The art of high-technology

management
Modesto A. Maidique and Robert H. Hayes

Despite some recent skepticism, there's plenty of evidence

that many US companies excel at managing technological
innovation - and that, accordingly, others need not go far
afield in search of inspiring role models. Surveyingfirmsin
five industries rated as "high-tech," the authors sought to
identify the common ingredients of success. Their findings,
often counterintuitive but broadly applicable, outline six
areas of special strength: business focus, adaptability,
organizational cohesion, entrepreneurial culture, sense of
integrity and "hands-on" management.

Over tbe past 15 years, tbe world's perception of tbe competence of

US companies in managing tecbnology bas come full circle. In 1967,
Jean-Jacques Servan-Scbreiber reported witb alarm in bis book
The American Challenge tbat US tecbnology was far abead of tbe
rest of tbe industrialized world. Tbis "tecbnology gap," be argued,
was continually widening because of tbe superior ability of Amer-
icans to organize and manage tecbnological development.

Today, tbe situation is perceived to bave cbanged drastically. Tbe

concern now is tbat tbe gap is reversing: tbe onslaugbt of Japanese
and/or European cballenges is tbreatening America's tecbnological
leadersbip. In bis book, America's Technology Slip, Dr Simon Ramo
notes tbe apparent inability of US companies to compete tecbnologi-
cally witb tbeir foreign counterparts. Moreover, in tbeir The Art of
Japanese Management, Ricbard Pascale and Antbony Atbos use as
a basis of comparison two tecbnology-based firms, Matsusbita and
ITT. Here, tbe Japanese firm is depicted as a model for managers,
wbile tbe management practices of tbe American firm are sbarply
criticized.

Nevertbeless, a number of US companies appear to be successfully

fending off tbese foreign cballenges. Tbese firms are repeatedly

SUMMER 1985 43
included on lists of "America's best-managed companies." Many of
tbem are competitors in the R&D-intensive industries, a sector of
the economy that has come under particular criticism. Ironically,
some of them have even served as models for highly successful
Japanese and European high-tech firms.

For example, of the 43 companies that Tom Peters and Robert H.

Waterman, Jr., judged to be "excellent" in In Search of Excellence,
almost balf were classified as "bigb technology," or as containing a
substantial bigb-tecbnology component.* Similarly, of tbe five US
organizations tbat William Oucbi, autbor of Theory Z, described as
best prepared to meet tbe Japanese cballenge, tbree (IBM, Hewlett-
Packard and Kodak) were bigb-technology companies. Indeed,
high-technology corporations are among the most admired firms in
America. In a Fortune study that ranked the corporate reputation of
the 200 largest US corporations, IBM and Hewlett-Packard ranked
first and second, respectively.^ And ofthe top 10firms,nine compete
in such high-technologyfieldsas Pharmaceuticals, precision instru-
ments, communications, office equipment, computers, jet engines
and electronics.

The above studies reinforce our own findings, wbicb bave led us to
conclude tbat US bigb-tecbnology firms seeking to improve tbeir
management practices to succeed against foreign competitors need
not look overseas. Tbe firms cited above are not unique. On tbe
contrary, tbey are representative of scores of well-managed small
and large US tecbnology-based firms. Moreover, tbe management
practices tbey bave adopted are widely applicable. Tbus, perbaps
tbe key to stimulating innovation in America is not to adopt tbe
managerial practices of the Europeans or Japanese, but to adapt
some of tbe policies of successful American bigb-tecbnology firms.

The study approach

Over tbe past two decades, we bave been privileged to work with a
host of small and large high-technology firms as participants,
advisors and researchers. We and our assistants interviewed
formally and informally over 250 executives, including over 30
* For the purpose of this article, the high-technology industries are defined as those that
spend more than 3 percent of sales on R&D. These industries, though otherwise quite
different, are all characterized by a rapid rate ofehange in their products and technologies.
Only five US industries meet this criterion: (1) chemicals and Pharmaceuticals; (2)
machinery (especially computers and office machines); (3) electrical equipment and
communications; (4) professional and scientific instruments; and (5) aircraft and missiles.
See National Science Foundation, Science Resources Studies Highlights, NSF81-331.
December31,1981,p.2.
' References appear on page 62.

44 THEMcKINSEYQUARTERLY
CEOs, from a wide cross-section of high-tech industries - biotecb-
nology, semiconductors, computers, pbarmaceuticals and aero-
space. About 100 of tbese executives were interviewed in 1983 as
part of a large-scale study of product innovation in tbe electronics
industry conducted by one of tbe autbors and his colleagues at
Stanford University. Our research has been guided by a fun-
damental question: Wbat are tbe strategies, policies, practices and
decisions tbat result in successful management of bigb-tecbnology
enterprises? One of our principal findings was tbat no company bas
a monopoly on managerial excellence. Even the best-run companies
make big mistakes, and many small, lesser regarded companies are
surprisingly sopbisticated about tbe factors tbat mediate between
success and failure.
It also became apparent from our interviews tbat tbe driving force
bebind tbe successes of many of tbese companies was strong
leadersbip. All companies need leaders and visionaries, of course,
but leadersbip is particularly essential wben tbe future is blurry
and wben tbe world is cbanging rapidly. Altbougb few bigb-tecb
firms can succeed for long witbout strong leaders, leadersbip itself
is not tbe subject of tbis article. Ratber, we accept it as given and
seek to understand wbat strategies and management practices can
reinforce strong leadersbip.

Tbe companies we studied were of different sizes ($10 million to $30

billion in sales); tbeir tecbnologies were at different stages of
maturity; tbeir industry growtb rates and product mixes were
different; and tbeir managers ranged widely in age. But tbey all bad
tbe same unifying tbread: rapid cbange in tbe tecbnological base of
tbeir products, implying novel products and functions and tbus
usually rapid growtb. But even wben growtb is slow or moderate,
tbe destruction of tbe old capital base by new tecbnology results in
tbe need for rapid redeployment of resources to cope witb new
product designs and new manufacturing processes. Tbus, tbe two
dominant cbaracteristics of tbe bigb-tecbnology organizations tbat
we focused on were growtb and cbange.

