1993-硬盘驱动器行业：一部商业与技术的动荡史

1993-硬盘驱动器行业：一部商业与技术的动荡史

Christensen C M. The rigid disk drive industry: A history of commercial and technological turbulence[J]. Business history review, 1993, 67(4): 531-588.

Christensen, C. M. (1993). The rigid disk drive industry: A history of commercial and technological turbulence. Business history review, 67(4), 531-588.

In its early years, the disk drive industry was led by a group of large-scale, integrated firms of the sort that Alfred D. Chandler, Jr., observed in his studies of several of the world’s largest industries. The purpose of this history is to explore why it was so difficult for the leading disk drive manufacturers to replicate their success when technology and the structure of markets changed. The most successful firms aggressively developed the new component technologies required to address their leading customers’ needs, but this attention caused leading drive makers to ignore a sequence of emerging market segments, where innovative disk drive technologies were deployed by new entrants. As the performance of these new-architecture products improved at a rapid pace, the new firms were eventually able to conquer established markets as well. As a consequence, most of the integrated firms that established the disk drive industry were driven from it, displaced by networks of tightly focused, less integrated independent companies.

早年，磁盘驱动器行业是由一些大型一体化企业主导的，小阿尔弗雷德-钱德勒（Alfred D. Chandler, Jr.）在研究世界上一些最大的行业时注意到了这一点。这段历史的目的在于探讨，为什么当技术和市场结构发生变化时，领先的磁盘驱动器制造商很难复制他们的成功。最成功的公司积极开发新的组件技术，以满足其主要客户的需求，但这种关注导致领先的硬盘制造商忽视了一系列新兴的细分市场，在这些市场中，新进入者部署了创新的硬盘技术。随着这些新架构产品性能的快速提升，新公司最终也征服了成熟市场。因此，建立磁盘驱动器行业的大多数一体化公司都被赶出了该行业，并被重点突出、集成度较低的独立公司网络所取代。

From the beginnings of the computer industry, engineers have wrestled with the challenge of storing and retrieving information. Users have wanted to store more information and access it more rapidly and to do so at decreasing cost. The effectiveness with which computer and peripheral equipment manufacturers responded to these demands has been an important factor in the growth of the computer industry and has enabled users to apply computing power in a broadening range of tasks.

从计算机行业诞生之初，工程师们就一直在与存储和检索信息的难题作斗争。用户希望存储更多的信息、更快速地访问这些信息并降低成本。计算机和外围设备制造商对这些需求的响应效率一直是影响计算机行业发展的重要因素，还使用户能够在越来越广泛的任务中应用计算能力。

Information storage and retrieval capabilities of computing systems comprised semiconductor (formerly magnetic-core) memory circuits and peripheral magnetic storage and retrieval devices, including tape, rigid (hard) disk, and floppy disk drives. Although the history of each of these technologies is rich, the history of the companies that developed rigid disk drives seems to have been particularly complex and tumultuous. The rigid disk drive industry grew from a research project begun in the San Jose laboratories of International Business Machines, Inc. (IBM) in 1956 to a $15 billion industry in 1990. Of the 138 firms known to have entered the industry in this period, 103 subsequently failed, and six others disappeared through acquisition or absorption by competitors.

计算机系统的信息存储和检索功能由半导体（以前的磁芯）存储电路和外围磁性存储和检索设备组成，包括磁带、硬盘和软盘驱动器。尽管每种技术的历史都很丰富，但开发硬盘驱动器的公司历史似乎尤为复杂和动荡。硬盘驱动器行业从1956年在美国国际商业机器公司（IBM）圣何塞实验室开始的一个研究项目发展到1990年价值150亿美元的产业。在此期间，已知进入该行业的138家公司中，有103家最终失败了，另有6家被竞争对手收购或吞并而消失。

New firms entered to lead the industry in four of its six technologically defined product generations. The demise of the leading firms of each generation seems to have been triggered by the emergence of new product architectures and of new market segments in which these architectures were used. This history of the rigid disk drive industry therefore is focused on the emergence of these new technologies and markets.

在六代由技术定义的产品中，有四代是由新公司进入并引领行业发展的。每一代领先企业的消亡似乎都是由新产品架构和使用这些架构的新细分市场的出现引发的。因此，这部硬盘驱动器行业发展史主要集中在这些新技术和新市场的出现上。

This article examines those firms that design and manufacture rigid magnetic disk drives for sale in the original equipment (OE) market to computer manufacturers. Because this study’s emphasis is on the interactions between technological developments and market forces in the disk drive industry, primarily the open-market disk drive activities of vertically integrated computer manufacturers such as IBM and Control Data—not their internal, intra-corporate disk drive transactions—are considered here.

本文探讨了设计和制造硬盘驱动器并在原始设备 (OE) 市场上将其销售给计算机制造商的企业。由于本研究的重点是磁盘驱动器行业的技术发展与市场力量之间的相互作用，因此这里主要考虑的是垂直一体化计算机制造商（如IBM和Control Data）的公开市场磁盘驱动器活动，而非其公司内部的磁盘驱动器交易。

By 1990 rigid disk drive production was a worldwide industry populated by multinational firms with headquarters in twelve countries spanning four continents. Several produced drives outside the country where their headquarters were located; for example, Rodime, a Scottish firm, produced drives in Florida; IBM and Quan tum both manufactured drives in California and Japan; the manufacturing operations of most U.S. firms were centered in Singapore; and some Japanese firms manufactured drives in the United States. References in this article to the“United States,” “Japanese,” or “European” disk drive industries thus relate to those groups of firms whose headquarters are in those countries.

到1990年，硬盘驱动器生产已成为跨国公司云集的全球性产业。这些跨国公司的总部遍布四大洲十二个国家。一些公司在总部所在国以外的地方生产硬盘；例如，苏格兰公司Rodime在佛罗里达生产硬盘；IBM和Quan-tum在加利福尼亚和日本生产硬盘；大多数美国公司的制造业务集中在新加坡；一些日本公司在美国生产硬盘。因此，本文中提到的“美国”、“日本”或“欧洲”磁盘驱动器行业指的是总部设在这些国家的企业集团。

The Emergence of the Industry

行业的兴起

Technological Definition of the Disk Drive at IBM

• Almost all development of magnetic information-recording technology through the mid-1960s occurred at IBM’s San Jose laboratories. Engineers there guided the company through six distinct generations of magnetic recording products prior to the emergence of a dominant product design: magnetic drums; 0.5-inch reel-to-reel tape; moving-head fixed-disk drives; rigid removable disks and disk packs; flexible (floppy) removable diskettes; and, finally, sealed, non-removable “Winchester” rigid disks.

IBM磁盘驱动器的技术定义

• 20世纪60年代中期，几乎所有磁性信息记录技术的开发都是在IBM的圣何塞实验室进行的。在主导产品设计出现之前，那里的工程师指导公司经历了六代不同的磁性记录产品：磁鼓、0.5英寸卷对卷磁带、移动磁头固定磁盘驱动器、硬性可移动磁盘和磁盘组、柔性可移动软磁盘（软盘）；最后是密封的不可移动的“温彻斯特”硬盘。

The earliest peripheral magnetic information-storage devices were magnetic drums—an architecture similar in concept to Thomas Edison's early phonograph cylinders. The drums were developed at IBM in the late 1940s and until the mid-1950s were the primary storage devices used with early computing machines. Magnetic drum technologies gave way to magnetic tape in the mid-1950s, and through 1963 tape was the dominant data storage and retrieval medium. The primary drawback of tape storage was access time: if a user needed to access data at the end of a tape, the drive had to read through the entire tape before it could access the desired file. Moreover, changing a record anywhere within the tape required re-recording the entire tape.

最早的外围磁性信息存储设备是磁鼓——一种在概念上与托马斯-爱迪生的早期留声机圆筒相似的结构。磁鼓于20世纪40年代末在IBM公司开发，直到20世纪50年代中期一直是早期计算机使用的主要存储设备。20世纪50年代中期，磁鼓技术让位于磁带技术。到 1963年，磁带成为主要的数据存储和检索介质。磁带存储的主要缺点是访问时间长：如果用户需要访问磁带末端的数据，驱动器必须读完整个磁带才能访问所需的文件。此外，更改磁带中任何位置的记录都需要重新记录整个磁带。

IBM’s efforts to address its customers’ thirst for increased processing speeds bore fruit in 1956, when the company shipped its first moving-head magnetic disk drive, called RAMAC—an acronym for Random Access Method for Accounting and Control. The use of rotating disks in the IBM RAMAC represented a major change in engineering thinking for the magnetic information-storage industry. In all previous generations, the need to position the read-write head precisely led engineers to fix it rigidly in place and to move the magnetic media—drums, tapes, or strips—past it. In contrast, the RAMAC drive had a movable head, positioned in the first product 0.0008 inch above the disk’s surface via a hydrostatic air bearing. The RAMAC drive was a mechanical marvel, with one moving arm and head for each of fifty 24-inch diameter disks in the drive.

1956年，IBM为满足客户对提高处理速度的渴求而做出的努力结出了硕果，公司推出了首款移动磁头磁盘驱动器，名为RAMAC（Random Access Method for Accounting and Control）。在IBM的RAMAC中使用旋转磁盘代表了磁性信息存储行业工程思想的重大变革。在前几代产品中，由于需要精确定位读写磁头，工程师们只能将其固定在原位，并将磁性介质——磁鼓、磁带或磁条——从其上方移动。与此相反，RAMAC驱动器的读写头是可移动的，通过静压空气轴承定位在磁盘表面上方0.0008英寸的第一个产品。磁头RAMAC驱动器是一个机械奇迹，硬盘中50个24英寸直径的磁盘各自有一个活动臂和一个磁头。

In 1962 the fourth generation of magnetic storage devices—the removable disk pack—surpassed the RAMAC’s fixed-disk architecture in capacity and flexibility. By interchanging packs of rigid disks, users could store far more data than was possible in a fixed-disk system. The disk pack was the industry’s dominant architectural design for more than a decade, and it was the product vehicle that most of the early participants in the original equipment market used to enter the disk drive industry. In 1971, IBM introduced the first drive using removable, flexible (floppy) diskettes to enable more efficient offline storage and loading of the proliferating number of specialized routines and programs for its mainframe computers, where frequency of use did not justify permanent residence for those programs in core memory. The original Model FS33 floppy disk drive was a read-only device, but a read-write version followed in 1973.

1962年，第四代磁性存储设备——可移动磁盘组——在容量和灵活性上超越了 RAMAC的固定磁盘架构。通过交换硬盘组，用户可以存储比固定磁盘系统多得多的数据。在长达十多年的时间里，磁盘组一直是行业的主流架构设计，原始设备市场的大多数早期参与者都是利用它作为产品载体进入磁盘驱动器行业的。1971年，IBM推出了第一款使用可拆卸软盘的驱动器，以便更有效地离线存储和加载其大型计算机中不断增加的专用例程和程序，因为这些程序的使用频率并不适合永久保存在核心内存中。最初的FS33型软盘驱动器是一种只读设备，但在1973年推出了读写版本。

IBM’s Model 3340—a sealed rigid 14-inch disk drive introduced in 1973 and dubbed the Winchester—was IBM’s crowning architectural achievement in magnetic storage. Over the next decade the Winchester design was adopted throughout the world industry. Competing firms have incrementally improved, but have not yet radically altered, the fundamental Winchester design. In the diskpack architecture, particulate contamination and the removability of the disks prevented close head-to-disk spacing, which inherently limited improvements in recording density. The Winchester drive addressed these issues by permanently sealing the disks with heads, motors, actuators, and electronics inside a dust-proof drive housing. This enabled IBM to reduce the height at which the head flew over the disk surface to 0.000008 inch—a height one-thousandth of the head-to-disk distance in the RAMAC drive. The cost per megabyte of Winchester drives was 30 percent less than the cost of disk-pack drives of equivalent capacity.

IBM的3340型磁盘驱动器于1973年推出，是一款密封的14英寸硬盘驱动器，被称为“温彻斯特”，它是IBM在磁性存储领域的最高架构成就。在接下来的十年中，温彻斯特的设计被世界工业界广泛采用。竞争公司逐步改进了温彻斯特的基本设计，但尚未从根本上加以改变。在磁盘组结构中，微粒污染和磁盘的可拆卸性阻碍了磁头到磁盘间距的缩小，从而限制了记录密度的提高。温彻斯特硬盘通过将磁盘、磁头、电机、致动器和电子元件永久密封在防尘硬盘外壳内来解决这些问题。这使得IBM能够将磁头飞过磁盘表面的高度降低到0.000008英寸，这一高度是RAMAC硬盘磁头到磁盘距离的千分之一。温彻斯特硬盘的每兆字节成本比同等容量的磁盘组硬盘低30%。

The Rise of Plug-Compatible Equipment Manufacturers

• Until IBM introduced its disk-pack architecture in 1962, it was the only disk drive manufacturer in the world. But in the early 1960s several companies that had been founded to make add-on tape storage systems that were plug-compatible (the computer equivalent of interchangeability of parts) with IBM equipment copied IBM’s disk-pack drive concept and began marketing plug-compatible drives directly to users of IBM computer systems—marking the beginning of an industry of independent disk drive manufacturers.

插头兼容设备制造商的崛起

• 在IBM于1962年推出磁盘组架构之前，IBM是全球唯一一家磁盘驱动器制造商。但在20世纪60年代初，几家为生产与IBM设备插头兼容（相当于计算机部件之间的互换性）的附加磁带存储系统而成立的公司模仿了IBM的磁盘组驱动器概念，开始直接向IBM计算机系统的用户销售插头兼容的驱动器——这标志着独立磁盘驱动器制造商行业的开端。

The market for IBM plug-compatible disk drives was pioneered by eleven firms, most of which diversified into disk technology from earlier positions in other magnetic recording product markets. Eight had been tape drive manufacturers: Telex Corp. (Tulsa, Okla.); Storage Technology Corp. (Boulder, Colo.); the peripherals division of Control Data Corp. (Minneapolis, Minn.); and five smaller Los Angeles-area firms: Century Data, International Storage Systems (ISS), Pertec, Wangco, and Kennedy.7 These firms were joined by Memorex, which had been the leading supplier of magnetic tape (but not drives) since it was founded in 1961, and by two start-up companies, Iomec and Caelus.

IBM插头兼容磁盘驱动器市场由11家公司开创，其中大多数公司都是从早期的其他磁记录产品市场转向磁盘技术的。其中8家曾是磁带驱动器制造商：Telex公司（俄克拉荷马州塔尔萨市）、Storage Technology公司（科罗拉多州博尔德市）、Control Data公司（明尼阿波利斯市）的外围设备部门以及洛杉矶地区的五家较小公司：Century Data、International Storage Systems (ISS)、Pertec、Wangco和Kennedy。加入这些公司行列的还有1961年成立以来一直是磁带（但不是驱动器）领先供应商的Memorex公司，以及两家新成立的公司Iomec和Caelus。

For the plug-compatible equipment makers, IBM was not just a competitor; it was the environment. Plug-compatible manufacturers could be product imitators, but not innovators. They sold tape and disk drives directly to users of IBM computers who needed additional or replacement data storage capability, pricing their products beneath IBM’s substantial price umbrella (they priced their drives between $8,000 and $12,000, compared to IBM’s Model 1311 price of $26,000) and offering whatever performance advantages they could within the constraints of IBM specifications. By the late 1960s the plug-compatible disk drive business was booming; the market had reached $100 million by 1970; $250 million by 1976; and nearly $700 million in its peak year, 1985. IBM and the group of plugcompatible manufacturers each consistently accounted for 20-30 percent of worldwide disk drive shipments through the 1970s. But the proportion of total drives made by plug-compatible manufacturers declined from 27 percent in 1976 to under 7 percent by 1987(see Table l). The firms that captured an increasing share of the world disk drive market during this period were independent suppliers of disk drives in the OE market. Table 1 also shows that overall industry revenues grew at an average 22 percent annual rate between 1976 and 1989. Unit shipments over the same period averaged 34 percent annual growth.

