ser-
viced. Three years later, IBM had become the world's dominant
computer maker and standard-bearer. IBM did not create the com-
puter. But in 1950, its flexibility, speed, and humility created the
computer industry.
However, the same assumptions that had helped IBM prevail in
1950 proved to he its undoing 30 years later. In the 1970s, IBM as-
sumed that there was such a thing as a "computer," just as it had
in the 1950s. But the emergence of the PC invalidated that as-
sumption. Mainframe computers and PCs are, in fact, no more one
entity than arc generating stations and electric toasters. The lat-
ter, while different, are interdependent and complementary. In
contrast, mainframe computers and PCs are primarily competi-
tors. And, in their hasic definition of information, they actually
contradict each otber: for the rnainframe, information means
memory; for the hrainless PC, it means software. Building gener-
ating stations and making toasters must he run as separate husi-
nesses, but tbey can he owned by the same corporate entity, as
General Electric did for decades. In contrast, mainframe comput-
ers and PCs prohahly catinot coexist in the same corporate entity.
IBM tried to combine the two. But hecause the PC was the
fastest growing part of the husiness, IBM could not suhordinate it
to the tiiainframe business. As a result, the company could not op-
timize the mainframe business. And hecause the mainframe was
still the cash cow, IBM could not optimize the PC business. In the
end, the assumption that a computer is a computer-or, more pro-
saically, that the industry is hardware driven-paralyzedlBM.
GM had an even more powerful, and successful, theory of the
husiness than IBM had, one that made GM the world's largest and
HARVARD BUSINESS REVIEW September-October 1994
most profitable manufacturing organization. The company did
not have one sethack in 70 years-a record unmatched in business
history. GM's theory comhined in one seamless web assumptions
ahout markets and customers with assumptions about core com-
petencies and organizational structure.
Since the early 1920s, GM assumed that the U.S. automohile
market was homogeneous in its values and segmented hy ex-
tremely stahle income groups. The resale value of the "good" used
car was the only independent variable under management's con-
trol. High trade-in values enahled customers to upgrade their new-
car purchases to the next category - in other words, to cars with
higher profit margins. According to this theory, frequent or radical
changes in models could only depress trade-in values.
Internally, these market assumptions went hand in hand with
assumptions ahout how production should be organized to yield
the biggest market share and the highest profit. In GM's case, the
answer was long runs of mass-produced cars with a minimum of
changes each model year, resulting in tbe largest numher of uni-
form yearly models on tbe market at the lowest fixed cost