Acorn ARM published Research Policy 2008

This article was published in an Elsevier journal. The attached copyis furnished to the author for non-commercial research and

education use, including for instruction at the author’s institution,sharing with colleagues and providing to institution administration.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Research Policy 37 (2008) 210–224

Available online at www.sciencedirect.com

Speciation through entrepreneurialspin-off: The Acorn-ARM story

Elizabeth Garnsey a,∗, Gianni Lorenzoni b, Simone Ferriani b

a University of Cambridge, Center of Technology Management, Mill Lane, Cambridge CB2 1RX, UKb University of Bologna, Management Department, Via Capo di Lucca 24, 40126 Bologna, Italy

Received 25 April 2006; received in revised form 13 September 2007; accepted 1 November 2007Available online 22 January 2008

Abstract

Can the concept of speciation explain evidence on how technologies branch and advance? Can evidence on innovation throughspin-off usefully inform the concept of speciation? These questions are addressed through a case study of detailed processes enablingthe shift of technology to new domains of application. An innovative IT firm developed its own semiconductor technology to remedysupplier deficiencies but it required a joint venture with a completely new business model to adapt and move the technology into newmarket domains. We propose the concept of techno-organizational speciation to delineate this phenomenon. Competing perspectiveson speciation (compatibility, niche and lineage approaches) are found to illuminate the evidence, while complementarities betweenthese conceptual dimensions are revealed by the case. Causal processes uncovered include the following: (1) Techno-organizationalspeciation through spin-off may be needed to launch a dominant technical standard, compatible with multiple applications. (2) Thiscan be achieved through niche creation from which develops a new business ecosystem. (3) Inherited knowledge together withorganizationally based learning foster the branching and renewal of technological lineages.© 2007 Elsevier B.V. All rights reserved.

Keywords: Spin-off; Technological speciation; Parent–progeny; Technological innovation

1. Introduction

The branching and diversification of related technolo-gies is a feature of technological advance. The firms thatincubate and carry innovations also branch and diver-sify as new firms are spun off from old (Klepper andSleeper, 2005). The idea that these two processes areclosely connected is intuitive, but the association has notbeen explored in detail in the literature on innovation.This study uses the concept of speciation to explain a case

∗ Corresponding author.E-mail addresses: [email protected], [email protected]

(E. Garnsey), [email protected] (G. Lorenzoni),[email protected] (S. Ferriani).

where the spin-off process stimulated the branching andadvance of technologies. The evidence informed theoryby delineating the concept of techno-organizational spe-ciation and illuminating its complementary dimensions.

In ecology, speciation describes the process throughwhich a new natural species may emerge as a sub-population adapts to new ecological conditions (Dupre,2001). Technological speciation refers to a new tech-nology that emerges as an existing technological formis adapted to a new selection environment (Levinthal,1998). For example, when the Internet expanded itsdomain from military and academic applications to theworld of consumers in the mid-1990s, it was trans-formed into the World Wide Web (enabled by newstandards) reflecting the very different features and

0048-7333/$ – see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.respol.2007.11.006


E. Garnsey et al. / Research Policy 37 (2008) 210–224 211

attributes demanded by consumers and firms in the newmarkets. The shift in domain of application allowed aniche technology to become a global phenomenon andstimulated a host of ancillary technologies.

The concept of speciation can be used to tracecontinuity alongside discontinuity. In the study of tech-nological innovation, continuity has been depicted interms of path dependence, the cumulative influence ofthe past on developments (Arthur, 1990). In contrast,discontinuities in technology are highlighted by the con-cept of the technological paradigm, which points tocognitive and practical disjunctures that set off the path-breaking trajectories in a new technological paradigm(Dosi, 1984). Technological speciation addresses a spec-trum rather than disjuncture: in the lineage of a newtechnological species there is continuity, but applicationto a new domain results in discontinuity. Over the last10 years research on technological speciation has pro-vided several accounts of innovations stemming fromthe transfer of specific technological know-how to anew domain (Levinthal, 1998; Adner and Levinthal,2002; Cattani, 2006). Little is known, however, aboutthe actual processes that underlie the transfer of exist-ing know-how into a new domain and allow subsequentadaptation to the selection forces operating there. Whatmicro-developments and decisions prompt the branchingof an existing technical lineage? What transformationsare required in order for the existing technological know-how to be adapted to new selection forces? How does thisprocess unfold over time?