对于插头兼容设备制造商来说，IBM不仅仅是一个竞争对手，更是一个环境。插头兼容设备制造商可以说是产品模仿者，但不是创新者。他们将磁带和磁盘驱动器直接销售给需要额外或替代数字储存能力的IBM计算机用户，将其产品定价在IBM强大的价格保护伞之下（与IBM的1311型售价2.6万美元相比，他们的驱动器定价在8,000美元到1.5万美元之间），并在IBM规格的限制下提供了它们所能提供的任何性能优势。到20世纪60年代末，插头兼容驱动器业务蓬勃发展；到1970年，市场规模已达1亿美元；到1976年，达到2.5亿美元。1985年达到最高峰，销售额达7亿美元。整个20世纪70年代，IBM和插头兼容制造商集团一直各占全球磁盘驱动器出货量的20-30%。但插头兼容制造商生产的磁盘驱动器占总量的比例从1976年的7%下降到1987年的不足7%。在此期间，在世界磁盘驱动器市场上占有越来越大份额的公司都是原始设备市场上磁盘驱动器的独立供应商。表1还显示，1976年至1989年期间，行业总收入的年平均增长率为22%。同期的单位出货量平均年增长率为34%。

The Development of the Disk Drive Industry for the OE Market

• It was the explosive growth of the minicomputer industry in the 1970s that spawned the original equipment market for disk drives. Throughout the 1960s and 1970s, the larger computer manufacturers—IBM, Digital Equipment, Control Data, Data General, Burroughs, Fujitsu, Hitachi, and Univac—made most of their own drives. It was the growth of computer manufacturers that were not as extensively integrated—firms like Wang, Prime, NCR, and Nixdorf in the 1970s, and Apple, Commodore, Compaq, Tandy, and Sun Microsystems in the 1980s—that created a major OE market for independent disk drive manufacturers. The original equipment market differed from the plug-compatible market primarily in that disk drive manufacturers sold to computer manufacturers, rather than to users of computers.

面向原始设备市场的磁盘驱动器产业的发展

• 正是20世纪70年代微型计算机行业的爆炸式增长催生了磁盘驱动器的原始设备市场。在20世纪60年代和70年代，大型计算机制造商——IBM、Digital Equipment、Control Data、Data Cen-eral、Burroughs、Fujitsu、Hitachi和Univac——制造了自己的大部分硬盘。恰恰是那些没有高度集成化的电脑制造商的增长——20世纪70年代的Wang、Prime、NCR和Nixdorf等公司，以及80年代的Apple、Commodore、Compaq、Tandy和Sun Microsystems等公司——为独立磁盘驱动器制造商创造了一个重要的原始设备市场。原始设备市场与插头兼容市场的主要区别在于，其销售对象是计算机制造商，而非计算机用户。

The robust OE market demand described in Table 1 attracted a host of entrant firms. In addition to the eleven makers of plugcompatible products previously mentioned, at least eighty-seven other firms entered the OE market fray between 1975 and 1989. For the purposes of this article, they can be grouped into five categories:

表1所述的旺盛的OE市场需求吸引了众多公司加入。除了前面提到的11家插头兼容产品制造商外，在1975年至1989年期间，至少还有87家公司进入了原配套设备市场。就本文的写作目的而言，可将这些公司分为五类：

Start-Up Firms were generally venture capital-backed companies founded to design and manufacture rigid disk drives. Most focused almost exclusively on the disk drive business. Although a few start-ups had entered between 1965 and 1970 to manufacture plug compatible drives, by 1973 this early group of firms had all been acquired by larger, diversified firms. The vast majority of start-ups entered after 1978.

初创公司通常是由风险资本支持的公司，成立的目的是设计和制造硬盘驱动器。大多数公司几乎只专注于磁盘驱动器业务。尽管1965年至1970年间，一些新成立的公司开始生产插头兼容驱动器，到1973年，这批早期公司已全部被规模较大的多元化公司收购。绝大多数新公司都是在1978年之后进入市场的。

Related-Technology Firms were diversified concerns that entered by adapting magnetic data-recording technologies they had developed in other product-market contexts to rigid disk drive products. Examples include Storage Technology Corporation (from tape drives) and Ampex (from audio and video recorders).

相关技术公司是通过将在其他产品市场上开发的磁性数据记录技术应用于硬盘驱动器产品而进入市场的多元化企业。例如存储技术公司（磁带驱动器）和Ampex公司（录音机和录像机）。

Related-Market Firms such as Memorex, Diablo, Perkin Elmer, and Calcomp were diversified concerns for which disk drives were one of several product lines made for the burgeoning computer industry. In contrast to the related-technology firms, for which disk drives represented a business technologically related to their previous activities, the related-market firms had expanded into disk drive production in a strategy related to markets and customers.

相关市场公司，如Memorex、Diablo、Perkin Elmer和Calcomp都是多元化企业，磁盘驱动器是它们为新兴的计算机行业生产的几种产品系列之一。与相关技术公司不同的是，相关市场公司的磁盘驱动器业务在技术上与它们以前的业务相关，它们通过与市场和客户相关的战略扩展到磁盘驱动器生产领域。

Forward Integrators began by manufacturing critical disk drive components such as read-write heads or controllers, and then integrated forward to the design and assembly of complete disk drives.

前向一体化公司从制造读写磁头或控制器等关键硬盘驱动器部件开始，然后向设计和组装完整磁盘驱动器的方向发展。

Vertically Integrated Computer Manufacturers historically produced a large proportion of the world’s disk drives (see Table 1). Some, such as IBM and Digital Equipment, generally manufactured for internal, captive consumption. Others, such as Control Data, Fujitsu, and Hitachi, always competed actively in the OE market in addition to supplying some captive needs.

垂直一体化计算机制造商在历史上生产了全球大部分磁盘驱动器（见表1）。其中一些厂商，如IBM和Digital Equipment，通常为内部自用而生产。其他厂商，如控制数据公司（Control Data）、富士通和日立除了供应一些自用需求外，还一直在OE市场积极竞争。

Table 2 describes the entry patterns of each of these groups of firms in the United States, as well as those of independent Japanese and European firms. Eighty-seven U.S.-based firms entered the OE market, compared to thirty-one Japanese and thirteen European entrants. The total number of active U.S. firms rose from twenty-two in 1976 to fifty-four in 1983, and then declined to twenty-seven by 1989. The number of Japanese participants peaked two years after the U.S. high point. European firms were never important factors in the world industry.

表2介绍了美国这两类公司的进入模式，以及日本和欧洲独立公司的进入模式。进入 OE市场的美国公司有87家，而日本公司有31家，欧洲公司有13家。活跃的美国公司总数从1976年的22家上升到1983年的54家，然后到1989年下降到27家。日本参与企业的数量在美国达到最高点两年后达到顶峰。而欧洲公司从来都不是世界工业的重要因素。

The bottom section of Table 2 charts the shares of the world original equipment/plug-compatible market claimed by these groups of firms. Clearly, the rigid disk drive industry was not an industry pioneered by classic venture capital-backed Silicon Valley start-up firms. The pioneering firms were the larger, diversified concerns that entered the disk drive industry on the basis of technological relatedness, market relatedness, or by backward integration from computer manufacturing. Most of these disk drive manufacturers were further vertically integrated into the manufacture of components such as heads, disks, and motors that were employed in their drives. This pattern of integration was consistent with those that Alfred D. Chandler, Jr., observed: in the industry’s formative years, the need to assure and coordinate the availability of key components gave vertically integrated firms a strong competitive advantage.

表2底部列出了这些公司在全球原始设备/插头兼容市场上所占的份额。显然，硬盘驱动器行业并不是由典型的风险资本支持的硅谷初创公司开创的行业。先驱企业都是规模较大的多元化企业，它们凭借技术相关性、市场相关性，或通过从计算机制造向后整合而进入磁盘驱动器行业。这些磁盘驱动器制造商大多进一步垂直整合，生产磁头、磁盘和驱动器所用电机等部件。这种整合模式与小阿尔弗雷德—钱德勒（Alfred D. Chandler, Jr.）所观察到的情况一致：在该行业的成长期，由于需要确保和协调关键部件的供应，垂直整合的企业具有强大的竞争优势。

Much of the entry and exit activity summarized in Table 2 occurred at the industry’s periphery, among firms that never became commercially viable. But the turbulence permeated the ranks of the industry’s largest firms as well. The combined world market share of the integrated firms, which was 84.8 percent in 1978, had declined to 7.9 percent by 1989. The older firms were essentially driven from the market by the start-ups. These start-ups, which claimed less than 1 percent of the world market in 1979, accounted for more than half of it a decade later. In 1989, seven of the world original equipment/plug-compatible market’s ten largest participants were U.S.-based start-up firms.

表2中总结的大部分进入和退出活动都发生在该行业的边缘，发生在那些从未在商业上可行的公司中。但是，动荡也渗透到了该行业最大的公司中。1978年，企业的全球市场份额合计为84.8%，到1989年已下降到7.9%。老公司基本上被新公司赶出了市场。这些新公司在1909年占世界市场的份额不到1%，但十年后却占据一半以上。1989年，在世界原始设备/插头兼容市场的十大参与者中，有七家是总部设在美国的新成立公司。

Table 3 offers a closer look at the ten largest disk drive manufacturers in the original equipment market in selected years between 1976 and 1989. The firms that led the industry in 1976 are shown in regular type; entrants to the industry after 1976 are listed in bold face. The entire initial population of leaders had disappeared by 1989. The decline of Control Data was particularly spectacular. Its market share, built largely on 14-inch disk-pack and Winchester drives, peaked at 62 percent in 1980-81. Its share had declined to 21 percent by 1988, before it was acquired by Seagate Technology, the new industry leader.

表3提供了1976年至1989年期间选定年份原始设备市场上最大的十家磁盘驱动器制造商的详细情况。1976年行业领先的公司用普通字体表示；1976年后进入行业的公司用黑体字表示。到1989年，最初的领先企业已全部消失。Control Data公司的衰落尤为惊人。其市场份额主要建立在14英寸磁盘组和温彻斯特硬盘上，1980-1981年达到62%的峰值。到1988年，它的市场份额下降到21%，随后被行业的新领头羊希捷科技收购。

The misfortune of the initially dominant integrated disk drive manufacturers reflected in Table 3 seems to have typified the fate of the other integrated U.S.-based manufacturers that entered the market in its later years. Only three (11 percent) of the twenty-eight related-technology, related-market, forward-integrator, and vertically integrated firms that entered between 1976 and 1989 survived to the end of the period. In contrast, sixteen (40 percent) of the forty start-up entrants survived.

表3中所反映的最初占主导地位的磁盘驱动器制造商的不幸遭遇，似乎也是晚年进入市场的其他美国一体化制造商的命运。在1976年至1989年间进入市场的28家相关技术、相关市场、前向一体化和垂直一体化公司中，只3家（11%）存活到这一时期结束。相比之下，40家新创企业中有16家（40%）存活了下来。

Without exception, the start-ups that grew to dominate the world industry were focused exclusively on manufacturing rigid disk drives—they made no other products. Furthermore, the start-ups that successfully entered later in the period were progressively less vertically integrated than those that had entered earlier. This trend is illustrated in Table 4, which lists the five companies with the largest total cumulative disk drive sales in the industry’s history. The early leading manufacturers, IBM and Control Data, were completely integrated into the manufacturing of critical components and even into the research activities required to support new component development. Seagate, the dominant firm through the 1980s, was integrated quite extensively into component manufacturing, but its commitment to research in support of advanced component development was not nearly as extensive as that maintained by IBM and Control Data. Conner Peripherals was much less vertically integrated: its only commitment to component development and manufacture was its acquisition in the late 1980s of a small firm that made drive heads that glided on the surface of disks. Quantum, the most recent entrant, was the least integrated of all—it sourced all of its components in the external supplier market and had a Japanese partner, Matsushita, manufacture its drives by contract. This reversal of fortunes—the inability of the integrated firms that created the industry to maintain their leadership—might in some sense be considered “un-Chandlerian”. The ability of the successful firms that Chandler studied to assure supplies of critical materials and to control distribution and sales through a vertically integrated infrastructure seems to have been the key to those firms’ sustained competitive success. In the cases studied here, integration progressively seems to have become a disadvantage as the industry matured. A primary objective of the research reported in the remaining pages of this history is to develop a deeper understanding of why this reversal of fortunes between integrated and start-up firms occurred.

在世界产业中占据主导地位的初创企业无一例外，都只专注于制造硬盘驱动器，而不生产其他产品。此外，与较早进入该行业的公司相比，较晚进入该行业的新公司的垂直一体化程度逐渐降低。表4显示了这一趋势，其中列出了行业历史上磁盘驱动器累计总销售额最高的五家公司。早期的领先制造商IBM和Control Data都是由两家公司合并而成。完全集成到关键部件的制造中，甚至集成到开发新部件所需的研究活动中。希捷公司（Seagate）是整个20世纪80年代的主导企业。康纳外围设备公司（Conner Peripherals）在元件制造方面的整合相当广泛，但其在支持先进元件开发方面的研究投入远不如IBM和Control Data 公司。Conner Peripherals公司的垂直一体化程度要低得多：它唯一致力于部件开发和制造方面的是在20世纪80年代末收购了一家小公司，该公司生产可在磁盘表面滑动的驱动头。量子公司（Quantum）是最近进入市场的公司，它的一体化程度最低——它从外部供应商市场采购所有部件，并与日本合作伙伴松下公司（Matsushita）签订合同生产驱动器。这种命运的逆转——创造了这一产业的一体化公司无力保持其领导地位——在某种意义上可能被认为是“非钱德勒式的”。钱德勒所研究的成功企业有能力确保关键材料的供应，并通过垂直一体化的基础设施控制分销和销售，这似乎是这些企业在竞争中取得持续成功的关键。在本文研究的案例中，随着工业的成熟，一体化似乎逐渐成为一种劣势。本历史报告剩余篇幅中的一项主要研究目标是，深入了解一体化企业和初创型企业之间命运逆转的原因。

The Development and Diffusion of New Technology in the Disk Drive Industry

磁盘驱动器行业新技术的发展与传播

In an industry as technologically turbulent as the disk drive industry, the relative abilities of large, integrated firms and small, focused firms to respond to rapidly changing technologies and markets is a natural issue to explore. Indeed, disk drive technology changed very rapidly between 1976 and 1989. Table 5 summarizes improvements in the performance of disk drives along several important dimensions over the period. Recording density increased by a factor of 20; average data access time fell 38 percent; and the physical volume occupied by the smallest available 20 megabyte (MB) drive in 1989 was 0.0015 of the volume required for the same capacity in 1977.

在磁盘驱动器行业这样一个技术动荡的行业中，大型一体化企业和小型专业化企业应对快速变化的技术和市场的相对能力自然是一个值得探讨的问题。事实上，从1976年到1989年，磁盘驱动器技术的变化非常迅速。表5总结了这一时期磁盘驱动器在几个重要方面的性能改进情况。记录密度提高了20倍；平均数据访问时间缩短了38%；1989年，最小的20兆字节（MB）硬盘所占的物理容量是为1977年相同容量所需体积的0.0015。

Enormous decline in the cost per megabyte of memory was an important industry growth driver (see Fig. 1). The average price per megabyte of memory in the OE market declined from $560 in 1977, when only 33 cumulative terabytes had been shipped since the industry's inception, to $6.60 per megabyte in 1989, by which time over 6,000 cumulative terabytes had been shipped. The experience curve fitted through the points declined along a 53 percent slope, meaning that every time cumulative output in the industry doubled, costs per megabyte declined to 53 percent of their former level.

