"This book addresses a tough problem that most successful companies will face eventually. It's lucid, analytical-and scary."

-Andrew S. Grove, Chairman and CEO, Intel Corporation

ine Re'volutlonary¡ '_wu; Nati6

mn-alt3estseTler That I

¡ ,i

ChangeCf1l1e-way W':J ""--Do ffusiness-

Contents

In Gratitude vii

Introduction xi

pjjrt One: Whq Great [ompanies [¡In Fail 1 How Can Great Finns Fail? Insights from the

Hard Disk Drive Industry 3

2 Value Networks and the Impetus to Innovate 33

3 Disruptive Technological Change in the

Mechanical Excavator Industry 69

4 What Goes Up, Can't Go Down 89

Part lUJo: Managing Oisruptive lechnnlogical [hilnge 111 5 Give Responsibility for Disruptive Technologies

to Organizations Whose Customers N eed Them 117

6 Match the Size of rhe Organization to the

Size of the Market 139

7 Discovering New and Emerging Markets 165

8 How to Appraise Your Organization's

Capabilities and Disabilities 185

9 Performance Provided, Market Demand, and

the Product Life Cyc1e 211

10 Managing Dísruptive Technologícal Change:

A Case Study 235

11 The Dilenunas of Innovation: A Surnmary 257

The Innovator's Dilemma Book Croup Cuide 263

Index 271

About the Author 287

vi • Con ten 1.'

Hum Con Greot Firms Foil? Insights from the Hard Disk Orive Industr~

ONE

W HEN 1 llEGAN my search for an ans\ver to the puzzle oCwhy

the best fiuns can faíl, a friend offered sorne sage advice.

"Those 'who study genetics avoid studying hurnans," he noted. "Be­cause new generations come along only every thirty years or so. it

takes a long time to understand the cause and efrect of any changes.

Instead, they study fruit fues, because they aré conceived, born,

matuIe, and die all witrun a single day. lf you want to undcrstand

why something happens in business, study the disk drive industry.

Those companies are the closest things to fruit flies that the business world 'will ever see.'·

lndeed, nowhere in the history of business has there been an

indusrry like disk drives, where change~ in technology, market struc­ture, global scope, and verdea! integration have been so pervJsÍ've,

rapid. and unrelenting. While this pace and complexity might be a

nightmare for managers, rny friend was right about its being fertile

ground for research. Few industries offer researchers the ¡ame oppor­

tunities for developing rheories about how different types of change cause certain types of firms to succeed or faíl or for testing those

theories as the industry repeats its cycles of change. This chapter sunmurizes the rustory of the disk dlive industry in

al! Íts cornplexity. Sorne readers wiil be interested in it for the sake oÍ history itself ¡ But the value of understanding this history is that out of its cornplexíty emerge a few stunningly simple and consistent faetors that have repeatedly detennined rhe success and failure of the indusuy's best firrm. Simply put, when che best finns succeeded, they did so because chey listened responsivdy to their customers and

invested aggressively in the technology, products, and manufacturing capabilities that satisfied their customers' next-generation needs. But, paradoxically, when the best firms subsequently failed, it was foc

rhe same reasons-they listened responsively te their customers and ínvesred aggressively in the technology, produets, and manufaeturing capabilities thar satisfied their customers' ne:)ct-generation needs. This is one oÍ the innovator's dilemmas: Blindly follovving the maxim that

good managers should keep close to their customers can sometimes be

a fatal mistake. The rustory of the di5k drive industl'Y provides a framework

for understanding when "keeping close to your customers " i5

good advice--and when ir is noto The robustness of this framework could only be explored by researching che industry's history in carefJJ detail. Sorne of that derail is recounted here, and elsewhere in this book, in the hope that readers who are immersed in the

detail of their own industries will be better able to recognize how similar patterns have affected their own fortunes and those of their competitors.

HOW DISK ORIVES WORK

Disk drives write and read infonnation thar computers use. They

eomprise read-write heads mounted at the end of an ann that swing; over the surtace of a rorating disk in much the s'Jme way that a phonograph neeclie and arm reach over a record; aluminum or glas; disks coated with magnetic material; aI least rwo clectric motors, a

spin motor that drives the rotanon ofthe clisks and an actuator motor that moves the head to the desired positíon over t.he disk; and a

4 • Why Creat Compaflies Can Pail

variety of electroruc circuits that control the drive's operation and its interface with the computer. See Figure L 1 for an illustration of a typical disk drive.

The read-write head is a tiny electromagnet whose polarit)' changes whenever the direction of the electrical current runrung through ie changes. Because opposite magnetic poles attrace, when the polarity of the head becomes positive, the polarity of the area

on the disk beneath the head switches to negative, and vice versa. By rapidly changing the direcríon of current flo,-"ing through the head's electromagnet as the disk spins beneath the head, a sequence of posirively and negatively oriented magnetíc domains are created in concentric tracks on the disk' s surface. Disk drives can use the positíve and negatíve domains on ehe disk as a binary numeric sys­tem-l and o--to "write" informaríon onto disks. Drives read infor­matíon from disks in essentially the opposite process: Changes in the

magnetic flux fields on the disk surface induce changes in the micro current f10wing through the head.

Figure 1.1 Pritnary Components of a Typical Disk Driye

Aeluator motor

Controller. other eleetroníe eireuitry underneath

Aluminum or glass

disk coated with magnetie material

Oplieal encoder to ensure clase head· lrack alignment

Spín motor (al base 01 spindle)

/ Hermelíe housing

How Can Creat Fírms Fail? • 5

EMERGENCE OF THE EARLlEST DISK ORIVES

A tcam of rcsearchers at IBM's San Jose research laboratories devel­oped the first disk drive between 1952 and 1956. Named RAMAC (for Random Acccss Mechod for Accouncing and Control), this drive was the size of a large refrigerator, incorporated [¡fty twenty-four­ínch disks, and couId srore 5 megabytes (MB) of ínfoffi1atíon (see Figure 1.2). Most of the fundamental archítecturaI concepts and component technologies that defined roday's donúnant disk drive

design were also developed at LBM. Tbese indude ies removable packs of rigid disks (introduced in 1961): the floppy disk drive (1971);

and the Winchester architecture (1973). All had a powerful, defining influence on the way engíneers in the rest of the industry defined

what disk drives .vere and what they couId do. A, IBM produced drives to meet its own needs, an independel1t

Figure 1.2 The First Disk Drive, Developed by mM

Source: Courtesy of Internacional Business Machines Corporation.

6 • 1+1¡y Crear COl1lpanics Call Faíl

disk drive industry emerged serving rwo distinct markets. A few firms developed the plug-compatible market (PCM) in the 19605, 5elling souped-up copies of IBM drives direcrly to IBM customers at dis­

count prices. A1though most ofIBM's competitors in computers (for example, Control Data, Burroughs, and Univac) were integrated venicaUy into the manufacture of their o\Vn disk drives, the emer­

gence in the 19705 of smaller, nonintegrated computer makers such

as NixdOlf Wang, and Prime spawned an original equipmem market (OEM) for disk drives as well. By 1976 abour $1 billion worth of disk drives were produced, of which captive production accounted for 50 percent and PCM and OEM for about 25 percent each.