In this study we address these questions by mov-ing beyond speciation in the incumbent firm, addressedin the literature, to examine an incumbent companythat spun out a new venture into a new domain, wheredifferent market forces operate. We found that the inher-ited technology mutated into a distinctive combinationof new technology and business capability. We pro-pose the concept of techno-organizational speciation todepict the emergence of a new species of technology andthe concurrent crafting of organizational arrangementsthat allow its successful exploitation in the new marketenvironment. We investigate this dynamic through theexemplar of Acorn Computers and its offshoot ARM,one of the most successful spin-offs in the history ofEuropean technology-based industry. The RISC chip(Reduced Instruction Set Computer Chip) launched byARM has had a major impact in the semiconductorindustry, but its core technology was developed in AcornComputers as a microprocessor for its PC products. AtARM, this microprocessor was licensed as a core adaptedto design needs in new market niches, and came to be anew platform technology.

The paper is organized as follows. After an intro-duction to relevant concepts and literature, we describethe design of our inquiry and go on to outline our casestudy, in which we present the antecedents to the spin-off of ARM and its outcome. In the interpretation of theevidence we use the concept of speciation as a patternrecognition device to identify generic issues raised bythe evidence. We offer an explanation of this instance oftechnological branching and diversification, and iden-tify organizational conditions conducive to the processof speciation through spin-off that are revealed by thecase.

2. Speciation, technological innovation andentrepreneurial spin-offs

The richness of the idea of speciation is not fortuitousbut reflects longstanding research using this concept inbiology and ecology (Ayala and Fitch, 1997). Darwin’sevidence from the Galapagos Islands fed understand-ing of the detailed process by which one species cangive rise to others through natural selection operatingin different habitats. Darwin viewed a species not as anentity but as a category to aid classification of the vari-eties of life forms (Darwin, 1887). There are inevitabledifferences of perspective on how best to classify liv-ing forms, which has resulted in controversy on speciesboundaries. These differences have outlived Darwin andare reflected in three main (and to some extent com-peting) approaches to ‘species’ in the natural world,and corresponding perspectives on the speciation pro-cess (Dupre, 2001). Assigning an empirical instance toa category is an endemic classification problem. Judge-ment is involved in setting such boundaries, whether ofnatural species or of technological trajectories (Dosi,1984). We found that three major approaches to iden-tifying natural species also provided relevant criteriafor drawing boundaries between types or species oftechnology.1

(1) Incompatibilities prevent members of a species’population from inter-breeding with those of other

1 While we apply the biological framework to understand the natureof the selection environment acting on possible technologies and,in particular, the niche structure of the resource space, we are notassuming a process of blind variation generation. Indeed, the roleof intentionality and choice in technology development clearly dis-tinguishes it from processes of biological change. Unlike the blinddeterminism of natural selection described by Darwin, technologyselection processes, when understood by participants, give rise to learn-ing on the part of actors and deliberate efforts to change their selectionor opportunity space.


212 E. Garnsey et al. / Research Policy 37 (2008) 210–224

species. Constraints on species’ replication are addressedby this concept of species (Panhuis et al., 2001). (2) Eco-logical perspectives on speciation highlight the selectionenvironment and the niches and ecosystems in whichdistinctive selection forces operate. (3) The lineageapproach categorizes species in terms of common ances-try and evolving lineage that splits as a new speciesemerges (Claridge et al., 1997).