每兆字节内存成本的大幅下降是行业增长的重要推动力（见图1）。OE市场上每兆字节内存的平均价格从1977年的560美元下降到1989年的6.6美元，而累计出货量也从该行业自成立以来仅有的33TB，到已超过6000TB。通过各点拟合的经验曲线沿着53%的斜率下降，这意味着该行业的累计产量每翻一番，每兆字节的成本就会下降到原来水平的53%。

A useful framework for characterizing changes in disk drive technologies was proposed by Rebecca Henderson and Kim Clark. They posit that, in general, a product’s fundamental technological approach is embodied in its components, which are designed into a system architecture. Henderson and Clark classify innovations by the degree to which they reinforce or render obsolete firms’ expertise along these two dimensions—component technology and architectural design. They conclude from this logic that there are four distinct types of innovation, as shown in Figure 2.

丽贝卡-亨德森（Rebecca Henderson）和金-克拉克（Kim Clark）提出了一个有用的框架来描述磁盘驱动器技术的变化。他们认为，一般来说，产品的基本技术方法体现在其组件中，而这些组件被设计成一个系统架构。他们根据创新在组件技术和架构设计这两个维度上强化或淘汰企业专业技术的程度对创新进行了分类。根据这一逻辑，他们总结出四种不同类型的创新，如图2所示。

Incremental innovation is defined as any change that builds on a firm’s expertise in component technology and that occurs within its established product architecture. An example of incremental innovation would be the development of a faster electric spin motor driving the rotation of disks. Modular innovation occurs when a new core technology, embodied in a component, is “plugged” into a fundamentally unchanged system architecture—as when a thin-film head is substituted for a ferrite head in a 5.25-inch Winchester disk drive. Architectural innovation leaves the core technological concepts of the components intact but changes the way they are designed to work together, as in the change from removable disk packs to the sealed Winchester disk drive architecture. Radical innovation, exemplified by the emergence of optical (as opposed to magnetic) disk drives, involves change along both dimensions.

渐进式创新的定义是，以公司在组件技术方面的专长为基础，在其既定产品架构内进行的任何变革。一个渐进式创新的例子是开发出更快的驱动磁盘旋转的电动旋转电机。模块化创新是指将体现在组件中的新核心技术“嵌入”到基本不变的系统架构中，例如在5.25英寸温彻斯特磁盘驱动器中用薄膜磁头取代铁氧体磁头。结构性创新保留了部件的核心技术要素，但改变了它们的协同工作方式，如将可移动磁盘组改为密封的温彻斯特磁盘驱动器结构。以光学（相对于磁性）磁盘驱动器的出现为例，激进式创新涉及两个方面的变化。

Examples of each type of technological change can be found in the disk drive industry between 1973 and 1989. There were modular changes in every significant disk drive component. Disks coated with thin metal films substituted for disks coated with particles of iron oxide; heads made through photolithographic processes substituted for heads made by winding copper wire around a machined ferrite core; and the codes in which data were recorded came to employ more economical, space-conserving concepts. And there were innumerable incremental improvements. In ferrite heads, for instance, the development of barium-doped ferrite greatly increased the physical strength of the material, permitting the heads to be ground to much finer, more precise dimensions without chipping or cracking. The development of lapping processes permitted manufacturers to grind the ferrite cores even more finely. And depositing a strip of metal in the gap separating the leading and trailing surfaces of the head increased the strength of the magnetic field generated by the ever-shrinking ferrite head.

1973年至1989年期间的磁盘驱动器行业中可以找到每种技术变革的例子。磁盘驱动器的每个重要部件都发生了一些变化。涂有金属薄膜的磁盘取代了涂有氧化铁颗粒的磁盘；通过照相平版印刷工艺制作的磁头取代了通过在机加工的铁氧体磁芯上缠绕铜线制作的磁头；记录数据的编码采用了更经济、更节省空间的概念。此外，还有无数的渐进式改进。以铁氧体磁头为例，掺钡铁氧体的开发大大提高了材料的物理强度，使磁头可以磨削到更精细、更精确的尺寸，而不会出现崩裂或开裂。研磨工艺的发展使制造商可以将铁氧体磁芯磨得更细。在磁头前表面和后表面之间的缝隙中沉积金属条，可以增强不断缩小的铁氧体磁头产生的磁场强度。

Between 1973 and 1989, five waves of major architectural change swept through the disk drive industry. The first, in which 14-inch Winchester disk drives substituted for removable disk-pack drives, has already been described. Each of the subsequent architectural generations was associated with a reduction in size within the Winchester paradigm—from 14 to 8 inches in 1978; from 8 to 5.25 inches in 1980; to 3.5 inches in 1985; and to 2.5 inches in 1989. Although each of these downsizings involved shrinking the size of the components used, they also each involved significant redesign of the way components interacted within the architecture.

从1973年到1989年，磁盘驱动器行业经历了五次重大架构变革。第一次浪潮中，14英寸温彻斯特硬盘驱动器取代了可移动磁盘组驱动器。随后的每一代架构都伴随着温彻斯特模式的尺寸缩减——1978年从14英寸缩减到8英寸；1980年从8英寸缩减到5.25英寸；1985年缩减到3.5英寸；1989年缩减到2.5英寸。每次缩减不仅涉及缩小所用组件的尺寸，还涉及重新设计组件在架构中的交互方式。

Many scholars of technological change have observed that new technologies are often introduced into industries by firms entering an industry, rather than by the established firms. In this examination of the extent to which this was true of the disk drive industry, established firms are defined as those that had previously manufactured drives employing earlier technology. Entrant firms are those whose initial product on entry into the industry employed the new technology being analyzed.

许多研究技术变革的学者都指出，新技术通常是由进入行业的公司而不是老牌公司引入行业的。新进入公司是指那些进入该行业的初始产品采用了正在研究的新技术的公司。

In the disk drive industry, the firms that led in the development and use of new component technologies were generally established firms. Entrant firms—whether start-ups or larger integrated concerns—rarely used new component technology in their initial, entry products. No such generalization can be made about points where architectural technological change entered the industry, however. Of the five transitions in architectural technology between 1973 and 1989, established firms led in introducing two, and entrant firms were the leaders in three. The following case histories of two innovations in component technology and two innovations in architectural technology will provide the reader with a sense of what technological innovations in components and architecture were like, how they originated and became diffused through the industry, and how they differed from each other in these respects. These examples were chosen because their histories are representative of a broader set of new technologies.

在磁盘驱动器行业，率先开发和使用新元件技术的公司一般都是老牌公司。新进入的公司，无论是初创公司还是大型综合企业，都很少在其最初的进入产品中使用新的元件技术。然而，对于架构技术变革进入该行业的时间点，却不能作出同样的概括。在1973年至1989 年间的五次架构技术变革中，老牌企业主导了两次变革，而新进入的企业则主导了三次变革。以下是两个组件技术创新和两个架构技术创新的案例史，读者可以从中了解到组件和架构技术创新是什么样的，它们是如何起源并在行业中传播的，以及它们在这些方面有什么不同。之所以选择这些例子，是因为它们的历史可以代表更广泛的新技术。

Leading Innovators in New Read-Write Head and Recording Code Technologies

• IBM began exploring the use of thin-film photolithography to etch an electromagnet onto the surface of a read-write head in 1965, in response to a preliminary technological forecast that the current method for making heads—winding copper threads around a core of ferrite material—would eventually become constrained by manufacturers’ inability to grind the ferrite to finer dimensions. The thin-film approach required a completely different set of engineering competencies, a completely different set of manufacturing equipment, and a very different manufacturing process flow than was required to manufacture ferrite heads competitively. Thin-film was, in terms coined by Michael Tushman and Philip Anderson, a competence-destroying technological innovation

新的读写磁头和记录代码技术领域的创新先行者

·IBM于1966年开始探索使用薄膜光刻技术在读写磁头表面蚀刻电磁铁，根据一项初步技术预测，由于制造商无法将铁氧体磨削到更精细的尺寸，目前的磁头制造方法——在铁氧体材料芯线周围缠绕铜线——最终将受到限制。与制造具有竞争力的铁氧体磁头相比，薄膜方法需要一套完全不同的工程能力、一套完全不同的制造设备和一套完全不同的制造工艺流程。用迈克尔-图什曼（Michael Tushman）和菲利普-安德森（Philip Anderson）的话来说，“薄膜技术是一种‘能力破坏型’技术创新”。

By the time IBM had established proof of the thin-film concept in the early 1970s, other leading integrated disk drive makers such as Control Data and Burroughs had also launched thin-film head development projects. Burroughs announced a model equipped with a thin-film head in 1976 but was never able to manufacture it. In 1979, however, IBM successfully introduced its Model 3340 with thin-film heads—fourteen years, and $300 million dollars, after it had initiated development. Thin film was a difficult, competence-destroying technology—and yet the firms that led in its development and use were the established practitioners of ferrite-head technology. From 1981 to 1986, when over sixty firms entered the rigid disk drive industry, only five of them (all commercial failures) attempted to do so using thin-film heads as a source of performance advantage in their initial products (see Table 6). All other entrant firms—even aggressive, performance-oriented firms such as Maxtor and Conner Peripherals—found it preferable to use ferrite heads in their entry products before tackling thin-film technology in subsequent generations.

20世纪70年代初，当IBM证实薄膜概念时，Control Data和Burroughs等其他领先的磁盘驱动器制造商也启动了薄膜磁头开发项目。巴勒斯公司（Burroughs）在1948年发布了配备薄膜磁头的型号，但一直未能生产。然而，1979年，IBM成功推出了配备薄膜磁头的3340型磁盘驱动器——在此之前，IBM已花费了14年时间和3亿美元。薄膜技术是一项高难度、能力破坏型技术。然而，引领该技术发展和应用的公司均为铁氧体磁头技术的成熟实践者。自1981年到1986年，有六十多家公司进入了硬盘驱动器行业，其中只有五家（均为商业失败者）尝试将使用薄膜磁头作为性能优势的来源（见表6）。所有其他新加入的公司——甚至迈拓（Maxtor）和康纳外设（Conner Peripherals）等积极进取、以性能为导向的公司发现，最好在入门产品中使用铁氧体磁头，再在后续产品中采用薄膜技术。

One explanation for why the leaders in component innovation were the industry’s large, established firms is Chandlerian: these innovations generally were complex, time-consuming, and expensive, and only leading incumbent firms commanded the resources required to undertake and coordinate such development. However, established firms’ leadership in developing and deploying new component technologies extended literally to every component-level innovation—even relatively simple, inexpensive ones—for which a history can be reconstructed. An example of such an inexpensive innovation was the substitution of Run Length Limited (RLL) recording codes for Modified Frequency Modulation (MFM) codes between 1982 and 1988. Development of RLL codes, which enabled a 30 percent improvement in recording density, was a software development project that consumed several hundred thousand to a few million dollars per firm. In spite of this relatively low barrier to development, the established, rather than the entrant, firms still led in this important innovation. Table 6 shows that thirteen firms introduced new models employing RLL technology in 1985. Eleven were established firms, meaning that they had previously offered models based on MFM technology; two were entrants, meaning that their initial products employed RLL codes. Although RLL technology represented a relatively cheap way to increase recording density (and therefore should have been an attractive technology to entrant firms), entrants lagged behind established firms in its introduction.

对于为何组件创新的领导者是该行业的大型老牌企业，有一种钱德勒式的解释：这些创新通常复杂、耗时且昂贵，只有领先的创新型企业才拥有进行和协调此类开发所需的资源。然而，老牌企业在开发和应用新组件技术方面的领先地位实际上扩展到了每一个组件层面的创新——即使是相对简单、廉价的创新，也能重塑其历史。一个这种低成本创新的例子是1982至1988年间，用运行长度受限（RLL）记录代码取代改进的频率调制（MFM）代码。RLL代码的开发使记录密度提高了30%，它是各公司耗资由几十万到几百万美元不等的一个软件开发项目。尽管开发的门槛相对较低，但在这一重要的创新中，老牌企业而非新进入的企业仍然处于领先地位。表6显示，1985年有13家公司推出了采用RLL技术的新机型。其中11家是老牌企业，意味着它们以前曾推出过基于MFM技术的机型；两家是新进入企业，意味着它们的初始产品采用了RLL代码。尽管RLL技术是一种相对廉价的提高记录密度的方法（也因此成为一种有吸引力的方法技术），新进入的公司在引进技术方面仍落后于老牌公司。

The history of other relatively simple but important component level innovations—such as embedded servo systems, zone-specific recording densities, and higher RPM motors—reveals a similar pattern: established firms led in the adoption of new component technology. Entrant firms were the component technology followers. This was true both when the new component technologies had to be developed in-house, as with thin-film heads and RLL codes, and when components, such as spin motors, could be procured from outside vendors. Generalizations that radically new technologies tend to be brought into industries by entrant firms; that established firms will excel primarily at the types of innovation that build on established technological competencies; or that established firms lead in component-level innovation because of their relatively greater ability to countenance greater complexity, risk, and expense seem to be inaccurate and insufficient to explain these patterns of innovation in the disk drive industry.

其他相对简单但重要的组件级创新（如嵌入式伺服系统、特定区域的记录密度和更高转速的电机）的发展历程也揭示了类似的模式：老牌企业率先采用新的组件技术。新进入的公司则是组件技术的追随者。无论是当新的组件技术必须在内部开发（如薄膜磁头和RLL代码）时，还是当组件（如旋转电机）可以从外部供应商采购时，情况都是如此。笼统地说新技术往往是由新进入企业带入行业的；老牌企业主要擅长建立在既有技术能力基础上的创新类型；或者老牌企业在组件级创新方面领先是因为它们相对而言更有能力承受更大的复杂性、风险和费用，这些说法似乎都不准确，也不足以解释磁盘驱动器行业的这些创新模式。

Leaders in Architectural Innovation

• In contrast to the pattern just noted, where established firms led in the development and introduction of new component technology, the pioneers of the 8-, 5.25-, and 3.5-inch generations of architectural technologies were entrant rather than established firms. For example, in 1978 an entrant to the rigid drive industry (Shugart Associates) offered the industry’s first 8-inch drive. By the end of 1979, six firms were offering 8-inch drives; two-thirds of them were entrants. Two years after the first 5.25-inch drive appeared, eight of the ten firms offering 5.25-inch drives were entrants. Entrants likewise dominated the early population of firms offering 3.5-inch drives (see Fig. 3). In general, between half and two-thirds of the established manufacturers of the prior architectural generation never introduced a model in the subsequent architecture, and those that did move into the new technology introduced their new-architecture models an average of two years behind the leading entrant innovators.

架构创新的领导者

·与上述模式不同的是，老牌公司在开发和引进新的组件技术方面处于领先地位，而8英寸、5.25英寸和3.5英寸这几代架构技术的先驱则是新进入者而非老牌公司。例如，1978 年，硬盘驱动器行业的新进入者（Shugart Associates）提供了行业内第一个8英寸驱动器。到1979年底，有六家公司推出了8英寸硬盘，其中三分之二是新进入者。第一个5.25英寸硬盘出现两年后，提供5.25英寸硬盘的十家公司中有八家是新进入者。同样，在早期提供3.5英寸硬盘的公司中，进入者也占据了主导地位（见图3）。一般来说，在上一代架构中，有一半到三分之二的老牌制造商从未推出过后续架构的型号，而那些进入新技术领域的制 造商推出新架构型号的时间平均比领先的创新进入者晚两年。

The reasons why established firms lagged entrants in introducing these new architectural technologies seem unrelated to intrinsic technological difficulty—new architectural designs typically cost less than $2 million to develop, and most employed widely available, proven components. If one adjusts for differences in the types of components employed, the new architectural designs of those established firms that did belatedly introduce them performed just as efficiently as the products of entrant firms introduced in the same year.