In part because of tbis split focus, tbe companies we studied often

appeared to display contradictory bebavior over time. Despite tbese
differences, in important respects tbey were remarkably similar
because tbey all confronted tbe same two-beaded dilemma: bow to
unleasb tbe creativity tbat promotes growtb and cbange witbout
being fragmented by it, and bow to control innovation witbout
stifiing it. In dealing witb tbis concern, tbey tended to adopt
strikingly similar managerial approacbes.

SUMMER 1985 45
Six themes of success
Wben we grouped our findings into general tbemes of success, a
sig^iificant paradox gradually emerged - wbicb is a product of the
unique challenge that high-technology firms face. Some of the
bebavioral patterns tbat tbese companies displayed seemed to favor
promoting disorder and informality, wbile otbers would bave led us
to conclude tbat consistency, continuity, integration and order were
tbe keys to success. As we grappled witb tbis apparent paradox, we
came to realize tbat continued success in a bigb-tecbnology en-
vironment requires periodic sbifts between cbaos and continuity.
Our originally static framework, tberefore, was gradually replaced
by a dynamic framework witbin wbose ebbs and fiows lay tbe
secrets of success.

Tbe six tbemes tbat we grouped ourfindingsinto were: (1) business

focus; (2) adaptability; (3) organizational cohesion; (4) entrepreneur-
ial culture; (5) sense of integrity; and (6) "hands-on" top manage-
ment. No one firm exhibits excellence in every one of tbese
categories at any one time, nor are tbe less successful firms totally
lacking in all. Nonetbeless, outstanding bigb-tecbnologyfirmstend
to score bigb in most of tbe six categories, wbile less successful ones
usually score low in several. Let us consider eacb of tbem in turn.

1. Business focus
Even a superficial analysis of tbe most successful bigb-tecbnology
firms leads one to conclude tbat tbey are bigbly focused. Witb few
exceptions, tbe leaders in bigb-tecbnologyfieldssucb as computers,
aerospace, semiconductors, biotecbnology, cbemicals, pbarma-
ceuticals, electronic instruments and duplicating macbines realize
tbe great bulk of tbeir sales eitber from a single product line or from
a closely related set of product lines. For example, IBM, Boeing,
Intel and Genentecb confine tbemselves almost entirely to compu-
ter products, commercial aircraft, integrated circuits and genetic
engineering, respectively. Similarly, four-fiftbs of Kodak's and
Xerox's sales come from pbotograpbic products and duplicating
macbines, respectively. In general, tbe smaller tbe company, tbe
more bigbly focused it is. Tandon concentrates on disk drives;
Tandem on bigb-reliability computers; Analog Devices on linear
integrated circuits; and Cullinet on software products.
Closely related products. Tbis extraordinary concentration does not
stop witb tbe dominant product line. Wben tbe company grows and
establisbes a secondary product line, it is usually closely related to
tbe first. Hewlett-Packard, for instance, bas two product families.

46 THEMcKINSEYQUARTERLY
eacb accounting for about balf of its sales. Botb families - electronic
instruments and data processors - are focused on tbe same
tecbnical, scientific and process control markets. IBM also makes
two closely related product lines - data processors (approximately
80 percent of sales) and office equipment - botb of wbicb empbasize
the business market.

Companies tbat took tbe opposite patb bave not fared well. Two of
yesterday's tecbnological leaders, ITT and RCA, bave paid dearly
for diversifying away from tbeir strengtbs. Today, botb firms are
trying to divest many of wbat were once bigbly touted acquisitions.
As David Packard, cbairman of tbe board of Hewlett-Packard, once
observed: "No company ever died from starvation, but many bave
died from indigestion."
A communications firm tbat became tbe world's largest conglomer-
ate, ITT began to slip in tbe early 1970s after an acquisition wave
orcbestrated by Harold Geneen. Wben Geneen retired in 1977, bis
successors attempted to redress ITT's lackluster performance
tbrougb a far-reacbing divestment program.^ So far, 40 companies
and otber assets wortb over $1 billion bave been sold off- and ITT
watcbers believe tbe program is just getting started. Some analysts
believe tbat ITT will ultimately be restructured into three groups,
with the communications/electronics group and engineered prod-
ucts (home of ITT semiconductors) forming the core of a "new" ITT.

RCA experienced a similar fate. When RCA's architect and long-

time chairman. General David Sarnoff, retired in 1966, RCA was
internationally respected for its pioneering work in television,
electronic components, communications and radar. But by 1980, tbe
tbree CEOs who followed Sarnoff had turned a technological leader
into a conglomerate with fiat sales, declining earnings, and a $2.9
billion debt. Tbis disappointing performance led RCA's new CEO,
Tbornton F. Bradsbaw, to decide to return RCA to its bigb-
tecbnology origins.^ Bradsbaw's strategy is now to concentrate on
RCA's traditional strengtbs - communications and entertainment
- by divesting its otber businesses.