All three perspectives on speciation are relevant fordifferentiating between technologies (Nelson, 2006).(1) Compatibility between technologies is a salient cri-terion for distinguishing between technological types.Differences are formalised in technology standards thatdefine compatibilities between designs. As a new tech-nology is altered to the point of incompatibility with theoriginator technology, this brings to an end the designreplication process that keeps a technological or otherspecies from extinction (Rohlfs, 2001). (2) Technolo-gies are subject to selection forces in different niches ormarkets, which operate as forces differentiating betweentechnologies as they evolve. (3) Processes of technologi-cal inheritance occur as a design is replicated and adaptedin one organization after another, creating identifiabletechnological lineages in applications and products. Allthree of these speciation perspectives illuminate our caseevidence, which in turn proved to clarify connectionsbetween the perspectives.2

Over the last 10 years a growing number of evolution-ary economists have adopted a speciation lens to explainand illustrate the emergence of new technologicaltrajectories in a variety of industries. Technologi-cal innovations as diverse as wireless communication(Adner and Levinthal, 2002), the steam engine (Frenkenand Nuvolari, 2004), fibre optics (Cattani, 2006) andbroadcast radio (Levinthal, 1998) share a common devel-opmental pattern. All these technologies underwent aprocess of evolutionary development within their origi-nal domain of application. At some juncture, however,that set of technologies was applied to a new domainof application in which they evolved in new directions.The technological advance necessitated by this event

2 To term a technology type a ‘species’ of technology and the emer-gence of a new type of technology in new markets a ‘speciation process’is not to base reasoning on a figure of speech. It is to tap into a richanalytical vein, highly salient to the analysis of branching and diversi-fying technologies. Evolving technologies are among a class of systemsthat are subject, in different ways, to common evolutionary processes(Beinhocker, 2006). It is inappropriate to seek equivalence across thesesystems beyond generic processes of variety creation, selection and dif-fusion through which designs are replicated, branch and diffuse overtime (Nelson, 2006).

was not discontinuous. But the different selection cri-teria and new resources available in the new domainof application resulted in a technology that becamequite distinctive, even though path dependence – thestrong influence of prior developments – characterizedthe routes taken from the technology of origin. Researchon technological speciation shows that what appear tobe radically new technologies have gradual and incre-mental features when scrutinised more closely (Cattani,2006).

The literature on spin-offs is clearly relevant tothe speciation process that occurs when a new firmendowed with inherited knowledge branches out fromits originator and strives to apply that knowledge ina new economic domain (Agarwal et al., 2004). Sur-prisingly, however, this literature has not made useof ideas introduced by the speciation framework, norhas the stream of work on technological speciationexplored the micro processes that set off the initialsplitting of a common lineage and occasion subse-quent adaptation of a sub-group to the new domain.Most spin-off studies have focused on the new ven-ture without tracing evolutionary antecedents in theorganization of origin (Mustar et al., 2006). Alterna-tively they focus on the organizational endowment atfounding without a detailed account of the early pro-cesses at the spin-off level (Shane and Stuart, 2002).For instance there is an extensive stream of researchthat emphasizes the effects of spin-offs’ origin on theirsubsequent survival rate and performance (Agarwal etal., 2004; Cooper and Gimeno-Gascon, 1992; Klepper,2001; Lindholm-Dahlstrand, 1997). However, littleattention has been paid to the dynamic relationshipbetween the heritage of a spin-off and the exploita-tion of its innovative potential through entry into a newmarket.

A key benefit of looking at the speciation processthrough a dyadic (parent–progeny) lens is that it illu-minates issues of both continuity and discontinuity. Thedevelopment of a lineage is continuous in that the ini-tial form of technology in the spin-off bears a closerelationship to the antecedent technology in the sourcedomain; however, the technology may develop a quite adistinct form as a result of the spin-off entry into domainswith distinctive selection criteria and resource pools, aswell as deliberate organizational efforts on part of theactors to change their opportunity space (Cattani andLevinthal, 2005). In what follows, we present a casestudy of the emergence of a new species of technol-ogy, enabled by an organizational form more adept atexploiting the new opportunity space than was its parentcompany.