老牌公司在引入这些新架构技术方面落后于新进入者的原因似乎与内在技术难度无关——新架构设计的成本通常较低。这些公司的新建筑设计的开发费用超过200万美元，而且大多数都采用了广泛使用的、经过验证的部件。如果对所使用的部件类型的差异进行调整，那么那些姗姗来迟的老牌公司的新架构设计与同年推出的新进入公司的产品一样有效。

The puzzle in these patterns of leadership is intensified by the asymmetry between the risk and expense of innovations in component and architectural technologies and the rewards reaped by the innovators. On the one hand, although new architectural designs were inexpensive and technologically straightforward, their impact was industry-shaking. Each of the leading entrant firms shown in the 1989 and 1992 columns of Table 3 entered the industry with an innovative product architecture employing generally available, proven componentry. On the other hand, despite the risk, time, and expense expended in developing new component technologies, the historical evidence suggests that component-level innovations were essentially defensive in character; they helped innovators remain competitive in their markets but, despite the costs and risks incurred, did not create a sufficient competitive advantage to change the market shares or profitability of innovating firms substantially

组件和架构技术创新的风险和成本与创新者的回报之间的不对称使这些领导模式的谜团更令人费解。一方面，尽管新的架构设计成本低廉、技术简单，但其影响却震动了整个行业。表3中1989年和1992年两栏显示的每一家领先企业都是以创新的产品结构进入该行业的，它们采用的是普遍可用的成熟组件。另一方面，尽管在开发新的组件技术方面历经风险、消耗时间和费用，但历史证据表明，组件层面的创新基本上是防御性的；它们帮助创新者在市场上保持竞争力，但尽管付出了成本和风险，并没有创造出足够的竞争优势以大幅改变创新公司的市场份额或盈利能力。

In contrast to the sluggish movement that the established firms demonstrated in the three architectural innovations just described, they aggressively led the industry in its other two architectural transitions—the substitution of the 14-inch Winchester drive for its 14-inch disk-pack predecessor between 1973 and 1977 and the emergence of the 2.5-inch drive, which began in 1989. After its invention by IBM in 1973, the 14-inch Winchester drive was quickly adopted by the other firms that had been leading suppliers of diskpack drives, such as Control Data, ISS, Burroughs, and EMM. Seven of the eight firms that had introduced 14-inch Winchester architecture drives by 1977 were established producers of the prior architectural generation.

与老牌企业在上述三项架构创新的低迷相比，它们积极引领行业的另外两项架构转型——1973年至1977年间用14英寸Winchester硬盘取代其14英寸磁盘组前身，以及1989年开始出现的2.5英寸硬盘。1973年IBM发明14英寸Winchester硬盘后，Control Data、ISS、Burroughs和EMM等曾是磁盘组硬盘主要供应商的其他公司很快也开始采用该硬盘。在1977年之前推出14英寸温彻斯特架构硬盘的八家公司中，有七家是上一代架构的硬盘老牌生产商。

The first firm to introduce the 2.5-inch drive in 1989 was an entrant, Prairietek—a spin-off of Miniscribe, the second-largest maker of 5.25-inch drives. However, within a year the two largest makers of 3.5-inch drives, Conner Peripherals and Quantum, had weighed in with their own 2.5-inch models, and by 1991 they had captured over 95 percent of the 2.5-inch market. Prairietek was bankrupt by 1992.

1989年，第一家推出2.5英寸硬盘的公司是Prairietek，它是5.25英寸硬盘第二大制造商Miniscribe的分拆公司。然而不到一年，两家最大的3.5英寸硬盘制造商Conner Peripherals和Quantum也推出了自己的2.5英寸型号，到1991年，它们已占据了2.5英寸市场95%以上的份额。Prairietek于1992年破产。

Sustaining vs. Disruptive Technological Change

可持续性与颠覆性技术变革的决斗

The pattern of technological leadership in these two drive architectures was remarkably similar to the one described earlier for thin-film heads: established firms aggressively led the industry, and there were fatal consequences for the entrants that attempted to lead.

Why were these so similar to the leadership patterns in component technology? And what was it about the 8-, 5.25-, and 3.5-inch architectures that facilitated the entrant firms’ leadership? The hypothesis raised by this research is that the innovations in component technology and in the 14-inch and 2.5-inch Winchester architectures all had a similar impact on the customers of the leading established firms. Each of these technologies sustained the trajectory of product performance improvement that these firms’ customers demanded and had come to expect. In contrast, the 8-, 5.25-, and 3.5-inch architectures disrupted the trajectory of performance improvement in established markets and hence had initial appeal only in new, emerging market segments.

这两种驱动器架构的技术领先模式与前面描述的薄膜磁头模式极为相似：老牌公司积极引领行业发展，而试图领先的新进入者则面临致命的后果。为什么这与组件技术的领导模式如此相似？8英寸、5.25英寸和3.5英寸架构又是什么促成了新进入公司的领导地位？本研究提出的假设是，组件技术以及14英寸和2.5英寸温彻斯特架构的创新都对领先老牌企业的客户产生了类似的影响。每项技术都保持了这些公司的客户所要求和期望的产品性能改进轨迹。与此相反，8英寸、5.25英寸和 3.5英寸架构打破了既有市场的性能改进轨迹，因此最初只对新兴细分市场具有吸引力。

These differences in the way the technologies affected established firms’ customers are displayed in Figures 4 and 5. Figure 4a charts the trend in recording density achieved in drives using conventional ferrite-head/oxide-disk technology compared to that achieved in drives employing advanced thin-film heads and disks. Figure 4b contrasts trends in density achieved with disk-pack drives with that of the 14-inch Winchester architecture. In both charts, the solid dots and the curves fitted through them chart the trajectory of improvement achieved in the earlier technology. Each seems to have followed the sort of S-curve pattern that Richard Foster shows to be typical of technology life cycles. Thin-film component technology and Winchester architectural technology each enabled the firms innovating with those technologies to sustain the rate of performance improvement that historically had become established. In this respect, these two charts are archetypical of all other innovations in componentry and of the 2.5-inch Winchester architecture.

图4和图5显示了这些技术对老牌企业客户影响方式的差异。图4a显示了采用传统铁氧体磁头/氧化物磁盘技术的硬盘与采用先进薄膜磁头和磁盘的硬盘在记录密度方面的趋势对比。图4b对比了磁盘组硬盘与14英寸Winchester架构硬盘的密度趋势。在这两张图中，实心点和穿过实心点的曲线描绘了早期技术的改进轨迹。薄膜组件技术和温彻斯特架构技术分别使采用这些技术进行创新的公司保持了历史上已经形成的性能改进速度。从这个角度来看，这两张图表是组件领域所有其他创新以及2.5英寸Winchester架构的典型代表。

In contrast, Figure 5 highlights the disruptive nature of the 8-, 5.25-, and 3.5-inch architectures by mapping the average capacity of all drives designed with each architecture compared to the capacity of drives in other architectures. In this graph, the dotted lines chart the average capacity of all drives within each architecture introduced in each year beginning in 1976. The top dotted line shows that the capacity of the average 14-inch disk-pack drive was about 250 MB in 1976. As the capacity trajectory of the disk-pack architecture began to level off, the Winchester architecture emerged to sustain the historical trajectory of 22 percent annual growth. In 1978, the first 8-inch Winchester drives were introduced. Physically much smaller, the average 8-inch drive packed only 20 MB of capacity—compared to an average of 350 MB for the 14-inch drives introduced that year. The 8-inch technology, in other words, did not sustain the established capacity trajectory of 14-inch drives—it created a new trajectory. Once the 20 MB starting point was established, however, the makers of 20 MB drives subsequently were able to boost the capacity of their drives along a very steep 50 percent annual improvement trajectory.

相比之下，图5通过绘制采用各架构设计的所有硬盘的平均容量与其他架构硬盘容量的对比图，突出显示了8英寸、5.25英寸和3.5英寸架构的颠覆性。在此图中，虚线表示从1976 年开始每年推出的每种架构中所有硬盘的平均容量。顶部虚线显示，1976年14英寸磁盘组硬盘的平均容量约为250MB。随着磁盘阵列架构的容量轨迹开始趋于平稳，温彻斯特架构的出现维持了每年22%的增长轨迹。1978年，首款8英寸温彻斯特硬盘问世。8英寸硬盘的物理尺寸要小得多，平均容量只有20 MB，而当年推出的14英寸硬盘平均容量为350 MB。换句话说，8英寸技术并没有延续14英寸硬盘的既定容量轨迹，而是创造了一个新的轨迹。然而，20 MB硬盘的起点一旦确立，20 MB硬盘的制造商就能沿着每年提高50%的陡峭轨迹提高硬盘容量。

Figure 5 shows that, when the 5.25-inch architecture emerged in 1980, it had a similar disruptive impact on the established trajectories of capacity improvement. The new 5.25-inch products offered only 5-10 MB of capacity that year, compared to 50 MB in the average 8-inch drive and 700 MB in the average 14-inch Winchester product. Again, once they had begun production, the 5.25-inch drive makers were able to push the capacities of their products along a trajectory roughly parallel to that of the 8-inch products. The 3.5-inch architecture, introduced in 1985, had a similar disruptive impact vis-a-vis the earlier products, and 3.5-inch drives also improved along a trajectory parallel to that of earlier architectures.

图5显示，当5.25英寸架构在1980年出现时，它对容量提升的既定轨迹也产生了类似的颠覆性影响。当年，新的5.25英寸产品仅提供5-10 MB的容量，而平均8英寸硬盘的容量为50MB，平均14英寸（Vinchester）产品的容量为700 MB。同样，一旦开始生产，5.25 英寸硬盘制造商就能沿着与8英寸产品大致平行的轨迹推动其产品的容量。1985年推出的3.5英寸架构与早期产品相比也产生了类似的颠覆性影响，3.5英寸硬盘也沿着与早期架构平行的轨迹不断改进。

Although each packed less total capacity than predecessor products, these disruptive architectures did not necessarily represent a sequence of successively inferior technological approaches—they simply offered a very different package of attributes than were offered in the prior architectural generation. The 5.25-inch drive, for example, was inferior to 8-inch products in capacity and speed. But it was smaller and less expensive; it could fit physically and economically in a desk-top computer. Similarly, the capacity and speed of the 3.5-inch drive were inferior to those of 5.25-inch products when it was first introduced. But it had other redeeming attributes—it was small and rugged enough to be used in the early portable and laptop computers. Hence, although each of these new architectures could be deemed inferior according to the standards used to assess the performance of predecessor technologies, the new drives had appealing properties for other purposes. Trajectory-disrupting architectures therefore tended to be used initially in new, emerging market segments rather than in the large, mainstream markets served by the leading disk drive manufacturers. In contrast, the trajectory-sustaining technologies were first used within the mainstream markets.

虽然每台产品的总容量都低于前代，这些颠覆性的架构并不一定代表着一连串相继出现的低劣技术方法，它们只是提供了与上一代架构截然不同的一揽子属性。例如，5.25英寸硬盘在容量和速度上都不如8英寸产品。但它的体积更小、价格更低；在物理和经济上都适合台式电脑。同样，3.5英寸硬盘刚推出时，其容量和速度也不如5.25英寸产品。但它还有其他可取之处——体积小、坚固耐用，可用于早期的便携式电脑和笔记本电脑。因此，尽管根据评估前代技术性能的标准，这些新架构中的每一种都可能被认为是劣质的，但新硬盘在其他用途上却具有吸引人的特性。因此，轨迹破坏型架构一开始往往用于新兴市场，而不是主要磁盘驱动器制造商所服务的大型主流市场。与此相反，轨迹维持型技术则首先用于主流市场。

The forgoing examination of the history of leadership in technological innovation in the disk drive industry has disclosed that established firms consistently led the industry in developing and adopting new technologies—whether in componentry or in architecture—that reinforced or sustained the trajectory of performance improvement that their customers expected. When new technologies disrupted established trajectories and redefined the metrics by which performance was measured, entrant firms were the leading innovators.

对磁盘驱动器行业技术创新领导力的历史回顾表明，老牌企业在开发和采用新技术方面始终处于行业领先地位，无论是组件还是架构，这些新技术都强化或保持了客户所期望的性能提升轨迹。当新技术打破了既定的发展轨迹并重新定义了衡量性能的标准时，新进入的公司则成为创新的主导者。

Market Demand vs. Technology Supply as Drivers of Change in Industry Leadership

市场需求与技术供给是行业领导力变革的驱动力

We have just seen that drives with disruptive new architectural technologies were generally deployed in new market segments, where computing products needed the new package of attributes that these drives offered. A fit between a particular product architecture and the product characteristics demanded in a specific market segment might have resulted in a series of "niche" markets, each with its own relatively unique definition of product performance. In the disk drive industry, however, such a market structure never emerged—because the rate of increase in performance that disk drive manufacturers were able to provide within each new product architecture was substantially greater than the rate of performance improvement demanded by customers.

我们刚才已经看到，具有颠覆性新架构技术的硬盘通常部署在新的细分市场中，这类市场中的计算产品需要这些硬盘所提供的新属性包。特定产品架构与特定细分市场所需的产品特性之间的契合可能会产生一系列“利基”市场，其中每个市场都有自己相对独特的产品性能定义。然而，在磁盘驱动器行业，这样的市场结构却从未出现过——因为磁盘驱动器制造商在每个新产品架构中能够提供的性能提升速度大大超过了客户要求的性能提升速度。

These differences in the rates of performance improvement are contrasted in Figure 5. The dotted lines emanating from points A, B, C, and D measure trends in the average capacity that disk drive manufacturers were able to provide within each successive disk drive architecture. These steep trends in performance improvement within each architecture were driven by the sorts of trajectory-sustaining improvements in componentry described earlier. The disruptions in these trajectories—the movements from point A to points B, C, and D—were the result of changes in architectural technology. In light of the preceding discussion on technology leadership, we can now assert that the firms that led the industry along each dotted-line technology trajectory were established firms, and that the leaders in jumping to new points of departure (B, C, and D) were entrants.

图5对比了性能提升率方面的差异。从A、B、C和D点发出的虚线测量了磁盘驱动器制造商在每个连续磁盘驱动器架构中能够提供的平均容量趋势。每种架构的性能提升都呈现出这种急剧上升的趋势，这是由前面提到的各种组件的轨迹持续型改进所推动的。这些轨迹的中断——从A点到B、C和D点的移动——是架构技术变化的结果。根据前面关于技术领先地位的讨论，我们现在可以断言，沿着每条虚线技术轨迹引领行业发展的公司都是老牌公司，而跳到新起点（B、C和D）的领先公司都是新进入者。

The solid lines in Figure 5 map the demand trajectories for hard disk capacity within the major successive market segments of the computer industry: mainframes, minicomputers, desktop personal computers (PCs), portable and laptop PCs, and notebook PCs. These lines show, for example, that in 1976 the median-priced mainframe computer was equipped with 160 MB of hard disk capacity. The hard disk storage capacity supplied with the median-priced mainframe increased about 17 percent a year through the period studied, so that by 1990 the typical mainframe was equipped with 1,300 MB of hard disk capacity. Points B, C, D, and E mark the years in which 8-, 5.25-, 3.5-, and 2.5-inch drives were first used in the minicomputer, desktop personal computer, portable, and notebook computer market segments, respectively. The hard disk capacity demanded per computer (charted by the solid lines starting at each point) increased at annual rates of approximately 30 percent in each of these segments.