Focused R&D. Anotber policy tbat strengtbens tbe focus of leading

bigb-tecbnology firms is concentrating R&D on one or two areas.
Sucb a strategy enables tbese businesses to dominate tbe researcb,
particularly tbe more risky, leading-edge explorations. By spending
a bigber proportion of tbeir sales dollars on R&D tban tbeir
competitors do, or tbrougb tbeir sbeer size (as in tbe case of IBM,
Kodak and Xerox), sucb companies maintain tbeir tecbnological

SUMMER 1985 47
leadersbip. It is not unusual for a leading firm's R&D investment to
be one-and-a-balf to two times tbe industry's average as a percen-
tage of sales (8 to 15 percent) and several times more tban
any individual competitor on an absolute basis.^

Moreover, tbeir commitment to R&D is botb enduring and consist-

ent. It is maintained tbrougb slack periods and recessions because it
is believed to be in tbe best long-term interest of tbe stockbolders.
As tbe CEO of Analog Devices, a leading linear integrated circuit
manufacturer, explained in a quarterly report wbicb noted that
profits had declined 30 percent: "We are sharply constraining the
growth of fixed expenses, but we do not feel it is in tbe best interest
of sbareholders to cut back furtber on product development... in
order to relieve sbort-term pressure on earnings."^ Similarly, wben,
as a result of a recession, sales fiattened and profit margins
plummeted at Intel, its management invested a record-breaking
$130 million in R&D, and anotber $150 million in plant and
equipment.^

Consistent priorities. Still anotber way tbat a company demon-

strates a strong business focus is tbrougb a set of priorities and a
pattern of bebavior tbat is continually reinforced by top manage-
ment: for example, planned manufacturing improvement at Texas
Instruments; customer service at IBM; tbe concept of tbe entre-
preneurial product cbampion at 3M; and new products at Hew-
lett-Packard. Belief in tbe competitive effectiveness of tbeir cbosen
tbeme runs deep in eacb of tbese companies.

A business focus that is maintained over extended periods of time

has fundamental consequences. By concentrating on what it does
well, a company develops an intimate knowledge of its markets,
competitors, technologies, employees, and of tbe future needs and
opportunities of its customers. Tbe Stanford Innovation Project
recently completed a tbree-year study of 224 US bigb-tecbnology
products (balf of wbicb were successes, balf of wbicb were failures)
and concluded that continuous, in-depth, informal interaction witb
leading customers tbrougbout the product development process was
the principal factor bebind successful new products.

In sbort, this coupling is tbe cornerstone of effective bigb-tecbno-

logy progress. Sucb an interaction is greatly facilitated by tbe
longstanding and close customer relationsbips tbat are fostered by
concentrating on closely related product-market cboices. "Custom-
er needs," explains Tom Jones, cbairman of Nortbrop Corporation,
"must be understood way abead of time."

48 THEMcKINSEYQUARTERLY
2. Adaptability
Successful firms balance a well-defined business focus witb tbe
willingness, and tbe will, to undertake major and rapid cbange
wben necessary. Concentration, in sbort, does not mean stagnation.
Immobility is tbe most dangerous bebavioral pattern a bigb-
tecbnology firm can develop; tecbnology can cbange rapidly, and
witb it tbe markets and customers served. Tberefore, a high-
technology firm must be able to track and exploit tbe rapid sbifts
and twists in market boundaries as tbey are redefined by new
tecbnological, market and competitive developments.

Tbe cost of strategic stagnation can be great, as General Radio (GR)

found out. Once the proud leader of the electronic instruments
business, GR almost single-bandedly created many sectors of tbe
market. Its engineering excellence and its progressive buman
relations policies were models for tbe industry. But wben its
founder, Melville Eastbam, retired in 1950, GR's strategy ossified.
In tbe next two decades, tbe company failed to take advantage of
two major opportunities for growtb tbat were closely related to tbe
company's strengtbs: microwave instruments and minicomputers.
Meanwbile, its traditional product line witbered away. Now all that
remains of GR's once dominant instruments line, which is less than
10 percent of sales, is a small assembly area wbere a bandful of
tecbnicians assemble batcbes of tbe old instruments.

It wasn't until William Tburston, in tbe wake of mounting losses,

assumed tbe presidency at tbe end of 1972 tbat GR began to refocus
its engineering creativity and couple it to its new marketing
strategies. Using tbe failure of tbe old policies as bis mandate,
Tburston deempbasized tbe aging product lines, focused GR's
attention on automated test equipment, balanced its traditional
engineering excellence witb an increased sensitivity to market
needs, and gave tbe firm a new name — GenRad. Since tben, GenRad
bas resumed rapid growtb and bas won a leadersbip position in tbe
automatic test equipment market.
Tbe GenRad story is a classic example of a firm making a strategic
cbange because it perceived tbat its existing strategy was not
working. But even successful bigb-tecbnology firms sometimes feel
tbe need to be rejuvenated periodically to avoid tecbnological
stagnation. In tbe mid-1960s, for example, IBM appeared to bave
little reason for major cbange. Tbe company bad a near monopoly in
tbe computer mainframe industry. Its two principal products - tbe
1401 at tbe low end of tbe market and tbe 7090 at tbe bigb end —
accounted for over two-tbirds of its industry's sales. Yet in one

SUMMER 1985 49
move tbe company obsoleted botb product lines (as well as otbers)
and redefined tbe rules of competition for decades to come by
simultaneously introducing six compatible models of tbe "System
360," based on proprietary bybrid integrated circuits.
During tbe same period, GM, wbose dominance of tbe US auto
industry approacbed IBM's dominance of tbe computer mainframe
industry, stoutly resisted sucb a rejuvenation. Instead, it became
more and more centralized and infiexible. Yet, GM was also once a
bigb-tecbnology company. In its early days wben Alfred P. Sloan
ran tbe company, engines were viewed as bigb-tecbnology products.
One day, Cbarles F. Kettering told Sloan be believed tbe bigb
efficiency of tbe diesel engine could be engineered into a compact
power plant. Sloan's response was: "Very well - we are now in tbe
diesel engine business. Y^ou tell us bow tbe engine sbould run, and I
will ... capitalize tbe program."^ Two years later, Kettering
acbieved a major breaktbrougb in diesel tecbnology. Tbis paved tbe
way for a revolution in tbe railroad industry and led to GM's
preeminence in tbe diesel locomotive markets.