The next dozen years unfolded a remarkable stor)' ofrapid growth. market turbulence, and technoJogy':'driven performance improve­ments. The value of drives produced rose to about $18 billion by 1995. By rhe rnid-1980s the PCM market had become insignitlcanc,

while OEM output gre,," to represent about three-founhs of \Vorld production. Of the seventeen firms populating the industry in 1976-a11 of which were relatively large, diversified corporations such as Diablo, Ampex, Memorex, EMM, and Control Data-all exeept

IBM's disk drive operation had faiJed or had been acquired by 1995. During this period an additional 129 firms entered the inclustry. and 109 of those also failed. l\side from IBM. Fujitsu, Hitachi, and NEC. all of the producers remaining by 1996 had entered the industry as

start-ups afrer 1976. Some have attributecl the high monality rate among the integrated

firms that creared the industry te its nearly unfathomable paee of technolo¡:,'icaJ change. Indeed, the pace of change has been breathtak­

ing. The number of megabits (Mb) ofinforrnation rhat the industry's

engineers have been abJe to pack into a square il1ch of disk surface has il1creased by 35 percent per )'ear, on average, from 50 Kb in 1967 to 1.7 Mb in 1973, 12 Mb in 1981, and 1100 Mb by 1995.

The physical size of rhe drives was !educed at a similar pace: The smallest avaiJable 20 MB drive shrank from 800 cubie inehes (in.3)

in 1978 to 1.4 in .. ' by 1993-a 35 percent annual rate ofreduction. Fignre 1.3 shows that the slope ofthe industry"s experience eurve

(which correlates the cumulative number of terabytes (one thousand

HoU' Can Great Fimu Fail? • 7

Figure 1.3

~ .!2 1,000.00 ;g N CID

~ e s '" e o

S2.

i ca O>

'" :¡¡

100.00

10.00

1.00

Disk Drive Price Experience Curve

197~~

'~ 1985

• • 1990 •

53%SIOpe/ •

.'~ .. 1994

0.10 -j-----,-----.,.-------,---------, 10 100 1,000 10,000 100,000

Cumulative Terabytes Produced

SOUfce: Data are from various issues of Disk/Trend Report.

gigabytes) of disk storage capacity shipped in the industry's history

to the constant-dollar price per megabyte of memory) was 53 per­

cent-meaning thar with each doubling of cumulative terabytes

shipped, cost per mega byte fell to 53 percent of its former leveL Trus

is a much steeper rate of price decline than the 70 percent slope

observed in rhe markets for most orher microelectronics producrs.

The price per megabyte has declined ar about .5 percent per quarter

for more than twenty years.

THE IMPACT OF TECHNOLOGICAL CHANGE

My investigarion imo why leading firrns found it so difficult to stay atop

the disk drive industry 1ed me to develop rhe "technology mud,lide

hypothesis": Coping \.\'ith the relendess onslaught of technology

change was akin to trying to climb a mudslide raging clown a hill. You

have to scramble wirh everything you've got ro stay on top of ir, and if you ever once stop to carch your breath, you get buriecl.

8 • ¡Vil}' Crea! Compallies Ca/l Fllil

To test this hypothesis, I assembIed and anaIyzed a database con­

sisting of the technical and perfomlance specificarions of every model

of disk drive íntroduced by every company in the world disk drive

industry for each ofthe years between 1975 and 1994.2 This database

enabled me to identify the firms thar led in introducing each new technology; to trace how new technologies were diffused through

rhe industry over time; to see which firms led and which lagged;

and to measure the impact each technological innovation had on

capaciry, speed, and other parameters of disk drive performance. By carefully reconstructing the history of each technological change in the industry, the change~ that catapuIted entrants to success or that

precipitated the failure of established Ieaders couId be identified.

This study led me te a very different view of technology change

than the work of prior scholars on this question had led me to expect. Essentially, ir revealed thar neither the pace nor the difficulry of technologic31 change lay at the root of rhe leading fil1llS' failures.

The technology mudslide hypothesis Y\las wrong. The manufacturers of most products have established a trajectory

of perfonnance improvement over time.3 Intel, for example, pushed

the speed of its microprocessors ahead by about 20 percent per year,

from its 8 megahertz (MHz) 8088 processor in 1979 to its 133 MHz

Pentium chip in 1994. Eh Lilly and Company improved the purity

of its insulin from 50,000 impure parts per million (ppm) in 1925 to 10 ppm in 1980, a 14 percent annual rate of improvement. When

a measurable trajectory of improvement has been established, de­

tel1llining whether a new technology is likeJy to improve a product's

perf0l111anCe relative to earlier products is an unambiguous question. But in other cases, the impact of technological change is quite

different. For instance, is a notebook computer better than a main­

frame? This is an ambiguous question because the notebook computer

established a completely new pelformance trajectory, with a defini­tion of perfol1llance that differ, substantial1y from the way mainframe

perfonnance is measured. Notebooks, as a consequence, are general]y sold for very different uses.

This study of technological change over the history of the disk

drive industry revealed two rypes of technology change, each with

HoU' Catt Creat Fírms Pail? • 9

ver}" differt'nt effects en the industry's leaders. Technologies cf the first sort sustailzed me industry's rate of improvement in produce perfonnance (total capaciry and recording density were the two most

COllunon measures) and ranged in difilculty from incremental to radicaL The industry's dominant firms :uways led in deve10ping and

adopting these tedmologies. By contrasto innovations of the second sort disnlpted or redefined performance trajeetories-and consistenrly

resulted in che tailure oftne industry's leadíng nnm.' The remainder of this ehapter illustrates the distínction between

sustaining and disruptive technologies by describing prominent eX<lln­pIes of eaeh and summ<lrízing the role rhese played in che industry's

development. This discussion focuses on differences in how estab­líshed fimls carne ro lead or lag in developing and adopting ne\v technologies, conl.pared ",ith entrant fimls. To arrive at these exal11-

pies, each new technology in the lndustry was examined. In anaIyzmg

wruch firms led and lagged at eaeh ofthese points of change, 1 defined estahlishcd j/1711-' tú be those rhat had been established in the industry prior to che advent of the technology in quescion, practicing tne prior techllolOg)r. 1 defined entrallt jirr/lS as ,hose chat were new co

(he industry at thar point of technologv change. Hence, a given fiml \VouJd be considered an entrant at one specific point in the industry's

history, for ex.lmple, al che emergence of che 8-inch drive. Yet the same firm ""ould be considered an establi,hed firm 'l.vhen technologies

thal emerged subsequent to the fum's entry \vere srudied.