图5中的实线描绘了计算机行业主要连续细分市场对硬盘容量的需求轨迹：大型计算机、 微型计算机、台式个人计算机 (PC)、便携式个人计算机以及笔记本计算机。例如，这些数据线显示，1976年价格中位数的主机配备了160 MB的硬盘容量。在研究期间，价格中位数主机所配备的硬盘存储容量每年增长约17%，到1990年，典型的主机配备了1300 MB的硬盘容量。B、C、D和E点分别是8英寸、5.25英寸、3.5英寸和2.5英寸硬盘首次用于微型计算机、台式个人计算机、便携式计算机和笔记本计算机市场的年份。在这些细分市场中，每台计算机所需的硬盘容量（以每一点为起点的实线表示）以每年约30%的速度增长。

The mismatch between what the markets demanded and what the technology could provide enabled technologies that initially were not performance-competitive in established markets eventually to intersect with the demand in those markets. For example, by 1986 the average 8-inch drive packed the capacity required in the medianpriced mainframe computer. The average 5.25-inch drive, which initially could satisfy only the performance demanded in desktop personal computers, by 1986 packed the capacity demanded by the typical minicomputer user. By 1989 the 5.25-inch architecture was invading the lower end of the mainframe market.

市场的需求与技术所能提供的服务之间的不匹配，使得最初在成熟市场上没有性能竞争力的技术最终与这些市场的需求产生了交集。例如，到1986年，8英寸硬盘的平均容量已达到中位价位大型计算机的要求。平均5.25英寸硬盘最初只能满足台式个人电脑对性能的要求，但到1986年，其容量已达到典型微型计算机用户的要求。到1989年，5.25英寸架构的硬盘开始进入低端大型机市场。

To understand better what happened at the points where these new technologies emerged and where technology and market trajectories intersected, I interviewed over sixty executives who played key roles in the marketing, engineering, and manufacturing functions of the industry’s leading firms at these points of intersection. Data from these interviews were used to recreate as clearly as possible the processes that led to the firms’ decisions to invest, or to delay investing, in key component and architectural technologies. Although each firm’s experience differed in detail, there was remarkable similarity across the firms in the forces that influenced their technology investment decisions and in the outcomes that each experienced at the emergence of the 8-, 5.25-, and 3.5-inch architectures. These findings are generalized in the following sequence of events.

为了更好地了解这些新技术的出现点以及技术和市场轨迹的交汇点发生了什么，我采访了六十多名在这些交汇点上的行业领先企业的营销、工程和制造部门担任重要职务的管理人员。这些访谈数据被用来尽可能清晰地再现公司决定投资或推迟投资关键组件和架构技术的过程。尽管每家公司的经历在细节上有所不同，但在影响其技术投资决策的力量方面，以及在8英寸、5.25英寸和3.5英寸架构出现时每家公司所经历的结果方面，却有着显著的相似性。这些发现可归纳为以下事件顺序。

First, engineers in established firms conceived of new architectural concepts enabling the disruptive-architecture drives; they fabricated prototype samples and tested the market appeal of the products, in the form of drawings or prototypes, with mid-level marketing counterparts in their firms. The marketing personnel in turn showed the prototypes or drawings to key customers, who showed little interest in the products, because the new drives packed less capacity and typically had slower access times than the larger architecture drives the customers currently sourced for design into their existing computer systems. In response to customers’ feedback, the established disk drive manufacturers scaled back the resources committed to the new-architecture projects or canceled the efforts altogether. They then intensified their efforts to fill the need their customers had clearly articulated—in the case of mainframe computer manufacturers, greater capacity and faster access times within the 14-inch architecture. To accomplish this, the established drive makers intensified their focus on improving and employing advanced component technologies.

首先，老牌企业的工程师们构思出新的架构概念，从而推动了颠覆性架构的发展；他们制作出原型样品，并以图纸或原型的形式，与企业的中层营销人员一起测试产品的市场吸引力。营销人员又向主要客户展示了原型或图纸，但客户对产品兴趣不大，因为与客户目前采购的用于现有计算机系统设计的大型架构硬盘相比，新硬盘的容量较小，存取时间通常较慢。针对客户的反馈，老牌磁盘驱动器制造商缩减了投入新架构项目的资源，或完全取消了这些项目。然后，他们加紧努力以满足客户明确提出的需求。就大型计算机制造商而言，他们希望在14英寸架构内获得更大的容量和更快的访问速度。为了实现这一目标，老牌硬盘制造商加强了对改进和采用先进组件技术的关注。

In response to the loss or scaling back of their program, members of the engineering team that developed the smaller architecture often defected in frustration, launching one or more start-ups to manufacture drives based on the new technology. The start-up companies also were unable to interest large computer makers in the new-architecture drives. They therefore had to find new market applications for their drives. There was substantial uncertainty in this search for customers—what the markets might be and how large they might become were unknown. The applications that materialized during these search processes were the minicomputer, the desktop personal computer (and later the engineering workstation), and the portable computer industries. Although these markets are easily understood today as natural applications for hard disk drives, this was not at all clear when the new markets were first emerging.

为了应对项目的失败或缩减，开发小型架构的工程团队成员往往会沮丧地叛逃，成立一家或多家新公司来生产基于新技术的硬盘。新成立的公司也无法让大型计算机制造商对新架构硬盘感兴趣。因此，它们不得不为自己的硬盘寻找新的市场应用。在寻找客户的过程中，存在着很大的不确定性——市场可能是怎样的，规模有多大，这些都是未知数。在这些寻找过程中，微型计算机、台式个人计算机（以及后来的工程工作站）和便携式计算机行业的应用成为现实。尽管这些市场如今很容易被理解为硬盘驱动器的自然应用，但在新市场刚刚出现时，这一点并不明显。

When the start-ups had established a beach-head business in these new applications, they found that, by incorporating advances in new component technologies, they could increase the capacity of their drives at a faster rate than was required to satisfy customers in their home markets. They therefore fixed their strategic sights on the established computer markets immediately above them. When the capacity of the new-architecture drives had increased to the levels required in these higher market segments, computer manufacturers in those markets began switching to the “new” (now established) technology. At this point they generally found that the drives not only provided the capacity they needed, but that their smaller size and relative architectural simplicity also made them less costly and more reliable.

当初创公司在这些新应用中建立了一个业务据点后，他们发现，通过整合新组件技术的进步，他们能够以比满足原市场需求更快的速度提高驱动器的容量。因此，他们将战略目光锁定在紧随其后的成熟计算机市场。当新架构硬盘的容量提高到这些高端细分市场所需的水平时，这些市场上的计算机制造商便开始转向“新”（现已确立的）技术。此时，他们普遍发现，硬盘不仅能提供所需的容量，而且体积更小、结构相对简单，因此成本更低，可靠性更高。

When the smaller drives began to gain share in these higher markets, the drive makers that had established themselves by supplying the larger architectures took the new technology “off the shelf” and introduced their own versions as a defensive response to a competitive attack from below. As a result, these models generally can nibalized established disk drive manufacturers sales of largerarchitecture products. Yet, because they launched their new architecture models so late, the established drive makers rarely were able to build significant incremental business in the market segment that had recently emerged. Although a few established drive makers were able to defend their prior market positions by launching the new architecture, most found that the entrant firms had developed insurmountable volume-based manufacturing cost advantages through their positions in the new market segment, and the older firms ceded the established markets to the invaders.

当较小的硬盘开始在这些高端市场上获得份额时，那些通过供应较大架构产品而建立起自己地位的硬盘制造商就会将新技术“下架”，并推出自己的版本，作为对来自下面的竞争性攻击的防御性回应。因此，这些型号通常会削弱老牌磁盘驱动器制造商对大型架构产品的销售。然而，由于推出新架构型号的时间太迟，老牌硬盘制造商很少能在新近出现的细分市场中建立重要的增量业务。尽管少数老牌硬盘制造商能够通过推出新架构产品来捍卫其先前的市场地位，但大多数厂商发现，新进入的公司已经形成了难以逾越的以产量为基础的制造成本优势，而老牌企业则将既有市场拱手让给入侵者。

To illustrate more deeply the process through which the established firms tended to deal with disruptive architectural transitions, the story of one industry leader, Seagate Technology, is documented in detail in the next section.

为了更深入地说明老牌公司如何应对颠覆性的架构转型，下一节将详细记录行业领导者希捷科技的故事。

Seagate and the Transition from 5.25- to 3.5-Inch Drives

希捷与5.25英寸硬盘到3.5英寸硬盘的变革

The 3.5-inch drive was first developed in 1984 by Rodime, a Burroughs spin-off. Sales of this architecture were not significant, however, until Conner Peripherals, a Seagate-Miniscribe spin-off, started operations in 1986. Conner’s small, lightweight drive architecture was much more rugged than its 5.25-inch ancestors; it handled functions electronically that had previously been managed with mechanical parts and used microcode to replace functions that previously had been addressed electronically. Over 90 percent of Conner’s record first-year revenues of $113 million came from Compaq Computer, which had funded most of Conner’s start-up with a $16 million investment.

3.5英寸硬盘于1984年由 Burroughs分公司Rodime首次开发。然而，直到1986年Seagate-Miniscribe公司分拆出来的Conner Peripherals公司开始运营，这种结构的销售量才有所起色。康纳公司的小巧轻便硬盘架构比其5.25英寸的前代更加坚固耐用；它以电子方式处理以前由机械部件管理的功能，并使用微码取代以前以电子方式处理的功能。康纳公司创纪录的1.13亿美元首年收入中，90%以上来自康柏电脑公司，该公司为康纳公司的启动提供了1600万美元的大部分资金。

The 3.5-inch Conner drives were used primarily in a new application—portable and laptop machines, in addition to a few “small footprint” desktop models—where customers were willing to accept lower capacities and higher costs per megabyte in order to get the smaller size and weight, greater ruggedness, and lower power consumption that 3.5-inch drives offered. Like most firms that entered the disk drive market on the basis of a new product architecture, Conner attracted little notice from the established drive makers, because it sold products on the basis of performance attributes that established firms did not appreciate to customers that the established firms did not know. Conner’s relationship with Compaq, which was itself a company new to the computer business and whose own volume trajectory was meteoric, enabled Conner to build a substantial volume base without attracting serious competitive attention.

3.5英寸康纳硬盘主要用于一种新的应用——便携式和笔记本电脑，以及一些“小尺寸”台式机——在这些应用中，客户愿意接受较低的容量和较高的每兆字节成本，以获得3.5英寸硬盘所提供的更小的尺寸和重量、更高的坚固性和更低的功耗。与大多数以新产品结构进入磁盘驱动器市场的公司一样，康纳公司几乎没有引起老牌驱动器制造商的注意，因为它以老牌公司不欣赏的性能属性为基础，向老牌公司不了解的客户销售产品。康纳公司与康柏公司的合作关系，使康纳公司建立了一个庞大的销量基础，而没有引起竞争对手的激烈竞争性关注。

The leading 5.25-inch drivemaker, Seagate Technology, clearly foresaw the 3.5-inch architecture’s advent. By early 1985, less than one year after the first 3.5-inch drive was introduced by Rodime and two years before Conner Peripherals started shipping its product, Seagate engineers had developed working 3.5-inch prototype drives. Their development had actually been instigated at the request of Seagate’s largest customer, IBM, whose product planners were considering replacing the 5.25-inch drives in the PC-AT desktop computer with the smaller 3.5-inch drive. When IBM evaluated the Seagate 3.5-inch prototypes and understood that their capacities were limited to 10 and 20 MB models, however, the product planners decided to go with next-generation 5.25-inch drives instead. They sensed that the AT’s customers wanted 40 and 60 MB of disk storage in the next AT models to be released rather than the less tangible benefits of a physically smaller drive housed within the large AT box on their desktops.

领先的5.25英寸硬盘制造商希捷科技公司清楚地预见到了3.5英寸结构的出现。1985年初，Rodime和Seagate推出首款3.5英寸硬盘还不到一年。在Conner Peripherals开始销售其产品的两年前，希捷的工程师就已经开发出了可正常工作的3.5英寸原型硬盘。这些硬盘的开发实际上是应希捷最大客户IBM的要求而启动的，当时IBM的产品规划人员正在考虑用较小的3.5英寸硬盘取代PC-AT台式机中的5.25英寸硬盘。然而，当IBM评估了希捷3.5英寸原型机并了解到其容量仅限于10 MB和20 MB型号时，产品规划人员决定改用新一代5.25英寸硬盘。他们意识到AT机的客户在下一代AT机型中需要的是40和60MB的磁盘存储空间，而不是那些不那么切实的好处，比如在他们桌面上的大AT机箱内安装一个体积更小的硬盘。

Having been abandoned at the altar by IBM, Seagate marketers subsequently worked to sell the 3.5-inch product to other customers. But the customers to whom the 3.5-inch drives were shown were also manufacturers of full-sized desktop computer systems. Like IBM, they were looking for capacities of 40 and 60 megabytes for their next generation machines and showed little interest in the smaller drive. In response to these lukewarm reviews from customers, Seagate’s program manager lowered his 3.5-inch sales estimates, and the firm’s executives subsequently shelved the 3.5-inch program. Their reasoning was that the markets for 5.25-inch products were larger and that an engineering effort on new 5.25-inch products would generate greater revenues for the company than would efforts targeted at new 3.5-inch products. “We needed a new model,” recalled a Seagate program manager, “which could become the next ST412 [a very successful product nearing the end of its life cycle], which at the time was generating $300 million per year in revenues. The 3.5-inch product just didn't fit the bill. The market was just too small.” Seagate executives were also convinced that, because of manufacturability challenges inherent in the 3.5-inch drive and the relatively smaller size of its market, the new architecture would never become competitive, on a cost-per-megabyte basis, with 5.25- inch products.

被IBM抛弃后，希捷的营销人员随后努力向其他客户销售3.5英寸产品。但是，3.5英寸硬盘的客户也是全尺寸台式计算机系统的制造商。与IBM一样，他们也在为下一代机器寻找40和60兆字节的容量，对较小的硬盘兴趣不大。面对客户的冷淡评价，希捷的项目经理降低了3.5英寸硬盘的销售预期，公司的高管随后搁置了3.5英寸硬盘项目。他们的理由是，5.25英寸产品的市场更大，与针对新的3.5英寸产品的工程设计相比，针对新的5.25英寸产品的工程设计将为公司带来更大的收益。希捷的一位项目经理回忆说：“我们需要一个新的型号，它可以成为下一个ST4I2 [一个非常成功的产品，其生命周期即将结束]，当时它的年收入为3亿美元。3.5英寸产品并不符合要求。”希捷高管还坚信，由于3.5英寸硬盘固有的可制造性挑战及其相对较小的市场规模，按每兆字节成本计算，新架构永远无法与 5.2英寸产品竞争。

From 1984 to 1989, when the 3.5-inch architecture was becoming firmly established in portable and laptop applications, Seagate had in no way lost its ability to innovate. It was highly responsive to its own customers. The capacity of its drives increased at about 30 percent a year—a perfect match with the market demand charted in Figure 5 and a testament to the firm’s focus on the desktop computing market. Seagate also introduced new models of 5.25-inch drives at an accelerated rate. During this period Seagate announced new products that employed most of the available new component technologies—including thin-film disks, voice coil actuators, RLL codes, and embedded, SCSI interfaces.