Organizational flexibility. To undertake sucb wrencbing sbifts in

direction requires botb agility and daring. Organizational agility
seems to be associated witb organizational fiexibility - frequent
realignments of people and responsibilities as tbe firm attempts to
maintain its balance on sbifting competitive sands. Tbe daring and
tbe willingness to take "you bet your company" kinds of risks is a
product of botb tbe inner confidence of its members and a powerful
top management - one tbat eitber bas effective sbarebolder control
or tbe full support of its board.

3. Organizational cohesion
Tbe key to success for a bigb-tecb firm is not simply periodic
renewal. Tbere must also be cooperation in tbe translation of new
ideas into new products and processes. As Ken Fisber, tbe arcbitect
of Prime Computer's extraordinary growtb, puts it: "If you bave tbe
driving function, tbe most important success factor is tbe ability to
integrate. It's also tbe most difficult part of tbe task."

To succeed, tbe energy and creativity of tbe wbole organization

must be tapped. Anytbing tbat restricts tbe fiow of ideas, or
undermines tbe trust, respect and sense of common purpose among
individuals is a potential danger. Tbis is wby bigb-tecb firms figbt
so vigorously against tbe usual organizational accoutrements of
seniority, rank and functional specialization. Little attention is

50 THEMcKINSEYQUARTERLY
given to sucb mecbanisms as organizational cbarts: quite often tbey
simply don't exist.

Younger people in a rapidly evolving tecbnological field are often as

good - and sometimes even better - a source of new ideas as are
older ones. Some bigb-tecb firms, in fact, apply tbe notion of a
"balf-life of knowledge" - tbe amount of time tbat bas to elapse
before balf of wbat one knows is obsolete. In semiconductor
engineering, for example, it is estimated tbat tbe balf-life of a newly
minted PbD is about seven years. Tberefore, any practice tbat
relegates younger engineers to secondary, nonpartnersbip roles is
considered counterproductive.

Similarly, product design, marketing and manufacturing personnel

must collaborate in a common cause ratber tban compete witb one
anotber, as bappens in many organizations. Any policies tbat
appear to elevate one of tbese functions above tbe otbers can poison
tbe atmospbere for collaboration and cooperation.

A source of division, and one wbicb distracts tbe attention of people

from tbe needs of tbe firm to tbeir own aggrandizement, are tbe
executive "perks" found in many mature organizations: pretentious
job titles, separate dining-rooms and rest-rooms for executives,
larger and more luxurious offices (often separated in some way from
tbe rest of tbe organization), and even separate or reserved places in
tbe company parking lot all tend to establisb "distance" between
managers and doers, and substitute artificial goals for tbe crucial
real ones of creating successful new products and customers. Tbe
appearance of an executive dining-room, in fact, is one of tbe
clearest danger signals.

Good communication. One way to combat tbe development of sucb

distance is by making top executives more visible and accessible.
IBM, for instance, bas an open-door policy tbat encourages mana-
gers at different levels of tbe organization to talk to department
beads and vice presidents. According to senior IBM executives, it
was not unusual for a project manager to drop in and talk to Frank
Cary (IBM's cbairman) or Jobn Opel (IBM's president) until Cary's
recent retirement. Likewise, an office witb transparent walls and
no door, sucb as tbat of Jobn Young, CEO at Hewlett-Packard,
encourages communication. In fact, open-style offices are common
in many bigb-tecb firms.

A regular feature of 3M's management process is tbe montbly

Tecbnical Forum, wbere tecbnical staff members from tbe firm

SUMMER 1985 51
excbange views on tbeir respective projects. Tbis empbasis on
communication is not restricted to internal operations. Sucb a firm
supports and often sponsors industry-wide tecbnical conferences,
sabbaticals for staff members and cooperative projects witb tecbni-
cal universities.

Tecbnical Forums serve to compensate partially for tbe loss of

visibility tbat tecbnologists usually experience wben an organiza-
tion becomes more complex and wben production, marketing and
finance staffs swell. So does tbe concept of tbe dual-career ladder
tbat is used in most of tbese firms; tbat is, a job bierarcby tbrougb
wbicb tecbnical personnel can attain tbe status, compensation and
recognition tbat is accorded to a division general manager or a
corporate vice president. By using tbis strategy, companies try to
retain tbe spirit of tbe early days of tbe industry wben scientists
played a dominant role, often even serving on tbe board.
Again, a strategic business focus contributes to organizational
cobesion. Managers of firms tbat bave a strong tbeme/culture and
concentrate on closely related markets and tecbnologies generally
display a sopbisticated understanding of tbeir businesses. Someone
wbo understands wbere tbe firm is going and wby is more likely to
be willing to subordinate tbe interests of bis or ber own unit or
function in tbe interest of promoting tbe common goal.

Job rotation. A policy of conscious job rotation also facilitates tbis

sense of community. In tbe small firm, everyone is involved in
everyone else's job: specialization tends to creep in as size increases
and boundary lines between functions appear. If left uncbecked,
tbese boundaries can become rigid and impermeable. Rotating
managers in temporary assignments across tbese boundaries belps
keep tbe lines fiuid and informal, bowever. Wben a new process is
developed at Texas Instruments, for example, tbe process develop-
ers are sent to tbe production unit wbere tbe process will be
implemented. Tbey are allowed to return to tbeir usual posts only
after tbat unit's operations manager is convinced tbat tbe process is
working properly.
Integration of roles. Otber ways in wbicb bigb-tecb companies try to
prevent organizational, and particularly bierarcbical, barriers
from rising are tbrougb multidisciplinary project teams, "special
venture groups" and matrix-like organizational structures. Sucb
structures, wbicb require functional specialists and product/
market managers to interact in a variety of relatively sbort-term
problem-solving assignments, botb inject a certain ambiguity into

52 THE McKINSEY QUARTERLY

organizational relationsbips and require eacb individual to play a
variety of organizational roles.