SUSTAINING TECHNOLOGICAL CHANGES

In th!:: history of the disk drive industry, most technology changes have sustained or reinforced established crajectorie:\ of product pelfor­manee improvement. Figure ¡.4. which compares che average re­

cording density of drives that employed suecessive generations of head and disk technologies, maps a11 example of this. The first curve pIots the density of drives that used conventional particuJare oxide disk technology and ferrite bead technology; the secand charts rhe

average density of drives thar used new-technology rhin-film ht!ads

JO· H'hy GT!!.Jt C<Jl1lpa.lies Can E¡¡l

Figure 1.4 Impact ofNew Read-Write Head Technologies in Sustainillg the Trajectory of Improvement in Recording Density

1,000 ]

:2 c.> .5

100

1975 1980

Magneto-resistive heads" JÍ~

r,)/~

Thin·film heads ',,-

1985

Year

1990 1995

Source: Data are from various issues of Di5k/Trend Report.

and disk:;; the third marks the improvements in density achievable

with the latest head technology, magneto-resistíve heads.5

The way such ne,v technologies as these emerge to surpass the perfonllance oE the old resembles a series oE mtersecting technology

S-curves.6 Movement along a given S-curve is generally the result

of incremental improvements withm an cxisting technologic:al ap­

proach, "vbereas jumping onto the next technoJogy curve implies adopting a radica1ly new technology. In the cases measured in Figure

1.4, incremental advances, such as grinding che felTite heads to fincr, more precise dimensions and ming smaller and more findy dispersed

oxide particles on che disk's surface, led to the improvements in

densiry from 1 to 20 megabits per square illCh (Mbpsi) between 1976

and 1989. As S-curve theory would pn:dict, che improvement in recording density obtainable wíth ferrite/oxide technolo!,ry began to

level off toward the end ofthe period, suggesting a maturing technol­ogy. The thin-film head and disk teclmologies' effecr on the industry

HolI' Cal1 Crea! Fi>-l1Is Fa;/.' • 11

sustained performance improvement at its historical rateo Thin-film

heads were barely established in the early 19905, when even more advanced magneto-resistive head technology emerged. The impact of magneto-resistive technology sustained, or even accelerated, the rat.e of performance improvement.

Figure 1.5 describes a sustaining technological change of a very different character: an innovation in product architecture, in which the 14-inch Winchester drive is substituted for removable disk packs, which had been the dominant design between 1962 and 1978. Just as in the thin-film for ferrite/oxide substitution, the impact ofWin­ches ter technology sustaíned the historically established rate of perfor­mance improvemenr. Similar graphs could be constructed for most other technoJogical innovations in the industry, such as embedded

servo systems, RLL arrd PRML recording codes, higber RPM mo­tors, and embedded interfaces. Some of these were straightforward

Figure 1.5 Sustaining lmpact of the Winchester Architecture on the Recording Density of 14-inch Disk Drives

:c 10 <> E: Winchester drives "-!!:! 5 C\l :¡ cr 3 (/) .. GJ

2 Q.

~ íii 'O 11) t:

~ 0.5 '" Removable dlSk pack drives

! ~ 0.3 .¡¡¡ c: GJ 0.2 Q .. !!! c:(

0.1 1965 1970 1975 1980

Year

Source: Data are from various issues of Disk I Trend Report.

12 • vl'7,y Creat Companies Can Faíl

technology improvements; others were radical departures. But all had a similar impact on the industry: They helped manufacturers to

sustain the rate of historical performance irnprovement that their custorners had come to expect. 7

In literally every case of sustaining technology change in che disk

drive industry, established firms led in development and cornmercial­ization. The emergence of new disk and head technoJogies illustrates

this. In the 19705, sorne manufacturers sensed that they were reaching

the límit on the number ofbits of information they could pack onto oxide dísks. In response, disk drive manufacturers began studying ways of applying super-thin films of magnecic metal on aluminum

to sustain the hísrorical rate of improvements in recordíng density. The use of tmn-film coatings was then highly developed in che

integrated circuit industry, but its application to magnetic disks sti.ll presented substantial challenges. Experts estimate that the pioneers of thin-film disk technology-IDM, Control Data, Digital Equipment, Storage Technology, and Ampex-each took more than eight years and spent more than $50 millíon in that ettort. Between 1984 and 1986, about two-thirds of the producers active in 1984 introduced drives \vith thin-film disb. The over.vhelmíng majority of these

were established industry incumbents. Only a few entrant firms at­tempted to use thin-film disks in theír inicial products, and most of those folded shortly after entry.

The same pattem \Vas apparent ÍIl rhe emergence of thin-film

heads. Manufacturers of ferrite heads saw as early as 1965 the ap­proaching linut to improvements in this technology; by 1981 many believed that the limi.ts of precisíon \VouId soon be reached. Re­searchers tumed to thin-film technology, produced by sputtering

thin fihns of metal on the recording head and then using photohthog­

raphy to etch much finer electromagnets than couId be attained with ferrite technology. Again, this proved extraordinari1y difficult. Burtoughs in 1976, IBM in 1979, and other establíshed firms first successfully incorporated thin-film heads in disk drives. rn the period bet\Veen 1982 and 1986, during which sorne síxty firms entered the rigíd disk drive industry, only four (all cornmercial failures) attempted

How Can Crear Firm, Fa!l? • 13

to do so using thin-film heads in their initial producrs as a so urce of perfomlance advantage. All other entrant firms-even aggressively

perlonnance-oriented firms such as Maxtor and Conner Períph­erals-found it preferable to leam theír \Val' usíng conventional territe heads first, before tackling thin-film technology.

As \vas the case with thin-film disks, the íntroduction of tlún­

film heads entaikd the sort of sustained ínvestrnent thar onIy estab­lished tirms could handle. IBM and its rivals each spent more than $100 míllion developíng thin-film heads. The pattem ,"vas repeated ín the next-generatíon magneto-resístive head technology: The in­dustry's largest firms-rBM, S eagate , and Quantum-led the race.

The establíshed finns were the leading ínnovators not just in developing risky, complex, and expensíve component technologies

such as thin-film heads and disks, but in liccrally everJ' otile,. one cf the stlstaifling iflllovafÍolls in ¡he hldustry's histar)'. Even in relatively simple

innovations, such as RLL recording codes (which took the índustry from double- to triple-densíty disks), establíshed firms were the su c­cessful píoneers, and entrant D=S ""ere the technoiogy follo\vers. Thi5 was also tme for those archítectural innovations-for example,

14-inch and 2.5-inch Winchester drives-whose impact \vas to sus­tain established improvement trajectories. Established firms beat out the entrants.

Figme 1.6 sUlnmarízes this partem of techllo1ogy leadership

among estz,blished and entrant firnls offering products based on new

sustaining technologies during the years when those technologies were emergmg. The pattern i5 stunningly consistent. Whether the

technoJogy \Vas radical or incremental, expensi\'e or cheap. software or hardware, component or architecture, competence-enhancing or

competence-destroying, the pattem was rhe same. When faced with sustaining technology change that gave exísting customers something more and better in what they wanted, the leading practitioners of the

prior rcchnology Jed the industry in the developmem and adoption of the new. Clear1y, the leaders in this industry· did not faí] because they became passive, arrogant, or risk-averse or because they couldn't keep up with the srunning rate of technological change. 1\1)' technol­

ogy mudslide hypothesís \vasn't correcto

14 • 1'¡"71y Grcal. Compatlícs Ca/1 Fail

Figure 1.6 Leadership ofEstablisbed Fínns in Sustaining Technologies

'* ~ 30 en Qj .:! 30

'1:l '" '1:l en c '" c-:lE :1: 25 :lE O 25 DIE DIE E= 1:::

~ ~ 20 'g ~ 20 =.5 e; ~ 15 e; ¡: 15

E.r. ES ii:jlO ii: -¡ 10 'O-rI --rI

o '" ... 8. 5 ... 1:1. S '" 1:1. "'1:1. .c ._ .c-E '" E '" '" O" O ",O" O ZW ZW

76 80 84 88 84 85

Vear Vear

'* 2°l !!J. en 35 g''g I '" '" 'g 'g 30 i ~ 16-1 :lEC) SS ~ :5 g' 25

~ ~ 12~ .¡¡¡¡ 'S '" 5 20 =:2 -¡ E <>

IL u ¡¡ & 15 '0< s--i J _....1

Q; ; ~ ~ 10 ~gj 4 ~~ ",.r. E- S Z u '" '" 1:

O z;:! O § 74 76 78 84 85 86

Vear Vear

CJ Entrant Firms ~ Established Firms

So urce; Data are froro various issues of Disk/Trend Report.