从1984年到1989年，当3.5英寸架构在便携式和笔记本电脑应用中逐渐站稳脚跟时，希捷丝毫没有丧失其创新能力。它对自己的客户反应非常迅速。希捷硬盘的容量以每年约 30%的速度增长——与图5所示的市场需求完全吻合，同时也证明了希捷对台式机市场的专注。希捷还加速推出了5.25英寸的新机型。在此期间，希捷发布的新产品采用了大部分现有的新组件技术，包括薄膜磁盘、音圈致动器、RLL代码和嵌入式 SCSI接口。

By 1987-88, 3.5-inch drives began to pack the capacity required in the desktop market (shown in Figure 5 where the dotted 3.5-inch capacity technology trajectory begins to intersect with the demandfor-capacity trajectory in desktop computing). At this point, desktop computer makers discovered that, in addition to providing the capacity they needed, the smaller architecture offered other advantages on the desktop. For example, the small drives allowed them to shrink the footprint and profile of their products and to eliminate a noisy cooling fan. And less mass in the 3.5-inch drive meant less vibration and less inertia, enabling drive makers to position the head more accurately over more densely spaced concentric recording tracks on the disk. These engineering implications had not been apparent when the 3.5-inch architecture was first presented to manufacturers.

到1987至1988年，3.5英寸硬盘开始满足台式机市场的容量需求（如图5所示，3.5英寸容量技术轨迹虚线开始与台式机计算的容量需求轨迹相交）。此时，台式机制造商发现，除了提供所需的容量外，较小的架构还为台式机提供了其他优势。例如，小型硬盘使他们能够缩小产品的占地面积和外形，并省去了嘈杂的冷却风扇。3.5英寸硬盘的质量更小，意味着振动更小，惯性更小，使硬盘制造商能够在磁盘上更密集的同心记录轨道上更精确地定位磁头。当3.5英寸架构首次提交给制造商时，这些工程影响并不明显。

To defend its position in the desktop market, Seagate finally began shipping 3.5-inch drives in early 1988—three years after it had designed the drives and after nearly $750 million in 3.5-inch products had been shipped cumulatively in the industry. As of 1991, however, almost none of Seagate’s 3.5-inch products had been sold to manufacturers of portable, laptop, or notebook computers. Its primary customers for 3.5-inch drives were its previous customers for 5.25-inch drives, the desktop computer manufacturers; indeed, many of Seagate’s 3.5-inch drives continued to be shipped with frames permitting them to be mounted in XT- and AT-class computers designed to accommodate the larger drives.

为了捍卫自己在台式机市场上的地位，希捷终于在1988年初开始出货3.5英寸硬盘，这距离希捷设计出这种硬盘已经过去了三年，而希捷的3.5英寸产品在业内的累计出货量已经接近7.5亿美元。然而，截至1991年，希捷的3.5英寸产品几乎没有卖给便携式电脑、笔记本电脑或笔记本电脑制造商。希捷3.5英寸硬盘的主要客户是其之前5.25英寸硬盘的客户，即台式计算机制造商；事实上，希捷的许多3.5英寸硬盘在发货时仍带有框架，可安装在XT和AT级计算机中。设计用于容纳较大硬盘的驱动器。

Seagate’s response to the development of the 3.5-inch drive architecture was not atypical; by 1988, only 35 percent of the drive manufacturers that had established themselves making 5.25-inch products for the desktop PC market had introduced 3.5-inch drives. What led the incumbent manufacturers to resist or ignore the new technology?

希捷对3.5英寸硬盘架构开发的反应并不典型；到1988年，在为台式PC市场生产5.25 英寸产品的硬盘制造商中，只有35%推出了3.5英寸硬盘。是什么导致现有制造商抵制或忽视新技术？

The barrier to the development of competitive new-architecture products does not appear to have been engineering-based. In each case, established manufacturers that did introduce the new architecture, though late, weighed in with products whose performance was actually superior to that of the entrants, and they subsequently improved their new architecture products at a faster rate than the entrants did.

开发具有竞争力的新架构产品的障碍似乎并不在于工程技术。在每种情况下，虽然引入新架构的老牌制造商推出新架构产品的时间较晚，但他们的产品在性能上实际上优于新进入者，而且它们随后改进新架构产品的速度比新进入者更快。

The fear of cannibalizing sales of existing products is often cited as a reason why established firms delay the introduction of new technologies. As the Seagate-Conner experience illustrates, however, when innovative technologies are initially deployed in new-market applications, the introduction of new technology may not be inherently cannibalistic. When established firms wait until a new technology has become commercially mature in its new applications, however, and launch their own version of the technology only in response to an attack on their home markets, the fear of cannibalization can become a self-fulfilling prophecy.

担心蚕食现有产品的销售额往往是老牌企业推迟引进新技术的原因。然而，正如 Seagate-Conner公司的经验所表明的那样，当创新技术最初应用于新市场时，新技术的引进可能并不必然造成蚕食。然而，如果老牌企业等到新技术在新应用领域的商业化成熟之后，才推出自己的技术版本，以应对其本土市场的冲击，这种对蚕食的恐惧可能会成为自我实现的预言。

The primary problem of the established drive manufacturers was that they were held captive by their customers, who seemed, as the Seagate history demonstrates, to have been as oblivious as the drive makers to the potential benefits and possibilities of the new architectures. In the industry’s two trajectory-sustaining architectural transitions—from disk-pack drives to 14-inch Winchester drives and from 3.5- to 2.5-inch drives—the industry leader in the prior architectural generation was able to maintain its leadership in the subsequent generation as well. In both cases, their customers led them across the architectural transition. For example, the same companies that had leading market shares in portable and laptop computers—Toshiba, Sharp, and Zenith—also became the leading manufacturers of notebook computers. The same performance attributes that were valued in laptop machines, such as ruggedness, low power consumption, and capacity per unit of weight and volume, were critical in notebook computers as well. Since the 2.5-inch drive addressed these needs more effectively than did the 3.5-inch architecture, Conner Peripherals was able to follow its customers smoothly across this architectural transition. Doing so required no change in strategy.

老牌硬盘制造商的主要问题是，他们被客户所束缚，正如希捷的历史所证明的那样，客户似乎和硬盘制造商一样，对新架构的潜在优势和可能性视而不见。在行业的两次轨迹持续型架构转型中——从磁盘组硬盘到14英寸硬盘，以及从3.5英寸硬盘到2.5英寸硬盘——上一代架构的行业领导者在下一代架构中也能保持领先地位。在这两种情况下，他们的客户都引领他们完成了架构转型。例如在便携式电脑和笔记本电脑市场占有率领先的东芝、夏普和真力时也成为笔记本电脑的领先制造商。笔记本电脑同样需要坚固耐用、低功耗、单位重量和体积的容量等笔记本电脑所看重的性能。 由于2.5英寸硬盘比3.5英寸架构更有效地满足了这些需求，Conner Perirherals得以跟随其客户顺利地跨越了这一架构转型鸿沟。这样做并不需要改变战略。

A Deeper View of the Success and Failure of Large, Diversified Firms

深入了解大型多元化企业的成功与失败

The first section of this article showed that large-scale, diversified concerns were the early dominant manufacturers in the original equipment market. Having established the alignment in the disk drive industry between disruptive innovations in architectural technologies and the emergence of new market segments on the one hand, and the appeal of sustaining technological innovations to established market segments on the other, it is now possible to examine at a deeper level the overwhelming defeat by the start-ups of the large, integrated firms that once dominated the U.S. disk drive industry. The decline of these firms does not seem to have been caused by technological conservatism, short time horizons, or an unwillingness to invest in new technologies, as their leadership in a range of relatively risky, difficult, competence-destroying component innovations shows. Nor does their failure seem to have resulted from ignorance of customers’ needs or by a sluggish response to them. It was their very attentiveness to their customers’ needs, in fact, that drove their leadership in component and trajectory-sustaining architectural technologies. Rather, it seems that these firms failed because they listened too attentively to their established customers and ignored new product architectures whose initial appeal was in remote markets. Eventually, because of the mismatch in the trajectories of technological improvement and market demand, the established firms fell victim to attacks by start-ups and the new architectural technologies the entrants employed.

本文第一部分指出，大型多元化企业是原始设备市场的早期主导制造商。在磁盘驱动器行业中，一方面是架构技术的颠覆性创新和新细分市场的出现，另一方面是持续性技术创新对既有细分市场的吸引力，在这两者之间建立了一致性之后，我们现在可以从更深的层面来审视曾一度主导美国磁盘驱动器行业的大型综合企业被初创企业压倒性击败的情况。这些公司的衰落似乎并不是由技术保守主义、时间跨度短或“不求上进”造成的。不愿意对新技术进行投资，因为他们在一个从一系列风险相对较高、难度相对较高、能力破坏型的组件创新中可以看出，这些创新的失败似乎并不是因为不了解客户的需求，也不是因为对客户需求的反应迟钝。事实上，正是他们对客户需求的关注，推动了他们在组件和轨迹维持型架构技术方面的领先地位。相反，这些公司之所以失败，似乎是因为它们过于关注老客户的需求，而忽视了最初在偏远市场具有吸引力的新产品架构。最终，由于技术改进的轨迹与市场需求不匹配，这些老牌公司成为了初创公司和新进入者所采用的新架构技术攻击的牺牲品。

Table 2 showed that in 1976 there were seventeen active firms in the OE market: four related-technology, ten related-market, and three vertically integrated. Only one firm from each of these groups of original participants was still producing disk drives in 1989. And of the ten related-market firms that entered the industry after 1976, not one survived.

表2显示，1976年，有17家公司活跃在OE市场上：4家相关技术公司、10家相关市场公司和3家垂直一体化公司。每组中只有一家公司1989年，在最初的参与者中，只有一家仍在生产磁盘驱动器。而在1976年后进入该行业的10家相关市场公司中，没有一家幸存下来。

Although it is tempting to ascribe the 100 percent failure rate of related-market entrant firms to the inappropriateness of such a corporate form in a technologically turbulent market—or to ascribe the success of the start-ups to capabilities or corporate forms better suited to this environment—it appears instead that the firms’ failure and success rates are influenced most strongly by the technology and market strategies they pursued, rather than by their size or corporate structure. Although the entrants to the disk drive industry pursued a wide variety of technology and market strategies, some distinct, central tendencies can be discerned.

虽然我们很容易将相关市场进入者公司百分之百的失败率归咎于这种公司形式在技术动荡的市场中不合适，或者将新成立公司的成功归咎于更适合这种环境的能力或公司形式，但这些公司的失败率和成功率似乎受到了它们所追求的技术和市场战略的最大影响，而不是它们的规模或公司结构。尽管进入磁盘驱动器行业的企业所追求的技术和市场战略多种多样，但还是可以看出一些明显的中心倾向。

There were seventy-seven independent (non-computer manufacturer) entrants to the U.S. rigid disk drive industry between 1971 and 1989. For analytical purposes, “successful firms” in this population were arbitrarily defined as those that achieved more than $50 million in revenues in constant 1987 dollars in any single year between 1977 and 1989—even if they subsequently withdrew from the market. Failed firms were defined as those that were active at any time during the 1977-89 period; that did not achieve sales greater than $50 million in any single year; and that ceased operations in or before 1989. Firms still operating in 1989 that had not yet achieved $50 million in annual revenues were classified as “no verdict.”

在1977年至1989年间，有七十七个独立的（非计算机制造商）新进入公司加入美国硬盘驱动器行业。出于分析目的，我们将这一群体中的“成功企业”定义为在1977年至1989年间的任何一年中，以1987年固定美元价格计算计算收入超过5000万美元的企业——即使它们后来退出了市场。失败企业的定义是：在1977至1989年期间曾活跃过；在任何一年中销售额都没有超过5000万美元；并且在1989年或之前停止运营。1989年仍在运营但年收入尚未达到 5000万美元的公司被归类为“尚未定论”。

To examine the impact that these firms’ market entry strategies had on their success, each firm’s initial product-market approach was characterized according to whether it used new or proven component technology in its first products and whether these products were targeted at emerging or established markets. “New technology” was defined as an innovation in component or architectural technology that had been on the market for three years or fewer, or that fewer than 20 percent of firms in the market had adopted. Hence, thin-film heads, though introduced in 1976, were not considered a proven component technology until 1986, when over 20 percent of the firms in the market finally were offering products using them. Firms were considered to have targeted a new, emerging market if their initial models employed a trajectory-disrupting architectural technology (8-, 5.25-, or 3.5-inch drives) that had been in the market for fewer than two years. They were considered to have targeted an established market if their initial products employed a trajectory-sustaining architecture (14- or 2.5-inch Winchester drives), or if their first product was an 8-, 5.25-, or 3.5-inch product that arrived in the market three or more years after that architecture was first introduced.

为了研究这些公司的市场进入战略对其成功的影响，根据其首批产品是采用新技术还是成熟的组件技术，以及这些产品是针对新兴市场还是成熟市场，对每家公司最初的产品-市场方式进行了定性。“新技术”的定义是在组件或架构技术方面的创新，这种创新在市场上出现的时间不超过三年，或者市场上只有不到20%的公司采用了这种创新。因此，尽管薄膜头于1976年推出，但直到1986年才被视为成熟的部件技术，当时超过20%的公司采用了这种技术。如果公司最初的产品型号采用了轨迹破坏型架构技术（8、5.25 或 3.5 英寸硬盘），且进入市场的时间少于两年，则该公司被视为以新兴市场为目标。如果这些公司的初始产品采用了轨迹持续型的架构（14或2.5英寸Winchester 硬盘），或其首款产品为8、5.25或3.5英寸产品，且在该架构首次推出三年或三年以上后才进入市场，则这些公司被视为以成熟市场为目标。

Figure 6 employs a 2 x 2 matrix to describe the entry technology strategies that characterized the various types of firms. Its horizontal axis denotes whether the firm targeted a new or an established market, and the vertical axis shows whether the entrant's first product incorporated new or proven component technology. The headings in each quadrant denote whether a firm was successful (S), failed (F), or is still operating but has not yet succeeded or failed (N) according to the definition of success noted earlier. T denotes the total number of firms that followed that quadrant's strategy. The numbers under each heading show the number of firms in that category, listed by type of entrant.

图6采用2*2矩阵来描述各类企业的进入技术战略。矩阵的横轴表示企业的目标市场是新市场还是成熟市场，纵轴表示新进入者的首个产品是采用了新技术还是成熟的组件技术。根据前面提到的成功定义，每个象限的标题分别表示企业成功（S）、失败（F）或仍在运营但尚未成功或失败（N）。T表示采用该象限战略的企业总数。每个标题下的数字显示了该类别中按进入者类型分列的公司数量。

The “Totals” statistics show that, overall, entering established markets with proven components was the most common entry technology strategy, followed by thirty-six (48 percent) of the entrants. Only 11 percent of the firms following this strategy succeeded, however. In contrast, twenty-three (30 percent) of the entrants used proven component technologies with architectures targeted at new market applications. Sixty-five percent of these firms succeeded. The success rate of firms on the right half of the matrix, at 56 percent, was five times greater than the rate of those that entered established markets.

总体而言，“总计”统计显示，以成熟的组件进入成熟市场是最常见的进入技术战略，有36家（48%）进入者采用了这一战略。然而，采用这种战略的公司中只有11%取得了成功。与此相反，23家（30%）新进入公司使用了针对新市场应用的成熟组件技术和架构。这些公司中有65%取得了成功。这些公司位于矩阵右半部分的公司的成功率为56%，是那些进入成熟市场的公司成功率的五倍。

Sixty of the seventy-seven entrants were reluctant to seek performance advantage through new component technology in their initial products—possibly reflecting the difficulty often involved in developing and using such innovations. Whether firms used proven or innovative component technology was not, however, a factor markedly related to differences in the success rate.