For example, AT&T uses a combination of organizational and

pbysical mecbanisms to promote integration. Tbe Advanced De-
velopment sections of Bell Labs are pbysically located on tbe sites of
tbe Western Electric plants. Tbis creates an organizational bond
between development activities and Bell's basic researcb, and an
equally important spatial bond between development and tbe
manufacturing engineering groups at tbe plants. In tbis way,
tbree-way communication is encouraged.

Long-term employment. Long-term employment and intensive

training are also important integrative mecbanisms. Managers and
tecbnologists are more likely to develop satisfactory working
relationsbips if tbey know tbey will be barnessed to eacb otber for a
good part of tbeir working lives. Moreover, tbeir loyalty and
commitment to tbe firm is increased if tbey know tbe firm is
continuously investing in upgrading tbeir capabilities.

At Tandem, tecbnologists regularly train administrators on tbe

performance and function of tbe firm's products and, in turn,
administrators train tbe tecbnologists in personnel policies and
financial operations. Sucb a firm also tends to select college
graduates wbo bave excellent academic records, wbicb suggest
self-discipline and stability, and tben encourages tbem to stay witb
tbe firm for most, if not all, of tbeir careers.

4. Entrepreneurial culture
Wbile continuously striving to pull tbe organization togetber,
successful bigb-tecb firms also display fierce activism in promoting
internal agents of cbange. Indeed, it bas long been recognized tbat
one of tbe most important cbaracteristics of a successful bigb-
tecbnology firm is an entrepreneurial culture.

Tbe ease witb wbicb small entrepreneurial firms innovate bas

always inspired a mixture of puzzlement and jealousy in larger
firms. Wben new ventures and small firms fail, tbey usually do so
because of capital sbortages and managerial errors. Nonetbeless,
time and again tbey develop remarkably innovative products,
processes and services witb a speed and efficiency tbat baffle tbe
managers of large companies. Tbe success of tbe Apple II, wbicb
created a new industry, and Genentecb's genetically engineered
insulins are of tbis genre. Tbe explanation for a small entrepreneur-

SUMMER1985 53
ial firm's innovativeness is straigbtforward, yet it is difficult for a
largefirmto replicate its spirit.

Entrepreneurial characteristics. First, tbe small firm is tj^ically

blessed witb excellent communication. Its tecbnical people are in
continuous contact (and often in cramped quarters). Tbey bave
luncb togetber, and tbey call eacb other outside working hours.
Thus, they come to understand and appreciate the difficulties and
challenges facing one another. Sometimes they will change jobs or
double up to break a critical bottleneck; often the same person plays
multiple roles. This overlapping of responsibilities results in a
second blessing: a dissolving of tbe classic organizational barriers
tbat are major impediments to tbe innovation process. Tbird, key
decisions can be made immediately by tbe people wbo first
recognize a problem, not later by top management or by someone
wbo barely understands tbe issue. Fourtb, tbe concentration of
power in tbe leader/entrepreneurs makes it possible to deploy tbe
firm's resources very rapidly. Lastly, tbe small firm bas access to
multiple funding channels, from the family dentist to a formal
public offering. In contrast, tbe manager of an R&D project in a
largefirmbas effectively only one source, tbe "corporate bank."

Small divisions. In order to recreate tbe entrepreneurial climate of

tbe small firm, successful large bigb-tecbnology firms often employ
a variety of organizational devices and personnel policies. First,
tbey divide and subdivide. Hewlett-Packard, for example, is sub-
divided into 50 divisions: tbe company bas a policy of splitting
divisions soon after tbey exceed 1,000 employees. Texas Instru-
ments is subdivided into over 30 divisions and 250 "tactical action
programs." Until recently, 3M's business was split into 40 divisions.
Although these divisions sometimes reach $100 million or more in
sales, by Fortune "500" standards tbey are still relatively small
companies.
Variety of funding channels. Second, sucb bigb-tecb firms employ a
variety of funding cbannels to encourage risk taking. At Texas
Instruments (TI) managers bave tbree distinct options in funding a
new R&D project. If tbeir proposal is rejected by tbe centralized
strategic planning (OST) system because it is not expected to yield
acceptable economic gains, tbey can seek a "Wild Hare Grant." Tbe
Wild Hare program was instituted by Patrick Haggerty, wbile be
was TI's chairman, to insure that good ideas with long-term
potential were not systematically turned down. Alternatively, if the
project is outside the mainstream of the OST, system managers or
engineers can contact one of dozens of individuals who hold "IDEA"

54 THEMcKINSEYQUARTERLY
grant purse strings and who can authorize up to $25,000 for
prototype development. It was an IDEA grant that resulted in TI's
highly successful "Speak & Spell" learning aid.

3M managers also have three choices: they can request funds from
(1) their own division, (2) corporate R&D, or (3) the new ventures
division. This willingness to allow a variety of funding channels has
an important consequence: it encourages the pursuit of alternative
technological approaches, particularly during the early stages of a
technology's development, when no one can be sure of tbe best
course to follow.

IBM, for instance, bas found tbat rebellion can be good business.
Artbur K. Watson, tbe founder's son and a long-time senior
manager, once described tbe way tbe disk memory, a core element of
modern computers, was developed: "[It was] not tbe logical outcome
of a decision made by IBM management; [Because of budget
difficulties] it was developed in one of our laboratories as a bootleg
project. A bandful of men . . . broke tbe rules. Tbey risked tbeir jobs
to work on a project tbey believed in."^

At Northrop the head of aircraft design usually has at any one time
several projects in progress without the awareness of top manage-
ment. A lot can happen before tbe decision reacbes even a couple of
levels below tbe cbairman. "We like it tbat way," explains Nortbrop
Cbairman Tom Jones.