FAILURE IN THE FACE OF DISRUPTIVE

TECHNOLOGICAL CHANGES

86

87 88

Most technological change in the disk drive industry has consisted of sustaining innovations of the sort described aboye. In contrast, there have be en only a few of the other son of technological change, called disruptive technologies. These were the changes that toppled the industry's leaders.

How Ca11 Crea! FintE Fail? • 15

The most important disruptive technologies were rhe architec­tural innovations that shrunk the size of the drives-from 14-inch diameter disks to diameters of8, 5.25, and 3.5-inches and then from 2.5 ro 1.8 inches. Table 1.1 ¡llustrates the ways these innovations were disruptive. Based on 1981 data, it compares the attributes of a

typical 5.25-inch. drive, a new archirecture that had been in the market for less than ayear, with those of a typieal 8-inch drive,

whieh at that time was the standard drive used by minieomputer manufaeturers. Along the dimensions of performance important to established minicomputer manufacturers-capacity, cost per mega­byte, and access time--the R-inch product was vastly superior. The 5.25-ineh architecture did not address the perceived needs of mini­computer manufacturers at that rime. On the orher hand, the 5.25-inch drive had rearmes that appealed to che desktop personal computer

market segment just emerging in the period between 1980 and 1982. It was smalJ and lightweight, and, priced at around $2,000, it could be ineorporated into desktop machines economically.

Generall)' disruptive ínnovations \Vere teehnologically straightfor­ward, consisting of off-the-shelf components put together in a prod­uct architeeture that was often simpler than prior approaches." They offered less ofwhat customers in established markets wanted and so eouId rarely be initially employed there. They offered a different

package of attributes valued only in emerging markets remo te from, and unimportant to, the mairntream.

Table 1.1 A Disruptíve Technology Change: The 5.25-inch Winchester Disk Drive (1981)

Attribute

Capaciry (megabytes) Pbysical ,"olume (eubíc ínebes) \1/ eight (pounds) Access time (mllliseconds) Cost per mega byte Unit cost

S-Inch Drives (Minicomputer

Market)

60 56ñ

21 30

$511 $300()

5.25-Inch Drives (Desktop Computer

Market)

10 lS0

6 160

$200 52000

Source: Data are from various issues oí Disk I Trend Report.

16 • rVhy Great Compal1íes Ca/I raíl

The trajectory map in Figure 1.7 shows how thi~ series of simple but disruptive technologies proved to be the undoing of sorne very aggressive, astutely managed disk drive companies. Until the mÍd-1970s, 14-inch drives witil rcmovable packs of disks accounted for

nearly all disk drive sales. The 14-inch Winchester architecture then emerged to sustain the trajectory of recording density improvement.

Figure 1.7 Intersecting Trajectories of Capacity Dcmanded versus Capacity Supplied in Rigid Disk Drives

10,000

1,000

iD ~ ?: 'u ., C. ni 100 U ... ., Q "E ., J:

10

1975 1980 1985 1990 Year

So .. rce: Clayton M. Christensen, «The Rigid Disk Drive Industry: A History ofCommerciaJ and Technological TurbuJence," Business History Reviell' 67, no. 4 (Winter 1993): 559. Reprinted by pennission.

HOlV Can Crear Firms Faíl? • 17

Nearly a11 of these drives (removabIe disks and Winchesters) were sold to mainframe computer manufacturers, and the same companies ¡hat led the market in disk pack drives led the industry's transition ro the Winchester technology.

The trajectory map shows that the hard disk capacity provided

in the median priced, typically configured mainframe computer sys­tem in 1974 waS about 130 MB per computer. This increased at a 15 percent annual rate over the next Efteen years-a trajectory

representing the disk capacity demanded by the typical users of new mainframe computeTS. At the same time, the capacity of the average 14-inch drive introduced for sale each year increased at a faster, 22 percent rate, reaching beyond the mainframe marker to the Iarge scientifJc and supercomputer markets.?

Between 1978 and 1980, several entrant firms-Shugart Associ­ates, Mi cropolis, Priarn, and Quantum-developed smaller 8-inch drives with 10, 20, 30, and 40 MB capacity. These drives were of no interest to mainframe computer manufacrurers, which at that time

were demanding dríves with 300 to 400 ME capacity. These 8-inch entrants therefore sold their disruptive drives ínto a new applicatíon­minicomputers. l >1 The customers-Wang, DEC, Data General, Prime, and Hewlett-Packard-did not manufacture mainframes, and

their customers often used software substantíally different from that uSt'd in mainframes. These firms hitherto had been unable to offer disk drives in their small. desk-side ITÚnicomputers because 14-inch models \Vere too big and expensive. Although initially the cost per

megabyte of capacity of8-inch drives was higher than thar of 14-inch drives, these new customers were vvilling to pay a preITÚum for other attributes that were important to them-especiaJly smaller size. Smallness had little value to maintrarne users.

Once the use of8-inch drives became establíshed in ITÚnicomput­ers, the hard disk capaciry shipped with the median-priced minicom­

puter grew about 25 percent per year: a trajectory deterrnined by rhe ways in which minicomputer owners Ieamed to use theír ma­

chines. At the same time, however, the 8-inch drive makers found that, by aggressively adopting sustaining innovations, they couId in­

crease the capacíty of their products at arate of more than 40 percent

18 • Wlly Crear COll1pallies Can Fa;!

per year-nearly double the rate of increase demanded by their original "home" minicomputer market. In consequence, by the mid-19805, 8-ioch drive makers were able to provide the capacities re­quired for lower-end mainfranle computers. Unit voJumes had grown

significantly so that the cost per megabyte of 8-inch drives had declined below that of 14-inch drives, and other advantages became

apparent: For example, the same percentage mechanicaJ vibration in an 8-inch drive, as opposed to a 14-inch dtive, caused much less

variance in the absolute position of the head over rhe disk. Within a three-to-four-year period, therefore, 8-inch drives began tO invade the market above them, subsútuting far 14-inch drives in the lower­end mainframe computer market.

As the 8-inch products penetrated the mainframe market, the establi,hed manufacturers of 14-ineh drives began to fail. Two-thirds of them never introdueed an 8-ineh model. The one-third that introduced 8-ineh models did so about two years behind the 8-inch

entrant manufacturers. Ultimately, every 14-ineh drive maker was driven trom the industry. 11

The 14-ineh drive makers \Vere not toppled by the 8-iIlCh enn'ants beeause of technology. The 8-inch products generalIy incorporated

standard off-the-shelf eomponents, and when those 14-inch driv!;' makers that did introduce 8-ineh models got around to doing so, their products \Vere very performance-competitive in capaeity, areal

densir)'. access time, and plice per megabyte. The 8-inch models

irmoduced by the estabhshed firms in 1981 were nearly identiea1 in perfom1ance te the average of those introduced that year by the entrant firms. In addition, the rates ofimprovemenr in key attributes (measured bet:\veen 1979 and 1983) were stunningly similar between established and entrant firmS.12

Held Captive by 17/eir Customers

Why \Vere rhe leading drive makers unable to launch 8-inch drives untiJ it was too late? Clearly, they vv'ere technologieally capable of producing these drives. Their fuilure resulted from deJay in making the strategic commitment to enter the emerging market in whieh

How Can Creat Finlls Paíl? • 19

the 8-lnch drives ínitially could he soldo Imerviews wirh marketing

and engíneering execuuves close ro rhese companies suggest that the

established 14-inch drive manufe.eruten; ",vere held captive by custom­

e!S. Mam[rame compurer J:1alluÜcmrers did not neec an 8-lnch drive.