在77家新加入的公司中，有60家公司不愿意在最初的产品中采用新的部件技术来寻求性能优势——这可能反映了开发和使用这种创新技术通常会遇到困难。然而，企业使用成熟的还是创新的部件技术，并不是与成功率的差异有明显关系的因素。

Figure 6 also shows that, although different types of firms tended to employ different strategies on entry, their success rates are related much more to entry strategy than to type of firm. Seventeen of the twenty-four related-market firms entered the industry with proven-component technology/established-market strategies. This is understandable. Since their basis of diversification was not technological expertise in magnetic recording, one would not expect them to seek competitive advantage by developing new component technology. They entered the disk drive industry because they believed that their marketing capabilities, derived from selling other computer peripheral products, could be the basis of success in the disk drive business as well. Related-market firms had a very low success rate, but firms of every sort that pursued the established-market/proven-components entry strategy had a dismal success rate, ranging from zero among related-technology firms to 15 percent for start-ups (see the lower-left quadrant of Figure 6). The proximate cause of the poor showing of related-market firms in this industry may not be that market-related diversification as a corporate strategy is inappropriate for a technologically turbulent industry, but that the entry strategy that a related-market corporate structure seems to have led them to pursue offered a low probability of success.

图6还显示，虽然不同类型的公司在进入时往往采用不同的战略，但它们的成功率与进入战略的关系要比与公司类型的关系大得多。在二十四家相关市场公司中，有十七家是以成熟的组件技术/成熟市场战略进入该行业的。这是可以理解的。因为它们的多元化基础并不是磁记录方面的专业技术，人们不会指望它们通过开发新的元件技术来寻求竞争优势。它们之所以进入磁盘驱动器行业，是因为它们相信，销售其他计算机外围产品所获得的营销能力，也可以成为磁盘驱动器业务成功的基础。相关市场公司的成功率很低，而采用成熟市场/成熟组件进入战略的公司的成功率也很低，相关技术公司的成功率为零，而初创公司的成功率则为15%（见图6左下象限）。相关市场企业在该行业表现不佳的近因可能不是市场相关的多元化战略不适合技术动荡的行业，而是相关市场企业结构所采用的进入战略导致其成功概率较低。

The entry-technology strategy toward which the related-technology firms tended was substantially different; it was generally based on proven components targeted at new markets. An astonishing 83 percent of related-technology entrants that employed this strategy succeeded, whereas none that targeted established markets with proven components succeeded. The start-ups showed more diversity in strategies, but their success rates follow the same tendencies as the other groups of firms. As with related-market and relatedtechnology firms, the highest success rate for the start-ups, 62 percent, occurred among those firms that fo-llowed a provencomponents/innovative-market strategy. Whether the firms used proven or new component technology does not seem to have been as strong a discriminator of success, although entering with new component technologies seems to have been more difficult for all types of firms. The forward integrators all entered established markets, and none succeeded.

相关技术公司倾向于采用的进入技术战略存在很大差异；这种战略一般以针对新市场的成熟组件为基础。令人惊讶的是，采用这种战略的相关技术进入者有83%取得了成功，而针对成熟市场的进入者则无一成功。初创企业的战略更加多样化，但其成功率与其他组别企业的趋势相同。与相关市场和相关技术公司一样，采用“成熟组件/创新市场”战略的初创公司的成功率最高，达到62%。公司采用成熟组件技术还是新组件技术似乎并不是决定成败的关键因素，但采用新组件技术似乎对所有类型的公司来说都更加困难。前瞻性整合者都进入了成熟市场，但没有一家成功。

The Dis-Integration of Firms in the Disk Drive Industry

磁盘驱动器行业企业的分化

The preceding sections have shown how customers consistently commanded the strategic attention of the established disk drive manufacturers. Their tendency to focus so sharply on current customers, in the face of growing demand for disk drive technology outside the established firms' traditional customer base, seems to have set in motion powerful dis-integrative forces that spun valuable people and technology out of the established integrated companies and into the marketplace—a very “un-Chandlerian” outcome for what was a very Chandlerian beginning in this industry.

前面的章节已经说明，客户是如何不断地吸引老牌磁盘驱动器制造商的战略注意力的。面对老牌公司传统客户群之外对磁盘驱动器技术日益增长的需求，老牌公司如此专注于现有客户的倾向，似乎似乎已经产生了强大的分解力，将宝贵的人才和技术从老牌的整合公司中剥离出来，进入市场——对于这个行业非常钱德勒式的开端来说，这是一个非常“非钱德勒式”的结果。

The spinning out of people and technology from the established companies occurred at the levels of both architectural and component technologies. The outcome of the processes through which new architectural technologies were spun out of the established firms, described earlier, are reflected in the corporate genealogies of the leading disk drive manufacturers shown in Figure 7. Although the chart does not show all of the start-ups that entered the industry, it captures all but four of the start-ups that actually generated revenue, and the firms shown accounted for 99.4 percent of the total cumulative revenues generated by the start-up group. All but one of the start-ups can trace their genealogies to IBM, and all were sired by three plug-compatible market manufacturers that were IBM offspring—Pertec Computer, Storage Technology, and Memorex. Of the three IBM offspring with progeny, the descendants of Memorex, via Shugart Associates, were overwhelmingly the most productive, accounting for over $9.14 billion, or 64 percent, of the cumulative revenues of the start-ups.37 (Cumulative revenues for their descen-dants, in $ millions, and the percentage of the total for which they accounted are listed in the IBM, Datapoint, Memorex, STC, Pertec, and Shugart boxes.) Moreover, six of the seven Shugart spin-offs that generated revenue were still in operation in 1991 and included the U.S. industry's four largest firms—Seagate, Conner, Quantum, and Maxtor.38 This compares with one of eight Pertec spin-offs still in operation (Micropolis) and none of the descendants of Storage Technology.

人员和技术从老牌公司中剥离的过程既发生在架构技术层面，也发生在组件技术层面。图7所示的主要磁盘驱动器制造商的公司谱系反映了前文所述的新架构技术从老牌公司分离出来的过程的结果。虽然该图并未显示所有进入该行业的初创公司，但它捕捉到了除四家公司外所有实际创收的初创公司，而且图中所示公司占初创公司群累计创收总额的99.4%。除一家公司外，所有初创公司的谱系都可追溯到IBM，而且所有初创公司的前身都是IBM的三家插头兼容市场制造商——Pertec Computer、Storage Technology和Memorex。在这三个IBM的子公司的后代中，Memorex的后代通过Shugart Associates取得了压倒性的成功，占初创公司累计收入的64%，即91.4亿美元。（其后代的累计收入（以百万美元为单位）以及占总收入的百分比列于IBM、Datapoint、Memorex、STC、Pertec和Shugart方框内。）此外，1991年Shugart的七个衍生公司中有六个仍在运营，其中包括美国工业界最大的四家公司——希捷、康纳、昆腾和迈拓。相比之下，在八家仍在运营的Pertec衍生公司中，只有一家仍在运营（Micropolis），而Storage Technology的子公司则无一仍在运营。

One might term the pattern of spin-outs shown in Figure 7 one of horizontal dis-integration, where independent firms split off from predecessor firms to focus on different market segments. There seems to have been a pattern of vertical dis-integration in the industry as well. Whereas the industry’s original dominant firms were thoroughly integrated into component manufacturing and the research required to support advanced component development, those firms’ focus on a single set of (primarily internal) customers induced the spin-out of a host of independent companies that supplied components to the progressively less integrated set of disk drive manufacturers arrayed in Figure 7. Although the story of the creation of a network of independent component manufacturers is complicated enough to merit its own history, a brief summary of how and why independent component manufacturing firms spun out of the industry’s initially dominant integrated firms will be helpful here.

人们可能会将图7中所示的拆分模式称为横向分解，即独立公司从母公司拆分出来，专注于不同的细分市场。该行业似乎也存在垂直分解的模式。由于行业中的原始主导企业已完全整合到组件制造和支持先进组件开发所需的研究中，这些企业专注于单一客户群（主要是内部客户），导致许多独立公司脱颖而出，为图7中整合程度逐渐降低的磁盘驱动器制造商提供组件。虽然创建独立零部件制造商网络的故事足够复杂，值得单独成文，但在此简要介绍独立零部件制造公司如何以及为何从行业最初占主导地位的公司中分离出来，将有所帮助。

Firms that entered the industry with a new architectural technology targeted at an emerging market generally shifted their technological sights toward improved component technology, because improved componentry was the engine of performance improvement within each established product architecture. This shift in technology strategy from architectural orientation to component orientation entailed a significant change in the economics of product development for IBM and the other early industry leaders, because no network of component supply firms existed. Development of new product architectures generally had been an engineering task, not an issue of research and development. But the development of new component technology required substantial investment in research and development, because component technology development is where basic scientific research, engineering, system design, product design, and process development all come together. In terms of the time, expense, and expertise required, component technology development was an enterprise fundamentally different from the design of new product architectures that employed available componentry. This difference between the pace and scale of component development and those of product design eventually made it impossible for the integrated firms to perform internally the coordinating roles that Chandler observed in other industries in funneling new component technologies into new product designs.

以针对新兴市场的新架构技术进入该行业的公司，一般都会将其技术转移到新兴市场，因为改进组件是每个既定产品架构内性能提升的引擎。技术战略从架构导向转向组件导向，这种技术上的转变对IBM和其他早期的行业领导者来说，意味着产品开发的经济性发生了重大变化，因为当时还没有组件供应公司网络。开发新的产品架构通常是一项工程任务，而不是研发问题。但是，开发新的组件技术需要在研发方面进行大量投资，因为组件技术开发是基础科学研究、工程、系统设计、产品设计和工艺开发等所有方面的结合。就所需时间、费用和专业知识而言，零部件技术开发与采用现有零部件的新产品架构设计有着本质区别。零部件开发与产品设计在速度和规模上的差异，最终导致整合型企业无法像钱德勒在其他行业中观察到的那样，在内部发挥协调作用，将新的零部件技术整合到新的产品设计中。

The following account of IBM’s development of the thin-film head illustrates four phases that are typical of the course of events that occurred in the development and diffusion of many components within most of the industry’s early leading integrated firms, including Control Data, Xerox (which supported its Diablo, Shugart, and Century disk drive divisions through its Palo Alto Research Center), Burroughs, Digital Equipment, and Seagate Technology.

以下关于IBM开发薄膜磁头的叙述主要阐述了四个阶段，这四个阶段是该行业大多数早期领先的整合型公司在开发和推广许多组件过程中发生的典型事件，这些公司包括Control Data、Xerox（通过其帕洛阿尔托研究中心支持其Diablo、Shugart和Century磁盘驱动器部门）、Burroughs、Digital Equipment和Seagate Technology。

The Disparate Cycles of Component and Systems Development: IBM and the Thin-Film Head

组件和系统开发的不同周期：IBM和薄膜磁头

• As the industry’s pioneer and dominant firm during this period, IBM led the shift in technology strategy from architecture development to component development. Improved component technology defended IBM’s large, growing, and very profitable mainframe business by providing its customers with steady performance improvement within the 14-inch Winchester architecture. Backward integration reduced the uncertainty that required components might not be available to meet customer and competitive requirements in the next product generation.

·作为这一时期的行业先驱和主导者，IBM引领着技术战略从架构开发向组件开发的转变。发达的组件技术为客户提供了14英寸Winchester架构内稳定的性能改进，从而保护了IBM不断增长且利润丰厚的大型主机业务。后向一体化减少了所需组件在下一代产品中无法满足客户和竞争要求的不确定性。

In the earliest stages of thin-film head development, IBM worked alone to expand its understanding of basic scientific issues such as the physics of magnetic recording and the properties of new materials. This phase began in 1965 in IBM’s advanced research facilities at Yorktown Heights, New York, and San Jose, California. The second phase (beginning roughly in 1971 and ending in 1976- 78) was stimulated by proof of the concept at IBM and by the spread of that information through published scientific papers and the trade press to other firms. Statements by respected IBM scientists that thin-film technology was important and feasible led a broader group of vertically integrated manufacturers—Burroughs, Control Data, Digital Equipment, Fujitsu, Hewlett Packard, Hitachi, and NEC—to initiate their own development efforts. IBM’s pathbreaking research resolved a lot of uncertainty for these other firms: once they knew that something could be done, they could focus with greater commitment on learning how to do it. The end of this second phase was marked by the building of early working prototype thin-film heads.

在薄膜磁头开发的最初阶段，IBM独自致力于拓展对磁记录物理和新材料特性等基本科学问题的理解。这一阶段始于1965年，地点在IBM位于的纽约州约克敦高地和加利福尼亚州圣何塞的高级研究机构。第二阶段（大致从1971年开始，到1977年结束）是在IBM公司对这一概念进行验证，并通过发表科学论文和行业媒体向其他公司传播这一信息后产生的。IBM颇具名望的科学家们认为薄膜技术是重要且可行的，这促使更多的垂直一体化制造商——伯勒斯、Control Data、数字设备、富士通、惠普、日立和NEC——开始采用薄膜技术。启动自己的开发工作。IBM的开创性研究为其他公司解决了许多不确定因素：一旦他们知道某件事情可以做到，他们就可以更加专注地学习如何去做。第二阶段的结束标志着早期工作原型薄膜磁头的建成。

In the third phase, the component design was refined, a manufacturing process was established, and the component was designed into a new disk drive model. In 1976 Burroughs was the first in the industry to announce a drive with a thin-film head, but it was never able to manufacture the head reliably and withdrew the drive from the market. IBM introduced its model 3370, equipped with thin-film heads, in 1979. Positioned at the highest-performance end of IBM’s line, the 3370 was a very successful product, even though the heads were extremely difficult and expensive to manufacture.

在第三阶段，对部件设计进行了改进，建立了制造流程，并将部件设计到新的磁盘驱动器型号中。1976年，Burroughs在业内率先宣布推出配备薄膜磁头的硬盘，但该公司始终无法可靠地生产磁头，并将硬盘撤出市场。1979年，IBM推出了配备薄膜磁头的3370型硬盘。3370定位于IBM产品线中最高性能的一端，是一款非常成功的产品，尽管其磁头制造难度极大，成本高昂。

To this point, the component technology leadership of the large, vertically integrated firms—particularly IBM—was unambiguous. In moving to the fourth phase, however, the story becomes troubled. Although IBM usually initiated each component development process in behalf of the industry and was often the first to introduce the new component in a high-end product model, IBM and the other vertically integrated manufacturers subsequently were very slow to employ the componentry they had developed in other new models in their product lines. The vertically integrated firms' commercial introduction of the new components in a limited number of high-end models typically stimulated the fourth phase in the emergence of new component technology, in which demand for the new componentry became intense among certain independent, non-integrated disk drive manufacturers. These independent firms, such as Maxtor and Micropolis, generally pursued technology strategies that pushed, through innovative (some would say risky) system design, what was called in the industry "the bleeding edge" of performance—a much more aggressive engineering posture than the vertically integrated manufacturers typically were inclined to adopt.

至此，大型垂直一体化公司（尤其是IBM）在组件技术方面的领先地位已不言而喻。然而，进入第四阶段后，情况就变得扑朔迷离了。尽管IBM通常代表整个行业启动每个部件的开发过程，而且往往是第一个在高端产品型号中引入新部件的公司，但IBM和其他垂直一体化制造商随后在其产品线的其他新机型中采用他们开发的部件时却非常缓慢。垂直一体化公司在有限的高端机型上对新部件进行商业化引入，通常会刺激新部件技术进入第四阶段，在这一阶段，某些独立的非一体化磁盘驱动器制造商对新元件的需求变得十分迫切。这些独立公司，如Maxtor和Micropolis，通常通过创新（有人会说是冒险）的系统设计来推行技术战略，在业内被称为性能的“尖端”——这比垂直整合制造商通常倾向于采用的工程姿态更具侵略性。

IBM viewed its proprietary access to advanced componentry as its primary competitive advantage and was reluctant to sell its components in the external marketplace, because it viewed the bleedingedge manufacturers that most needed thin-film heads as indirect competitors. Sensing the opportunity to match component supply and demand more closely, several venture capitalists recruited key IBM and Xerox engineers into new start-up firms to produce and sell the new-technology components to bleeding-edge disk drive makers in the original equipment market. This was possible because the vertically integrated firms typically enjoyed little patent protection for the components: much of the key technology consisted of process know-how. The industry’s leading thin-film disk manufacturer, Komag, and the leading thin-film head manufacturer, Read-Rite, both started in this manner. In 1992, within eight years of their founding, these firms together logged over $1 billion in revenues.