3. Methods of inquiry

3.1. Research setting and design

We sought a case for which we could track key pro-cesses in a parent–progeny transition in considerabledetail, one recent enough to enable us to interview peo-ple who were directly involved in the transition. AcornComputers, founded in Cambridge England in 1978, hada unique record in spinning out high tech ventures, oneof which, ARM, became a world leader in micropro-cessor design. Acorn Computers was not a commercialsuccess, but it gave rise directly or indirectly to morethan thirty start-ups and has been compared to Sili-con Valley’s Fairchild Semiconductors (Athreye, 2004).In 1984, Acorn ran into serious difficulties and wasacquired by Olivetti, but retained its operational auton-omy. Among the reasons for Acorn’s problems had beendelays and defects in the microprocessors supplied byFerranti. To avoid recurrence of these difficulties, Acornset out to develop in-house expertise in microprocessordesign in its Advanced R&D department, which was tobe the basis for the 1990 spin-off of Advanced RISCMachines Ltd. (ARM).

A study of Acorn and ARM had been undertakenin real time, providing us with contemporary evidencefrom interviews dating back to the 1980s (Fleck andGarnsey, 1987; Garnsey and Fleck, 1988). This togetherwith contemporary documentation reduced the risk ofretrospective meaning being imposed on events fromknowledge of outcomes (Golden, 1992; Yin, 1994:92). To gain a longer-term perspective we conductedfollow-up interviews with key personnel involved in thetransition process and those responsible for developingRISC technology in ARM. Interviews included thoselisted in Table 1. Repeat interviews were conducted withthose closely involved in the transition.3

New product development is allocated resourcesaccording to selection pressures within firms (Bower,1970; Burgelman, 1994), whereas the market consti-tutes an external selection environment for the firm’sinnovations. To operationalize the concept of the firmproviding an internal selection environment, we usedinformation from respondents on how decisions aboutresource allocation to new projects were taken. To oper-ationalize the concept of external selection environment,we asked in interviews about market conditions affecting

3 Semi-structured interviews lasted between 1 and 3 h. Interviewsbefore 1995 focused on Acorn and those after 1995 focused on thespin-off and ARM’s development.

the firms’ technologies. To uncover how the competencedeveloped at Acorn was furthered at ARM, we askedinformants about ARM’s technological base and how theARM business model was devised. Extensive archivaland secondary data helped to clarify the internal as wellas environmental conditions under which the spin-offdriven speciation process unfolded. Other sources pro-vided background information on the role of Olivettiin rescuing Acorn from the severe financial crisis thatbefell the company in the mid-1980s (Ciborra, 1994;Piol, 2005) and on how ARM came to provide the lead-ing architecture in low-power consumption embeddedprocessors (Atack and van Someren, 1993).

3.2. Rationale for a dyadic case history

A longitudinal case study design across the twoorganizations was appropriate for studying speciationthrough spin-off (Yin, 1994). A dyadic approach, rare inboth spin-off and speciation literature, was essential touncover antecedents and microprocesses of innovation.A rationale for presenting the evidence as a chronol-ogy detailing antecedents and outcomes of the spin-offis provided by the cognitive scientist Jerome Bruner,who showed that narrative is uniquely suited to captur-ing the complexities of change as experienced by humanactors (Bruner, 1990). We applied constructs from thespeciation perspective to facilitate pattern recognition ofinterconnected factors and generic processes that shapedthis instance of innovation.

A single in-depth case does not aim to be represen-tative but to delineate concepts and identify connectingideas, using evidence from a detailed empirical exem-plar to inform theory. A single case has provided a newperspective on wider issues in key studies that includePenrose’s Hercules Powder case (Penrose, 1960), theinspiration for her Theory of the Growth of the Firm(1959), Burgelman’s account of strategic change at Intel(1994) and Schein’s work on Digital Equipment Cor-poration (Schein, 2003). The Acorn-ARM case helpsus delineate a framework and identify recurrent causalmechanisms in the continuities and discontinuities thatshape technological advance.