In fact. they expliCltly did notwant it: they warlted drives ,<vítn increased

caparit)' at a lower cost per megabyte. The 14-inch drive manufacturers

were listeningand respondingto lheirestablished custo!11ers. And their

customers-'m a way thar was not apparent ce eirher rhe disk drive

manuf.1.crurers Ol' theír cOl11pttter-making custOmers--\vere puliíng

tht>m akmg a traiectory of 22 perrent capacity faowth in a 14~mch

platfonn that \\"ould \11timately prove fatal.!.'

Figure 1.7 mar s ::he dispa,.He tr;Úl.'ftorie, of perfo:r:nance im­

provem~,1[ demandee in rhe computer product sei-,'ment, that

emerged latero compared to the capacit:-' rhar cnange5 in componenr

technolog\' and rcfint'llll.'m, in system desÍgtl made a\'ailable wlmin

ea eh succe,sive archítecrure. The sohd li:1es em:mating from poínts

A, D, e, D. and E measure the di:¡k drive rapacit); pro,'ided with

tl1e median-p,iced computer in ('aeh category, ,,:hile me dotted Enes

fron) the same points measure the ave:-age capac:ity oC all disk drives

introduced fi)r salt' in each architecture. for eaeh year. These transi­

tiom are brietly described be/o\...-.

TI/e Aduc/1t ¡2( the 5.25-i¡¡ch Drf¡·c

In 1980, Scap.:ate Technolog\' introduced 5.25 ínch di~k drives. Thelr

c:¡paciti e, of 5 and 10 rvm were (ji no interest to minicorr..puter

manufacturers, wha \Vere demanding drÍves of 40 ana 60 MD [mm rheir suppliers. Seagate a;ld other finns that emered witb 5.25-inch

nrivesin thf period 1980 to 1983 (for e:xample, Miníscribe, Computer

l\1emories, and Internacional Memories) had lO pioneer new apphca­

tíons for tbei:' products an¿ turned primarily to desktop personal

compllter l;lakers. By 19<10. the me ofhard drives in desktop comput­

ers was an obvious applicatíon for maf;'11etic recordlng. lt was !lot at

:111 elear in 1980, !-JO\H'H'·r. when the marh.et was jusr emerging, rhat

manv people cauld ever afford or use a hanl drive on the desktop. The eady 5.25-incl: drive maKers found thís application (one rn.:ght

even sal' that they enabled ir) by trial and error, selling drives to

whomever \VouId buy them. Once the use ofhard drives was established in desktop pes, the

disk capacity srupped with the median-priced machine (that is, the

capacity demanded by the general PC user) increased about 25 per­

cent per year. Again, the teehnology improved at nearly twice the rate demanded in the new market: The capacity of new 5.25-inch drives mcreased about 50 percent per year bet\:veen 1980 and 1990.

As in the 8-inch for U-ineh substitution, the first fimls to produce 5.25-inch drives were entrants; on average, established firms lagged behind entrants by two years. By 1985, oniy half of the firms produc­ing 8-inch dri ... ·es had introduced 5.25-inch models. The other half

never did. Growth in the use of 5.25-inch drives oecurred in two waves.

The first follo. ved creation of a new application for rigid disk drives: desktop compuring, in wbich product attributes such as physical size,

relarively unimportant in established applications, were high1y valued. The second wave followed substirution of 5.25-inch disks for larger drives in established minicomputer and mainframe computer markets, as the rapidly increasing capacity of 5.25-inch drives intersected the more slmvly grov;i.ng trajectories of capacity demanded in these markets. Of the four leading 8-inch drive makers-Shugart Associ­ates, Micropolis, Priam, and Quantum-only Micropolis survived

to become a significant manufacturer of 5.25-inch drives, and that was accomplished only V\>ith Herculean managerial effort, as described

in chapter 5.

The Pattem 15 Repeated: 17lc Emergence cfthe 3.5-ínch Drive

The 3.5-inch drive '.vas first developed in 1984 by Rodime, a Scottish entrant. Sales of this architecture were not significant, however, until Conner Peripherals, a spinoff of 5.25-inch drive makers Seagate and

Miniscribe, started shípping product in 1987. Conner had developed a small.lightweight drive architecture that was much more rugged than its 5.25-inch ancestors. lt handled electronically functions that had previously been managed with mechanical parts, and it used

How Can Creat Firms Fai!? • 21

llllcrocode ro replace nmctions that had previously been addressed

c1ecrroIllcally. Nearly aH ofConner's firsr vear re\'enue, of$l i3 mi1-llOn H carne [ron: Compaq Computer. \\-;úcn had aiJea COI1ller'S start­

up \yith a 530 r:,illion ínyestmem. The Conncr drives were used pri­

man:y in .lnt."wapplicatlon-ponabk and laprep machines. in addition

to "SJ11~1U foorpnm" dé"skrop modeh-\Yhere customers were willing

tü accept lower capacitics ano fúgher COst5 per megabyte to ger Jighter

\ycight, greater mggedness, and lower power consumption.

Seagate engineers ",ere llot obli\ious to the ccmillg oÍ the 3.5-­

inrh architecmre. lndced, in earl.y 19R5. less than one year afier

Rooime introdllCt'd the t1rst 3.5-inch dri\-e and IWO ye-ars b~/o,.e

Conner Peripherah 5tarred shipping irs produce. Seagar,t pe:-sonnel

showed working 3.5-incn protorype dr¡ves to cu,romers for evalua­

cion, The inihative túr the ne\\' drives came íTem Seagate's engi­

neering organizatlon. OppmÍtio¡] t0 the program came primaril)'

trom the markt"tíng organization and Seagate's executive teanl; they

arguea rnat the market wamed higher capacit\, drives at a lower cos!

per megabyte and rhar 3.5-itH.,h ¿rin:s could ne\er be built at a lower

co~t per rnegabyLC rhan 5.2.3-incn driye,.

Seagate \ mark<>ter., resr",d rhe 3.5-inch Prot0tlT<:,s \vith customers

i:1 rhe desktop cOlllputing Durket ir already sen'ed---manufacturers

like IBM, and \'ake-addd re,e11cIS uf fuJl-sized desktop computer

s)'stems. Not su:-pmi;lgiy, they inilicated líttle Ímerest in the ;;l11aller

dnve. They \Vere ¡ookmg for capacitle, of 40 and 60 megabyte~ for

thelT next-generation machines. while che 3.5-il1ch architeaure could pro\'ide onl\' 20 MB-;1no ar highcr costs.: 3

In re~pome to lukn\"aml Te\'ie\Y~ fi"om cU"rotners. Seaga!''''s ~ro­gTam manager JO\\'ered hi, 3.5 .. i11Ch sales estimares, and the tim1 's

executi"t,s cancded the programo Theír reasoning? The r:J.arkets f()[

5.25-in(h producrs wcre Iarge:-, and the sales generated hy spe;¡ding

the engineering erron on new S.25-inch produn5 \\'ould create

greater r\"'ven\1es fOl' che compau:' d:an \yould errons targtted at new 3.5-lnch produas.