IBM公司将其对先进组件的专有使用权视为其主要竞争优势，并不愿将其组件出售给外部市场，因为它认为最需要薄膜磁头的尖端制造商是间接竞争对手。一些风险投资家意识到，应该更紧密地匹配零部件的供需，于是招募了IBM和施乐公司的关键工程师，成立新的初创公司，为原始设备市场中的尖端磁盘驱动器制造商生产和销售新技术零部件。之所以能做到这一点，是因为垂直一体化公司通常很少对组件进行专利保护：大部分关键技术都是工艺诀窍。业内领先的薄膜磁盘制造商Komag和领先的薄膜磁头制造商Read-Rite都是以这种方式起家的。1992年，这些公司的总收入超过了10亿美元。

When the components became available from these start-up firms, even on an irregular, unpredictable, low-yield basis, disk drive manufacturers tended to utilize the new technology throughout their product lines at a pace that was roughly commensurate with their distance from the bleeding edge. And in general, the integrated manufacturers such as IBM tended to design conservatively, some distance from the bleeding edge. As a result, although the vertically integrated firms were the first to develop and introduce the new components, they were the slowest to incorporate them across the breadth of their product lines (see Figs. 8a and 8b). This was most strikingly the case with IBM. Although it spent over $300 million developing the thin-film head, it was the last firm in the industry to use the technology broadly in its entire product line. And although IBM spent over $100 million developing thin-film disks, it did not use them in any commercial product until 1988.

当这些初创公司提供元件时，即使是在不规则、不可预测、低产量的基础上，磁盘驱动器制造商也倾向于在其整个产品线中采用新技术，其速度与他们和尖端的距离大致相称。而一般来说，IBM等制造商倾向于保守设计，与尖端保持一定距离。因此，尽管垂直一体化企业最先开发和引进新元件，但它们在其产品线的广度上采用这些元件的速度却最慢（见图8a和8b）。IBM公司的情况最为突出。虽然它花费了3亿多美元开发薄膜磁头，但却是业内最后一家在整个产品线中广泛使用该技术的公司。虽然IBM花了1亿多美元开发薄膜磁盘，但直到1988年才将其应用于商业产品。

The unfortunate combination of the integrated firms’ leadership in component development and their followership in component use seems to have been the result of an inexorable decoupling of component development from systems design as the industry matured. This was a problem that IBM worked hard to remedy but that in many ways derives from the fundamentally different natures and time scales of the two processes. Initially, because it was such a dominant presence in the computer markets, IBM was able to control the pace at which it introduced new models into the market—roughly every four years in the 1970s. IBM was therefore reasonably capable of coordinating the emergence of new long-lead-time components with the design of the product systems in which they initially would be used. As the computer industry matured, however, competitive market forces increasingly became the drivers of the product development cycle. As IBM lost its ability to control that cycle, coordination between component development and system design became inherently more difficult.

整合型公司在部件开发方面的领先地位和在部件使用方面的追随者地位的不幸结合，似乎是随着工业的成熟，部件开发与系统设计不可避免地脱钩的结果。IBM公司一直在努力解决这个问题，但在很多方面，这两个过程的性质和时间尺度根本不同。起初，由于IBM在计算机市场上占据主导地位，因此它能够控制向市场推出新机型的速度——在20世纪70年代，大约每四年一次。因此，IBM有理由能够协调新的长周期部件的出现和它们所在产品系统的设计。然而，随着计算机行业的成熟，市场竞争力量日益成为产品开发周期的驱动力。由于IBM失去了控制这一周期的能力，组件开发与系统设计之间的协调变得更加困难。

When the component development cycle was decoupled from the product development cycle by the competitive market, it simply became impossible for anyone to predict accurately, a decade in advance, whether the company’s product position in the market would demand that a specific new component technology be used in its products. The market shifted, the conventional technology progressed, and—possibly the most unpredictable of all—the company’s product-market strategy changed. When fully committed development of thin-film head and disk technology was initiated in the early 1970s, almost no one could have imagined that, a decade later, 80 percent of IBM’s drives would be used in relatively non- demanding desktop personal computer applications.

当组件开发周期与产品开发周期在市场竞争中脱钩后，任何人都不可能提前十年准确预测公司产品在市场中的地位是否要求在其产品中使用特定的新组件技术。市场在变化，传统技术在进步，可能最难以预料的是，公司的产品市场战略也发生了变化。20世纪70年代初，当公司开始全力开发薄膜磁头和磁盘技术时，几乎没有人会想到，十年后，IBM 80%的硬盘将用于要求相对较低的台式个人电脑应用中。

In response to the decoupling forces, IBM’s managerial practice evolved toward a sort of “free market” system. When new component technologies were developed, they were made available to (but not forced on) product design engineers. The designers of new disk drive systems were free to choose whatever component technologies IBM had in its arsenal to meet the performance objectives of their product. Under this arrangement, IBM’s market position—the demands of its customers—determined the pace at which the company employed advanced component technologies in its new models. Figure 9 charts over time the product positions (average megabytes of all models in the 5.25-inch architecture) of Micropolis, Maxtor, and IBM. It shows that in 1984 Maxtor’s average 5.25-inch model held 125 megabytes, Micropolis’s held 64, and IBM’s held 16.5. There is nothing normative about this observation—these firms were simply serving different markets. Maxtor was selling to the memorystarved engineering workstation market, whereas IBM was making 5.25-inch drives for its XT and AT personal computers

为了应对脱钩力量，IBM的管理实践逐渐演变为一种“自由市场”体系。当新的组件技术被开发出来时，它们会被提供给(但不会强制)产品设计工程师。新的磁盘驱动器系统的设计人员可以自由选择IBM产品库中的任何组件技术，以满足其产品的性能目标。在这种安排下，IBM的市场地位——即客户的需求——决定了公司在新机型中采用先进组件技术的速度。图9显示了Micropolis、Maxtor和IBM的产品地位（5.25英寸架构中所有型号的平均兆字节数）。图中显示，1984年MaxTor 5.25英寸机型的平均容量为125兆字节，Micropolis为64兆字节，IBM为16.5兆字节。这一观察结果并不规范——这些公司只是服务于不同的市场。迈拓面向内存匮乏的工程工作站市场销售硬盘，而IBM则为其XT和AT个人电脑生产5.25英寸硬盘。

Figure 9 also shows the percentage of each firm’s 5.25-inch models in a given year that employed thin-film heads, thin-film disks, and RLL codes. Each of these technologies was developed at IBM but, in 1984, none of them was used in the IBM or Micropolis product lines, whereas Maxtor used thin-film disks in all of its models, with ferrite heads and MFM codes. In 1985, Micropolis adopted thin-film disks on 100 percent of its new models, and Maxtor and Micropolis converted completely to RLL codes. But IBM still did not use these technologies—it continued to support its market position with established technologies, which were much less costly and risky to use. In 1986, Maxtor and Micropolis both began using thin-film heads, while IBM was still able to satisfy its requirements with conventional technology. Finally, when its average 5.25-inch drive approached the 80-100 MB range in 1987, IBM began using RLL codes in 88 percent of its new models (Micropolis had adopted RLL codes when its products had penetrated this same range two years earlier). When its 5.25-inch drives reached even more demanding territory in 1988, IBM adopted thin-film disks on 62 percent of its new models. Although this step was taken four years after Maxtor had adopted thin-film disks, it occurred when IBM's drives reached the same capacity territory that Maxtor and Micropolis had occupied when they first used thin-film disks. Finally, though IBM had first used thin-film heads in a few high-end 14-inch drives as early as 1979, thin-film heads were not used in its 5.25-inch models until 1990.

图9还显示了各公司在特定年份的5.25英寸型号中采用薄膜磁头、薄膜磁盘和RLL代码的百分比。这些技术都是IBM开发的，但在1984年，IBM和Micropolis的产品线中都没有使用这些技术，而MaxTor在其所有型号中都使用了薄膜磁头、铁氧体磁头和MFM代码。1985年，Micropolis 100%的新型号都采用了薄膜磁盘，而MaxTor和Micropolis完全转用RLL代码。但IBM仍然没有使用这些技术，而是继续使用成本更低、风险更小的成熟技术来支持其市场地位。1986年，MaxTor和Micropolis都开始使用薄膜磁头，而IBM仍能用传统技术满足其需求。最后，1987年，当5.25英寸硬盘的平均容量接近86-100 MB时，IBM开始在其88%的新机型中使用RLL代码（两年前，当Micropolis的产品进入这一范围时，它也采用了RLL代码）。当其5.25英寸驱动器在1988年达到更高要求时，IBM在其62%的新型号上采用了薄膜磁盘。尽管这一步是在迈拓采用薄膜磁盘四年后才迈出的，但当时IBM的驱动器容量已经达到了MaxTor和Micropolis最初使用薄膜磁盘时的水平。最后，尽管IBM早在1979年就首次在几款高端14英寸驱动器上使用了薄膜磁头，但直到1990年才在5.25英寸型号中使用薄膜磁头。

By drawing horizontal lines across Figure 9, one can create product performance zones that seem to have mandated the use of particular component technologies. IBM’s computer business had devolved to a performance position far from the market's leading edge by the mid-1980s, and it had not yet entered the zones requiring use of the component technologies it had developed. On this basis, it would be difficult to argue that IBM’s failure to utilize more broadly the component technologies it had paid so dearly to develop was the result of conservative or inept technical management. Simply put, the firms that needed the new component technologies used them; the firms that did not, did not.

通过在图9上画水平线，我们可以创建产品性能区，这些性能区似乎必须使用特定的组件技术。到20世纪80年代中期，IBM的计算机业务已经落后于市场的领先地位，而且还没有进入要求使用其开发的组件技术的区域。在此基础上，很难说IBM没有更广泛地利用它花费巨资开发的组件技术是保守或技术管理不善的结果。简单地说，需要新组件技术的公司使用了这些技术，而不需要的公司则没有使用。

Because IBM turned out not to need all of the component technology it had developed when it became available and yet had a policy of not selling components outside the company, the independent component suppliers that spun out of IBM generally grew to become larger producers of the components than did IBM itself. As a result, the firms that incurred the development costs of the new component technology were not those that generated the revenues (see Table 7 for the case of thin-film heads). IBM, which spent onethird of the industry’s R&D dollars for this technology, has produced altogether only 8 percent of the industry’s cumulative output of thinfilm heads. The independent start-ups, after incurring only 15 percent of development costs, have captured most of the market. A similar table could be constructed for thin-film disk technology as well. Although the disparity between costs incurred and units produced may represent an unfortunate subsidization of the competition from the point of view of the IBM shareholders, it is clear that the disk drive industry—and a great many entrepreneurs and venture capitalists within it—have benefited greatly from IBM’s extraordinary technological largesse.

因为IBM在获得其开发的全部组件技术后，发现并不需要全部技术，但公司又规定不得向公司外部出售组件，因此从IBM分拆出来的独立组件供应商通常成长为比IBM本身更大的组件生产商。结果，承担新组件技术开发成本的公司并不是那些创造收入的公司（参见表7中的薄膜磁头案例）。IBM公司为这项技术投入了业界研发经费的三分之一，但总共只生产了业界累计薄膜磁头产量的8%。而独立初创公司只承担了15%的研发成本，却占据了大部分市场。类似的表格也可以用于薄膜磁盘技术。虽然从IBM股东的角度来看，成本与产量之间的差距可能意味着对竞争的一种不幸的补贴，但很明显，磁盘驱动器行业以及其中的许多企业家和风险投资家都从IBM超常技术慷慨解囊中获益匪浅。

Conclusion

结论

Coordinating the very different enterprises of component development and product system design in an increasingly segmented market became a nearly impossible challenge for the integrated firms as the disk drive industry matured through the 1980s. And in the face of strong, diverse market demand, managers in vertically integrated firms found it difficult to protect or retain valuable component and architectural technologies that their customers did not want. As a result, by the 1990s nearly all of the industry’s firms had decoupled their vertically integrated operations to some degree, enabling groups at each stage of the value chain to sell their output in the original equipment market. Control Data, Fujitsu, Hitachi, and NEC had followed a policy of selling completed disk drives in the OE market from the late 1970s. By 1990, IBM, Digital Equipment, and Hewlett Packard all had followed suit, selling disk drives aggressively in the OE market to their computer system competitors, rather than continuing to cede that business to defecting engineers. And some firms, such as Digital Equipment and Seagate, became leading suppliers of thin-film disks and heads, not just to their own downstream disk drive operations, but to direct disk drive competitors as well. Hence, an industry whose foundation and growth were built through the activities of large-scale, integrated organizations became in its more mature years an industry where market mechanisms forced the decoupling and specialized focus of enterprises that once were extensively integrated. This process created an industry structure where market mechanisms and interfirm transactions, rather than managerial coordination within large-scale firms, became the means for coordinating the development and manufacture of disk drives.

20世纪80年代，随着磁盘驱动器行业的成熟，在日益细分的市场中协调组件开发和产品系统设计这两个截然不同的业务，对整合型公司来说几乎是不可能完成的挑战。面对强劲而多样化的市场需求，垂直一体化公司的经理们发现很难保护或保留客户不需要的有价值的元件和架构技术。因此，到20世纪90年代，该行业几乎所有的公司都在一定程度上实现了垂直一体化业务的脱钩，使价值链每个阶段的企业都能在原始设备市场上销售其产品。从20世纪70年代末开始，Control Data、富士通、日立和NEC就一直奉行在原始设备市场上销售成品磁盘驱动器的政策。到1990年，IBM、Digital Equipment和惠普公司都纷纷效仿，在原始设备市场上积极向计算机系统竞争对手销售磁盘驱动器，而不是继续将业务拱手让给叛逃的工程师。一些公司，如Control Data公司和希捷公司，成为薄膜磁盘和磁头的主要供应商，不仅为自己的下游磁盘驱动器业务提供服务，也为直接的磁盘驱动器竞争对手提供服务。因此，一个通过大型综合组织的活动建立起基础和发展起来的行业，在其较成熟的年代里，市场机制迫使曾经广泛整合的企业脱钩并专注于专业化。在这一过程中，市场机制和企业间交易而非大型企业内部的管理协调，成为协调磁盘驱动器开发和制造的手段。

阅读体会：

本文简要介绍并探讨了硬盘驱动器行业的历史，特别是通过对比持续性与破坏性两种不同的技术创新模式，挖掘了该行业的颠覆性技术创新对企业的市场领导地位的影响。文章指出，大型垂直一体化企业在前中期是行业的主导者，但随着技术和市场需求的快速变化，这些企业难以维持其领导地位。

文章分析了组件技术和架构创新的不同影响，以及这些创新如何被不同类型企业所采纳。老牌企业通常在组件技术创新方面处于领先地位，而新进入市场的企业则更可能在架构创新方面取得突破，因为这些创新只有较少的技术负担，能够提供全新的产品解决方案，满足新兴市场的需求。由于老牌领先企业陷入仅满足当前顾客需求的短期视角，其组件技术开发恰好为新进入企业提供了技术基础，便利了这些企业的创新。在市场的作用下，新进入的企业通过颠覆性技术创新，抢先占领新兴市场份额并快速进行迭代，最终专业化新进入者成功颠覆老牌垂直一体化企业。

此外，本文首次提出了颠覆性创新的概念，结合硬盘驱动器行业组件技术及架构技术的创新史来理解，提高了我对于颠覆性创新的基本认知。