Tolerance offailure. Moreover, tbe successful bigb-tecbnology firms

tend to be very tolerant of tecbnological failure. "At Hewlett-
Packard," Bob Hungate, general manager of tbe Medical Supplies
Division, explains, "it's understood tbat wben you try sometbing
new you will sometimes fail." Similarly, at 3M, those who fail to
turn their pet project into a commercial success almost always get
another chance. Richard Frankel, the president of the Kevex
Corporation, a $20 million instrument manufacturer, puts it this
way: "You need to encourage people to make mistakes. You have to
let them fiy in spite of aerodynamic limitations."

Opportunity to pursue outside projects. Finally, these firms provide

ample time to pursue speculative projects. Typically, as mucb as 20
percent of a productive scientist's or engineer's time is "unprogram-
med," leaving bim or ber free to pursue interests tbat may not lie in
tbe mainstream of tbe firm. IBM Tecbnical Fellows are given up to
five years to work on projects of tbeir own cboosing, from bigb-speed
memories to astronomy.

SUMMER 1985 55
5. Sense of integrity
Wbile committed to individualism and entrepreneursbip, at tbe
same time successful bigb-tecb firms tend to exbibit a commitment
to long-term relationsbips. Tbe firms view tbemselves as part of an
enduring community tbat includes employees, stockbolders, cus-
tomers, suppliers and local communities: tbeir objective is to
maintain stable associations witb all of tbese interest groups.

Although these firms have clearcut business objectives, such as

growth, profits and market share, they consider them subordinate
to higher order ethical values. Honesty, fairness and openness —
that is, integrity — are not to be sacrificed for sbort-term gain. Sucb
companies don't knowingly promise wbat tbey can't deliver to
customers, stockbolders, or employees. Tbey don't misrepresent
company plans and performance. Tbey tend to be tougb but
fortbrigbt competitors. As Herb Dwigbt - president of Spectra-
Pbysics, one ofthe world's leading laser manufacturers - says: "The
managers that succeed here go out of their way to be etbical." And
Alexander d'Arbeloff, co-founder and president of Teradyne, states
bluntly: "Integrity comes first. If you don't bave tbat, notbing else
matters."

Tbese policies may seem Utopian, even puritanical, but in a

bigb-tecb firm tbey also make good business sense. Tecbnological
cbange can be dazzlingly rapid; tberefore, uncertainty is bigb, risks
are difficult to assess, and market opportunities and profits are hard
to predict. It is almost impossible to get a complex product into
production, for example, without solid trust between functions,
between workers and managers, and between managers and
stockbolders (wbo must be willing to see tbe company tbrougb tbe
possible dips in sales growtb and earnings tbat often accompany
major technological shifts). Without integrity the risks multiply
and the probability of failure (in an already difficult enterprise)
rises unacceptably. In sucb a context, Ray Stata, co-founder of tbe
Massacbusetts Higb Tecbnology Council, states categorically: "You
need an environment of mutual trust."
Tbis commitment to etbical values must start at tbe top, otberwise
it is ineffective. Most of tbe CEOs we interviewed consider it to be a
cardinal dimension of tbeir role. As Bernie Gordon, president of
Analogic, explains: "Tbe tbings tbat make leaders are tbeir
pbilosophy, ethics and psychology." Nowhere is this dimension
more important than in dealing with tbe company's employees.
Paul Rizzo, IBM's vice cbairman, puts it tbis way: "At IBM we bave
a fundamental respect for tbe individual . . . people must be free to

56 THEMcKINSEYQUARTERLY
disagree and to be beard. Tben, even if tbey lose, you can still
marshal them behind you."
Self-understanding. This sense of integrity manifests itself in a
second, not unrelated, way - self-understanding. The pride, almost
arrogance, of these firms in their ability to compete in tbeir cbosen
fields is tempered by a surprising acknowledgement of tbeir
limitations. One bas only to read Hewlett-Packard's corporate
objectives or interview one of its top managers to sense tbis
extraordinary blend of strengtb and bumility. Successful bigb-tecb
companies are able to reconcile tbeir "dream" witb what they can
realistically achieve. This is one of the reasons why they are
extremely reluctant to diversify into unknown territories.

6. "Hands-on" top management

Notwithstanding their deep sense of respect and trust for indi-
viduals, CEOs of successful high-technology firms are usually
actively involved in the innovation process to such an extent that
they are sometimes accused of meddling. Tom McAvoy, Coming's
president, sifts through hundreds of project proposals each year
trying to identify tbose that can bave a "significant strategic impact
on tbe company" - tbe potential to restructure tbe company's
business. Not surprisingly, most of tbese projects deal witb new
technologies. For one or two ofthe most salient ones, he adopts tbe
role of "field general": he frequently visits the line operations,
receives direct updates from those working on the project, and
assures himself that the required resources are being provided.

Sucb direct involvement of tbe top executive at Corning sounds

more cbaracteristic of vibrant entrepreneurial firms, sucb as
Tandon, Activision and Seagate, but Corning is far from unique.
Similar patterns can be identified in many larger bigb-tecbnology
firms. Milt Greenberg, president of GGA, a $180 million semicon-
ductor process equipment manufacturer, states: "Sometimes you
just bave to sbort-circuit tbe organization to acbieve major cbange."
Tom Watson, Jr. (IBM's cbairman) and Vince Learson (IBM's
president) were doing just tbat wben tbey met witb programmers
and designers and otber executives in Watson's ski cabin in
Vermont tofinalizesoftware design concepts for tbe System 360 - at
a point in time wben IBM was already a $4 billion firm.
Good bigb-tecb managers not only understand bow organizations -
and in particular engineers — work, tbey understand tbe fun-
damentals of tbeir tecbnology and can interact directly witb tbeir

SUMMER 1985 57
people about it. Tbis does not imply tbat it is necessary for tbe senior
managers of sucb firms to be tecbnologists (altbougb tbey usually
are in tbe early stages of growtb): neitber Watson nor Learson were
tecbnical people. Wbat appears to be more important is tbe ability
to ask lots of questions, even "dumb" questions, and dogged patience
in order to understand in-depth such core issues as: (1) how the
technology works; (2) its limits, as well as its potential (together
with the limits and potential of competitors' technologies); (3) wbat
tbese various tecbnologies require in terms of tecbnical and
economic resources; (4) tbe direction and speed of cbange; and (5)
the available tecbnoiogical options, tbeir cost, tbe probability of
failure and the potential benefits if they prove successful.