Ir: retraspen, it aprears that Seagate eXl,clltives read th(' mar­

kct-at Ieast their own market-\'('n' accurately. \Vith esrabiished

applications and produet archirtetures of their o'\vn, such as ¡he

IBM XT and A.T, rhese customers sa',," no yalue in the improved ruggedness or che reduced size, \,,-eight, and power consumptlOll of

3.5-inch products. Seagate tinally began shipping 3.5-inch drives Ín eady 1988-the

samIO' year in \yhich the performance trajectory of 3.5-inch drives

(shown in Figure 1.7) inte-rsected [he traJectory of capacity demanded

in desktop computers. By that time, rhe industry had shipped, cumu­lati\"ely, neady S750 million in 3.5-inch products. Interestingly, ac­cording to índustry obser"ers. as of 1991 almost none oE Seagate's

3 .5-ínch products had been sold to manufacturers of portable/laptop/ notebook computers. In other \Yords. Seagate's primary customers

\\"ere still desktop computer manufacturers, and many ofits 3.5-Ínch dri,-es \"ere shipped wíth frames for mounring them in computers

designed for 5.25-inch drives. The fear ofcannibalizing sales of existing products is ofren cited as

a reason ,,..hy established finns dela" the introducnon oE ne\\" technol­ogies. As rhe Seagate-Conner experience illustrates, however, if n e'" technologies enable new market applicatíons to emerge, the inrroduc­

tion of new rechnology may not be inherent1y cannibalistic. Bur when

established finns \"ait until a ne\\" rechnology has become commercially mature in its ne\v apphcatíons and launch their own version of the rechnology only in response to an attack on their home markets, the

fear of cannibalizarion can become a self-fulfilling prophecy. Although \\'e haye been looking at Seagate's response to the devel­

opmenr ofthe 3.5-1nch driye architecture. its behaviorwas not atypical; by 1988, only 35 percent ofthe drive manufacturers thar had established

themselves making 5.25-inch products for ,he desktop PC market had Íntroduced 3.5-inch drives. Similarto earlierproductarchitecture tran­sitions, the barrier to de\'elopment of a competitiye 3.5-inch product does nor appear to ha\'e been engineering-based. As in the 14- to 8-inch transitlon, the nev,;-architecture dnyes introduced by the incum­bent, established tlrms during rhe transitiorrs from 8 to 5.25 inches and from 5.25 to 3.5 inches \vere fuliyperformance-competltlve w¡th those of entrant dri,·es. Rather, the 5.25-inch drive manufacturers seem to

have been nllsled by their customers, notably IBM and its direct com­

petitors and resellers, \"ho themselves seemed as obli .... ious as Seagate

Hot/l Can Creal Firms Fail? • 23

to the potentü! benetits and pO$sibllitíes of porcable compucing and the

new disk cirive architecture thar mlghr :-aClut,lre ir.

Prairietd:, CD1IIICr, alld tlu: 2.5-illch Drí!JC

In 1989 ari industry emram in Longrnont, Colorado, Prairietek,

upstaged rhe índustry b\· annouTlcÍng a 2.3 .. inch drive, capturing

nearly all S3!l milhon of this r.ascent :narket. But Conner Peripherah

announceá its own 2.5-illCh product in cady 1990 alld by the end

of that year had daírncd 95 percent of tbe 2.5-illCh drive marker.

Prairietek dedareci bankr1.lptcy in late 1991, by \óICh time eaeh of

the oúH.'r .o.S-inc!: drivemakers·--Quanmm. Seagate. \Vestem Dif.'"Í­

tal. and Maxtor-had introduced 2.5- inch drives of rhei: own.

\Vbt had changed? Had rhe incumbem leading finns finaU)'

learned rhe kssons ofhIStüTV? Not realk l\lchough Figt:re 1.7 shows

tbe 1.5-inch drive had sígnitlcanclv less capaclty than ,he 3.3-inch

drives, lhe portable computíng markers imo whích [he smaller drins

were sold vaIued other attributes: \veigbt, mggedness, 1m\' power

consumption, sITIall physical size. and so on. Along t{¡ese dimensiom. the 2.5-inch cirivt: offered impron:-d performance over that of the

3.S-inch product: Ir \vas a slIstaÍilÍlIgtechnolog,"):. In facr, the computer

makers who bought Conner's 3.5-mch dnve---laptop cOlIlpurer

manufacturen such as Toshíba, Zerútb. ano Sharp-wcre dw Jeading

makers of noteboOK computers, and these finlls neened tbe sm.iller

2.5-inch drive architecrure .. Hence. Cünner and ¡es competitors in

the 3.5-i:KÍl market followed their clI,rorners seamlessly across the

transiúon to 2.5· inch d;i\'~'S.

In 1 Sl92. however, the 1.8-inch drive emerged, , .... íth ;¡ disonctly

disruptíve character . .I\.Jtl:1.ough ir5 stor)' \\iJ.! be recounted in detall

la ter, it sufllces to state nen' that by 1995. lt "-'as eiltrant trr:1S (hat

controlled 98 percent of the S]30 mi:lioll 1.8-inch drive markeL

Moreover. rhe la:g:f'st ínj(]¡)! :narket íor 1.8-mch duves wasn't in

compllting ar aH. lr \Vas in portabk heart monitocng devices!

Figure 1.8 summarizes this pattem 01 entraar hITJ1S' ;eadership in

disr>1ptive tedmology. It shows, .ror examp1e, chat (\\'0 years af::eí

the 8-ínch drive was introduced, two-rhirds of the únr,s producing

;J.; • H Ji)' Grecu Ci'mpm:ic.< Cal! Paji

Figure 1.8 Leadership ofEntrant Finns in Disruptive Technology

Year

83 84

Year

Year

04---------~------~

85 91

D Entrant Firms D Established Firms

92 93

Year

Source: Data are from various issues of Disk/Trend Report.

ir (tour of six), \vere emranrs .• A.nd, t',YO years after the {irst 5.25-

inch drive \Vas imrodtlced, 80 percem of the finns producíng these

dísruptive drives \Vere enttants.

SUMMARY

There are several parterns in the hisrory of innovatíon in the disk drive índustry. The first is that the disruptive innovations were tech­nological1y straightforward. They general1y packaged known technologies in a unique archítecture and enabled the use of these

products in appEcarions where magnetíc data storage and rettieval previously had not been technologicaJly or economical1y feasible.

Hou> CaH Crear Pirms Fail? • 25

The second p<l~tem is rhat the purpasc of advanccd ct'chnology

de"dopmem III che industry \\'a5 always ro sl/sraill cstablished trajecro­

ries of performance improyernent: ro reaen rhe higher-pertormance,

highcr-marg:tn domún ofrhe upper righr ofthe trajectory map, Many

01" thesc technologies "'eré radicalJy ne,\" and dífficult, but rhey \Vere

nor dmupti\'e, The cu~tomer> oí the leadmg dí,k Llrive sllppliérs led

thcm tO\Yard rhe,e achle\'ements, Smtaining technologies. as a result.

did nar precipitare failure,

The thnd patteDl shmys tbar. despite the established fums' techno­logical prowess in leading sustaining inno\'ations, from the simplest

tO tht' musr radical, che finm tbe led rhe industrv in every insrance

of developing and J.dopting dlSrupriVC techno!ogies w:::re emra!1ts lO

che Ínousn:y. nOl ir;, incumbent Jeaden, This book began by posing a puzzle: \Vny ,,'as ir rhat fum5 chat

could be estcemed as aggn'5sí\'e. milOvatlve, customer-sensitive orga­

mzatiom (ou¡::! ignore or attend bebtedly lO rechnologicaJ innovatiollS

with enonnous str:negic impon;¡ncer In the context of the preceding

analym o1,he d,¡sk dri\ (' industry. rhis que,,¡on can be sharpe!1ed con­

slderabl)', Tu::: esrablished fimls \\'ere. in tacto agf;Tessh'e, innovatÍvc.

and cuswfficr-semíti\'e in their approaches LO smtaining innm'ations

oC e,'eryson, Bm the problem establi$hed tlnns sccm unable lO eonfront

successfull~, i; thar of dOllhlll'o1Td vis10n and mobiliry, in tenns of thr

trajectory map, Fíndmg ne\\ applic:mons and markt>ts for these new

product> s{:ems to be a c.lpability thar each of these fimls exlúbited

once, upon enrry. and rhen app.trently Jost. Ir ,YaS as ,ftÍle leadíng firms

\Vere heid capti\'e by their cU5tomers, enabling attackmg entram tlmls

ro topple eh!! incumben! indum-y !t~3der> eaeh ame a disrupth'e tt'ch­nology emerged, l' Why chis happ<"ned, alld is stiU happening, ís che

~ubject of the ne;.;t charrer,

APPENDIX 1.1: A NOTE ON THE DATA AND METHOD

USED TO GENERATE FIGURE 1.7

The trajectones mappea in Figure 1.7 wen' calculated .!S follows,

Data on rhe eapaci10' pro\"ided \,,'ith computers was obtamed from

16 • ¡¡'Ir}" Grt<lt CiJl/!pcmies Ccm Fa;1

Data Sow'ces, an anuual publication listíng the technical specífications

of all computer mode1s availabJe from e\'ery computer manufacturero For instances in which particular models were ayaJ.!able with ditterent

features and configurations. the manufacturer provided Dala 501/1"ces

wlth a "typical" system configuraríon with dehned random access

memory (RAM) capaCl'}', perllmnance specifications of peripheral equípment (inclurung disk drin:,s).list prices, and year ofintroducrion.

For instances in v,'hich a gi\'en computer model \Yas offered for sale' O\'er a sequence of \"ears, the hard disk eapacity prO\'ided in the typical configurarion '}'pícally increased. Dattl SOllr(es used [he caregories

mainframe. mini/m.idrange. desktop personal, portable and laprop,

and notebook. As of 1993. 1.8-ínch driYe> were nor being used in hand-held compurers, so no dara on rhat patemia! market eXlsted.

For Figure 1.7. ter each yea.r and each class of computers, a11 models avaiJable ter sale were ra.nked by price and the hard dü,k capacit\' pro\'ided \virh rhe median-prieed model identified. Thc best-fit línes through the resultanr time series \vere plorred as the

solid lines in Figure 1.7 for expository simplificaríon ro indicate rhe trend in t\'pical machines. In realit\', of cou,-se. there is a wide band

around [hese lines. The fTO/lTier of pertornlance---rhe híghest capacity

otlered with the mosr expensive computers-\vas substanti:lily higher

than the t\'pical \'a!ues sho,,'n.

The dorted lines in Figure 1.7 represent the best-fit line through the un\\"eíghted J\"Crage capaci'}' of all disk drives introduced for sale in each given architecrure fol' each year. This data \Yas rakcl1 írom

DisklTrelld Report. Again, for expository símplitication, only this average line Ís shown. There 'was a \nde band ofcapacities introduced for sale in each ;-ear, so thar the fronner or highesr capaci'}' drive introduced in each rear was substantially aboye the average shown.

Stated in anather \\"ay, a distÍnction must be made berween rhe full

range of products available ter purchase and rhose in t\'pical system.~. The upper and lowcr bands around the median and ayerage figures

shown in Figure 1.7 are generally para11el ro rhe lmes sho\'m.

Because higher capacity drives were available in the market than

were offered with rhe median-priced systems. the solid-line trajector­ies in Figure 1.7, as 1 state in the text, represem the capacities

HOll! Cal! Crear Firm5 Fail? • 27

"demanded" in eac:h market. In other \vords, the capacity per ma­chine was nor constraíned by technological availability. Rather, 11

represents the selection oC hard disk capacity by compurer users,

giVCll the prevaíling cost.

NOTES

1. A more complete lúswr\' of the disk drive ind~stry can be found in Clayton M. Christensen, "The Rígid Disk Drive

Industry: A History of ComrnerciaJ and Technologícal Tur­

buJence," BliSÍllCSS Hístory Re1'Íeu' (67), Wínter. 1993, 531·-588. 'lhis history foemes only 011 rhe faanufacrurers of rigíd disk drives or haTo dril'es-products on which data

are stored on rigid metal platters. Companíes .:nanufacmring floppy dl';k drivts (!emovabk diske-ttes o:'i1t:xible mybr coated with ¡ron oxide 011 which data are stOred) historically \Vere

dífferent nnm from those making hard d1Sk drive;.

2. Much of the da:.. for this analysis carne from Di"kITrend Report, a highly respectcd mnuallllarket research publicatíon, augmemed v:ith more detaileJ pracillct-speciiication shects

o!Jtained fmm the disk drive manufacturer, rhem,elves. I am

gratl'flll to lne editors and m.ff ar Dísk:'Trt:nd. lne., for their panem and generaus assistance in this pro.iect.

3. The COIlcept of trajt:ctories of technological progress was ex­amíned by Gioyanri Dosi in "Tedmologicai P;lradlg¡ns ana

T echnologícal T rajectories," RC.'earch Polie}' (11), 1982. H7-162,

4. Thc \vay:,:n which tlle findings ofthis ,tud)' differ from those of some earlier scholars of technolof,'Y change whíle building tlpon those of others are discussed in greater detaíl in chapter 2.

S. Th(' nm technology for makiIlg heads bui]t m elcc-rromagnet by wrapping a fine thread of copper \A'ire .l.wllnd a core of

iron oxide (íerritc); hence rhe termJmí!e head. Incremental

irnprovements to LJ"is approach involved leaming to grh'1d the

28 • rv/¡l' Crear Campa/lies Cml Fai]

ferrite to finer and finer dimensions, using better lapping techniques. and strengrhening the ferrite by doping it wirh barium. ThÍ/l-Jilll1 heads \wre made photolithographícally,

using rechnology similar to rhat used in making íntegrated

eircuits on silieon wafers to eteh the electromagnet on tbe sudaee of rhe head. Tbis was difficult beeause it itlvolved mueh rhicker layers of material rhan \Vere eommon in le manufacturing. The third technology, adopted starting in the

rnid-1990s. was called l1Iil,íilleto-resistÍl'E heads. These \Vere also made ,úth thin-film photolithography, bur used rhe principIe Óat changes in rhe magnetic flux fidd on the disk surfaee changed rhe eleerrieal reslStivity of the circuitry in the head.