This depth of understanding is difficult enough to achieve for one

set of related technologies and markets; it is virtually impossible for
one person to master many different sets. Tbis is anotber reason
wby business focus appears to be so important in bigb-tecb firms. It
matters little if one or more perceptive scientists or technologists
foresees tbe impact of new tecbnologies on tbe firm's markets, if its
top management doesn't internalize tbese risks and make tbe major
cbanges in organization and resource allocation tbat are usually
necessitated by a tecbnological transition.

The paradox of high-tech mans^ement

Tbe six tbemes around wbicb we arranged our findings can be
organized into two, apparently paradoxical groupings: business
focus, organizational cobesion and a sense of integrity fall into one
group; adaptability, entrepreneurial culture and bands-on top
management fall into tbe other group. The first three imply
stability and conservatism; tbe second tbree are synonymous witb
rapid, sometimes precipitous cbange. Tbe fundamental tension is
between order and disorder. Half the success factors pull in one
direction; the other half tug the other way.

This paradox has frustrated many academics who seek to identify

rational processes and stable cause-effect relationsbips in bigb-tecb
firms and managers. Sucb relationsbips are not easily observable
unless a certain constancy exists. But in most bigb-tecb firms, tbe
only constant is continual cbange. As one insigbtful student of tbe
innovation process pbrased it: "Advanced tecbnology requires tbe
collaboration of diverse professions and organizations, often with
ambiguous or bigbly interdependent jurisdictions. In sucb situa-
tions, many of our bigbly touted rational management tecbniques
break down."^ One recent researcber, bowever, proposed a new

58 THEMcKINSEYQUARTERLY
model of tbe firm tbat attempts to rationalize tbe confiict between
stability and cbange by splitting tbe strategic process into two
loops, one tbat extends tbe past, tbe otber tbat periodically
attempts to break witb it. ^°

Establisbed organizations are, by tbeir very nature, innovation

resisting. By defining jobs and responsibilities and arranging tbem
in serial reporting relationsbips, organizations encourage tbe
performance of a restricted set of tasks in a programmed, predict-
able way. Not only do formal organizations resist innovation, tbey
often act in ways tbat stamp it out. Overcoming sucb bebavior —
wbicb is analogous to tbe way tbe buman body mobilizes antibodies
to attack foreign cells - is, tberefore, a core job of bigb-tecb
management.

Meeting the challenge

Higb-tecb firms deal witb tbis cballenge in different ways. Texas
Instruments, long renowned for tbe complex, interdependent ma-
trix structure it used in managing dozens of product-customer
centers (PCCs), recently consolidated groups of PCCs and made
tbem into more automonous units. "Tbe manager of a PCC controls
tbe resources and operations for bis entire family... in tbe simplest
terms, tbe PCC manager is to be an entrepreneur," explained Fred
Bucy, TI's president."
Meanwbile, a different trend is evident at 3M, wbere entrepreneurs
bave been given a free rein for decades. A recent major reorganiza-
tion was designed to arrest snowballing diversity by concentrating
its sprawling structure of autonomous divisions into four market
groups. "We were becoming too fragmented," explains Vincent
Ruane, vice president of 3M's electronics division.^^ Similarly,
Hewlett-Packard recently reorganized into five groups, eacb witb
its own strategic responsibilities. Altbougb tbis simply cbanges
some of its reporting relationsbips, it does give tbe company, for tbe
first time, a means of integrating product and market development
across generally autonomous units.

Tbese reorganizations do not mean tbat organizational integration

is dead at Texas Instruments, or tbat 3M's and Hewlett-Packard's
entrepreneurial cultures are being dismantled. Tbey signify first,
tbat tbesefirmsrecognize tbat botb (organizational integration and
entrepreneurial cultures) are important, and second, tbat periodic
cbange is required for environmental adaptability. Tbese tbree
firms are demonstrating remarkable adaptability by reorganizing

SUMMER 1985 59
from a position of relative strengtb - not, as is far more common, in
response to financial difficulties. As Lewis Lebr, 3M's president
explained: "We can cbange now because we're not in trouble."

Sucb reversals are essentially antibureaucratic, in tbe same spirit

as Mao's admonition to "let a bundred fiowers blossom and a
bundred scbools of tbougbt contend." At IBM, in 1963, Tom Watson,
Jr. temporarily abolished the corporate management committee in
an attempt to push decisions downward and tbus facilitate tbe
cbanges necessary for IBM's great leap forward to tbe System 360.
Disorder, slack and ambiguity are necessary for innovation, since
tbey provide tbe porosity tbat facilitates entrepreneurial bebavior
- just as do geograpbically separated, relatively autonomous
organizational subunits.

But tbe corporate management committee is alive and well at IBM

today. As it sbould be. Tbe process of innovation, once begun, is botb
self-perpetuating and potentially self-destructive: altbough tbe top
managers of bigb-tecb firms must sometimes espouse organization-
al disorder, for tbe most part tbey must preserve order.

Winnowing old products

Not all new product ideas can be pursued. As B. Cbarles Ames,
former president of Reliance Electric, states: "An entbusiastic
inventor is a menace to practical businessmen."^^ Older products,
upon wbicb tbe current success of tbe firm was built, at some point
bave to be abandoned: just as tbe long-term success of tbe firm
requires tbe planting and nurturing of new products, it also
requires tbe conscious, even rutbless, pruning of otber products so
tbat tbe resources tbey consume can be used elsewbere.