By measunng ehanges in resístivity rather than ehanges in the dírecnon ofcurrent flow, magneto-resistive heads Viere muen more sensitive, and henee pennitted denser data reeording, than prior technology. In the evolution of disk technology. the earliest disks were made by coatíng fine needle-shaped particles of iron oxide--literally rus e-o ver rhe surface of a flat, polished aluminum platter. Hence, these disks were called oxide dísks. Incremental improvemems ro this technology in­

\'oh'ed making finer and finer iron oxide particles, and dispers­mg them more unifonnJy, ",ith fewer uncoated voíds on rhe aluminum platter's surta.ce. This \Vas supplanted by a sput­rering technology, also bOITowed from semiconductor pro­cessing. thar coated dle aluminum pIarter \Virh a thin liIm of metal a fe-w angstrorns thiek. The thinness of this layer; ¡rs cominuous, rarher than particulate nature; and rhe process's flexibiJity in deposíting magnetic materials \Vith higher coer­cívity. enabled denser recording Oll chin-fi1n1 disks than was t!::asible on oxide disks.

6. Richard J. Foster. 11ll1O~'ation: The Attacker's Advantage (New York: Summít Books, 1986).

7. The examples of technology change presented in Figures 1.1 and 1.2 introduce so me ambiguity ro the unqualified tenn discotltimlÍty, as used by Giovanm Dosi (see "Technological Paradigms and Technological Trajectories," Rcsearch Políq

Ho", Can Creat Firm.' Fail? • 29

[11] 1982), Michael L. Tushman and Phili.p Anderson (see

"1 cchnological D.!scontínuities and OrganizltÍonal Environ­

ments. ,. Admillistrutive SdCI1(e Q:Wl't('l'Íy [31], 1(86). and others.

TIlé' innovatlons in head md disk tech.nology describt'd in

Fígme 1.4 represent ]liJSlti!'e disco:1tinuities 111 an estabbhed

technologícal trajectory, while t,he tr~jectorv-disruptiI1p: tech­

nologics charted in Figure 1.7 represen uq;aríllf rusconúnu­

itie;. As will be showl1 helo\\', es:ablished flnm secmcd quite

capable oí' leadíng the indusrry over positive díscontillUlties,

~ut generally 105t their indusrry ¡cad when faced with negative

¿lscon::ínuities. 8, Thl<; tendency consistently appean across a range ofindustries.

Richard S. RosenblooD1 and Clanon M. Chri~rc'men (in

'TechnologicaJ Discondnuitíe~. Organizational Capabilities.

and Strategic Commitments," IlIduJ'triaí ,md Corporate Chang¡; [3], 1994, 655-(85) ~ugge>t a much broader ser ofindusrries in

which leading firms ma~' have been toppled by technologically

straightforward disrupti ve innovaw om rhan í$ covered in this

book.

9, A sl:mmary oftbe da:.). and procedures used to generate Figure

1. 7 is inciuded in Appendix 1.1,

¡ n. Thc minlcomputer n:arket was not ne\'.' in 1978, but it '-vas

a ncw apphcation tor \Vinchester-techuology disk dríves.

J 1, This s\atemem appEes OJ'l.,iy to mdependem drive makers com­

peting in (he OEM rnarket. So:ne of the vercically in~egrated

computer makers, such as JBM, have survived across these

generatloll'i with che benefit of a captive internal markel. Evcn

mM, bo\\·cver. adciressed rht> sequence 0:- difIerent emerging

l1larkets j()r disk drives by crearing amonomous "stan-up"

di~k dn"e o:-ganizations ro address ea eh one, It> San Jose

organizarío:1 focused 011 high- end (primarily rnainfmme) ap­

plications, A separate div1síon in Rochc',rer, MN, focusen

on míd-range comp1.1ters ;)nd \vorkstaoo!1S. rRM created a

different organization in Fujisawa, ]apa..'1, to prodllce drives

tor che Jesktop personal computer market, 11, T111s result is very ditTerent from thar obse:rved by Rebecca

M. Henderson (see TIle Failure ~r Established Firms in tlle Face oI Technological Chal1ge: A StHdy of [he Semiconductor Photalítho­

graphic Alignmellt lndustry. russertation, Harvard University.

1988). who found the ne"v-architecture aligners produced by the established manufacturers to be inferior in performance to those produced by entranr fimIs. One possible reason for ,hese differem results is thal the successful entrants in the photolithographic aligner industry studied by Henderson

brought to the ne'N product a well-developed body of techno­logícal kno\yledge and eJ.-perience developed and refined in orher markets. In the case studied here, none of the entrants brought such well-deyeloped knowledge "vith thern. Most, in fact, were de novo start-ups composed of managers and engíneers ,,,ho had defecred from established drive manufac­turing firms.

13. This fillding is similar to the phenomenon observed by Joseph L Bower, \,-,ho saw that explicit customer demands have tremendous power as a source of impetus in the resource allocation process: "When the discrepanc)' (the problem to

be solved by a proposed investment) \Yas defined in ternlS of

cost and quality, the projects languished. In all four cases, the definition process moved toward completion when capacíty to meet sales "yas perceived to be ínadequate .... [n ,hort, pressure [rom the market reduces both the probability and the COS! ofbeing i.vrong." Alrhough Bower specifically refers to manufacturing capacity, the same fundamental phenome­non-the power ofthe known needs ofknown customers in marshaling and dírecting the Ínvestments of a finn-affects

response to dísruptive technology. SeeJoseph L Bower, ¡Han­agi'¡g ¡he Resollrce Aflocatían PfDcess (Homewood, IL: Richard D. Irwín, 1970) 254.

14. In booking $113 million in revenues, Conner Peripherals set

a record for bookíng mOre revenues in its first year of operation than any manufacturing company in United States history.

15. This finding is consistent with what Robert Burgehnan has

observed. He noted that one of the greatest difficulties en-

How Can Crea! Fínns Fail? • 31

COtl:ltered by co;porarc entrcFene:':TS h2~ beeJ: finding the

nghr "!wta test $ltes" \Yhere produns cOllld be ¡nleractivdy

dt~w!oped and rd:ined with Cl'SLOmers. (;,:neJJ1r, a ne'.\" \'cn­

ture's eutréc to rh" ~m=o::)e~ \,;), pnn-¡ceci by the saksp<"r.,on

represemíag the tlm¡'s estab~ished product lim'" Tlus helped

rh" tir!!! cie\'elop ne\\" products fór es:abbh"d n1.1rk~t; but

n,n to ldemii}' new applicarions to!' new l<:chno;ogy. See

Robc:rl A. Burgelman :md Leollard 5ay]es, lllsidr Corpor<ltr

11l110Filn,11l (Nc'w York: The' free Pn':;s. 1986) 76-80.

}6, 1 bdiev(' t!us ;¡nJckmg Grms bve a:1 adv<H1tagc

in disnlptive innovJtiQrls bU( no: in stl~~a¡ning ones-darHies, but is nOI in conrllcr \Vah, remer's asse:-riom abom [he attack­

er' s ady.mtagl', The nHto:-:c.al exa;nplcs r oster mc, W mbstm­tÍatt> his rhrory gencralk sC'cm ro na\'c hcc!1 disruprive

mmn-arior:s Ser Richard J, Fü~wr. lWh1 l'allOl1: 'TIli: /lullikcr's

,'l<i".llIuge (1'>;n' Yo,k: Sumrmt Books, 1986),