Tbis attitude demands bard-nosed managers wbo are continually

managing tbe functional and divisional interfaces of tbeir firms.
Tbey cannot be swayed by nostalgia, or by tbe fear of disg^pointing
tbe many committed people wbo are involved in tbe development
and production of discontinued products. Tbey must also overcome
tbe natural resistance of tbeir subordinates, and even tbeir peers,
wbo often bave a vested interest in tbe products tbat brougbt tbem
early personal success in tbe organization.

Yet firms also need a certain amount of continuity because major

cbange often emerges from tbe accretion of a number of smaller, less
visible improvements. Studies of petroleum refining, rayon and rail
transportation, for example, sbow tbat balf or more of tbe produc-

60 THE McKINSEY QUARTERLY

tivity gains ultimately acbieved witbin tbese tecbnologies were tbe
result of tbe accumulation of minor improvements.^"* Indeed, most
engineers, managers, tecbnologists, and manufacturing and
marketing specialists work on tbe little steps tbat improve or
extend existing product lines and processes.

Ambivalent management
Tbe successful bigb-tecbnology firm, tben, must be managed
ambivalently. A steady commitment to order and organization will
produce one-color Model T Fords. Continuous revolution will bar
incremental productivity gains. Many companies bave found tbat
alternating periods of relaxation and control appear to meet tbis
dual need. Surprisingly, sucb ambiguity does not necessarily lead to
frustration and discontent. In fact, interspersing periods of tension,
action and excitement witb periods of refiection, evaluation and
revitalization is tbe same sort of irregular rbytbm tbat cbaracter-
izes many favorite pastimes - including sailing, wbicb bas been
described as "long periods of total boredom punctuated witb
moments of stark terror."

Knowing wben and wbere to cbange from one stance to tbe otber,
and baving tbe power to make tbe sbift, is tbe core of tbe art of
bigb-tecbnology management. James E. Webb, administrator of
tbe National Aeronautics and Space Administration (NASA) dur-
ing tbe successful Apollo ("man on tbe moon") program, recalled
tbat "we were required to fiy our administrative macbine in a
turbulent environment, and . . . a certain level of organizational
instability was essential if NASA was not to lose control."^^

In summary, tbe central dilemma of tbe bigb-tecbnology firm is

tbat it must succeed in managing two confiicting trends: continuity
and rapid cbange. Tbere are two ways to resolve tbis dilemma. One
is an old idea: managing different parts of tbe firm differently -
some business units for innovation, otbers for efficiency.

A second way — a way wbicb we believe is more powerful and

pervasive - is to manage differently at different times in tbe
evolutionary cycle of tbe firm. Tbe successful bigb-tecbnology firm
alternates periods of consolidation and continuity witb sbarp
reorientations tbat can lead to dramatic cbanges in tbe firm's
strategies, structure, controls and distribution of power, followed by
a period of consolidation. Tbomas Jefferson knew tbis secret wben
be wrote 200 years ago: "A little revolution now and tben is a good
tbing."

SUMMER 1985 61
 REFERENCES

CE. Makin, "Ranking corporate reputations," Fortune, January 10, 1983, pp. 33-44.
Corporate reputation was subdivided into eight attributes; quality of management,
quality of products and services, innovativeness, long-term investment value, financial
soundness, ability to develop and keep talented people, community and environmental
responsibility, and use of corporate assets.
' After only eighteen months as Geneen's successor as president, Lyman Hamilton was
summarily dismissed by Geneen for reversing Geneen's way of doing business. See G,
Colvin, "The re-GeneeningofITT,"Forturae, January 11,1982, pp. 34-39.
' "RCA; still another m&ster"Business Week, August 17,1981, pp. 80-86.
"R&D scoreboard,''Busmess Week, July 6,1981, pp. 60-75.
' R. Stata, Analog Devices, Qaarter/y/feport, 1st Quarter, 1981.
' "Why they are jumping ship at Intel," Busmess Week, February 14, 1983, p. 107; and M.
Chase, "Problem-plagued Intel bets on new products, IBM's financial help," Wall Street
Journal, February 4,1983.
A.P. Sloan, My Years with General Motors, New York, Anchor Books, 1972, p. 401.
' Quoted by D.A. Shon, "Champions for radical new inventions," Harvard Business Review,
March-April 1963, p. 85.
' L.R. Sales and M.K. Chandler, Managing Large Systems: Organizations for the Future,
New York, Harper & Row, 1971.
R.A. Burgelman, "A model ofthe interaction of strategic behavior, corporate context and
the concept of corporate strategy," Acactemj' ofManagement Review, 1983, p.61-70.
S. Zipper, "TI unscrambling matrix management to cope with gridlock in major profit
centers,".Efec(roreicAfeu)s, April 26,1982,p. 1.
M. Barnfather, "Can 3M find happiness in the 1980s?", Forbes, March 11, 1982, pp.
113-116.
Quoted in "Exxon's $600-million mistake,"/'ortane, October 19,1981.
See, for example, W.J. Abernathy and J.M. Utterback, "Patterns of industrial innovation,"
Technology Review, June-July 1978, pp. 40-47.
Quoted in a presentation by Elmer B. Staats before the annual meeting ofthe Academy of
Management in August 1978.

Modesto Maidique is associate professor of engineering manage-

ment at Stanford University; Robert Hayes is professor of business
administration at tbe Graduate Scbool of Business Administration
at Harvard University. Tbis article is reprinted by special permis-
sion from tbe Winter 1984 issue of tbe Sloan Management Review.
Copyrigbt © 1984 by tbe Sloan Management Review Association.
All rigbts reserved.

62 THEMcKINSEYQUARTERLY