4. Summary of the case study

Acorn Computers was among the first European com-panies to design and market a microcomputer. Thespin-off from Acorn occurred through the coincidenceof pressures from within Acorn and the needs of a lead-ing US company, Apple Computers, which required amicroprocessor for a new hand-held product. Apple’s



Table 1Interviews

Interviewee Role Interviews Interview year

Hermann Hauser Co-founder, Acorn 5 1986, 1987 1988, 2004, 2005Andy Hopper Co-founder, Acorn 2 1988, 2005Chris Curry Co-founder, Acorn 1 1986Peter Wynn Acorn, Financial Director 1 1986Jeff Tansley Technology Manager, Acorn 2 1988Jim Merriman Chief Operating Officer, Acorn 1 1988Brian Long Managing Director, Acorn 3 1988Sam Wauchope Managing Director, Acorn 1 1990David Lee Managing Director, Acorn 2 1999Sam Boland CEO, Acorn 1 2001Malcolm Bird Technology Director, Acorn 3 2004Alex von Someren Acorn engineer, wrote 1993 ARM RISC Chip Programmer’s Guide 1 2004James Urquhart VLSI Design Manager, Acorn; Co-founder and Chief Operating Officer, ARM 2 2006, 2007Robin Saxby Co-founder and CEO, ARM 3 1995, 2005, 2006Ian Phillips Principal Staff Engineer, ARM 2 2006, 2007Elserino Piol Deputy Chairman, Olivetti 2 2005, 2006

managers had identified the potential of Acorn’s RISCchip technology, but did not want to license directly fromAcorn as a competing PC company. Apple Computersproposed a new joint venture. A group of engineersfrom Acorn’s microprocessor development team whohad outstanding hardware, software, system and inte-grated circuit expertise, provided the personnel for thenewly founded Advanced Risk Machines. Althoughthe expected market for the Newton notepad did noteventuate, this experience confirmed the existence ofan opportunity to license ARM’s RISC technology inrapidly expanding markets for embedded systems.

The transition from Acorn to ARM can be concep-tualised as two partly overlapping phases (Fig. 1): (1)The birth of Acorn Computers in 1979 and the sub-sequent development of Acorn’s RISC chip meetingcurrent requirements from the early 1980s. (2) The spin-off of the RISC technology into ARM Holdings in 1990and its transformation as a new techno-organizationalspecies. In what follows, evidence on the unfolding pro-cess through which this transition occurred is presentedand the evidence is then analysed as a speciation process.

5. Speciation through spin-off: the Acorn-ARMstory

5.1. 1978–1990: Acorn and the RISC technology

In 1978, after a PhD in physics at the University ofCambridge, the Austrian born Hermann Hauser joinedChris Curry, who had been working with Clive Sinclair,to launch Cambridge Processor Unit (CPU). Initially theaim was to provide computer consultancy services and so

fund the development of microcomputer kits to be soldby mail order. Acorn Computer Ltd. (a trading name usedby CPU) was able to call on skills at Cambridge Univer-sity in microcomputer design and was soon pioneeringin computer architecture, silicon chip design, operat-ing system design and local area networking (Fleck andGarnsey, 1987).

In 1981, Acorn was placed on the list of companiesinvited to tender for a BBC contract which would allowthe selected computer to carry the BBC emblem in returnfor royalties. Hauser succeeded in persuading a groupthat included Roger Wilson and Steve Furber to build aworking prototype in 4 days, in time for the BBC con-tract. He managed to convince them by “by telling eachof the team that the others thought this was possible”(Hermann Hauser, personal interview). Together theyachieved what they individually thought to be impos-sible: they came up with an innovative microcomputersystem in a very short span of time. The favourable reportof the BBC’s technical advisors and the energy of theyoung team with its strong university connections ledthe BBC group to select Acorn’s prototype. The BBCendorsement drove forward sales and allowed the R&Dmembers of the group to develop new, more powerfulmodels.

The group of people in and around Acorn Comput-ers was dedicated to technical excellence and believedin a holistic approach to computer design. Hardware,software, and architecture had to be mastered. As a con-sequence the group developed knowledge in every aspectof computer design. “The DNA at Acorn at the time,and it was a wonderful time, was dedication to technicalexcellence [. . .] it probably was one of the last times in



Fig. 1. From Acorn to ARM: the speciation process.

our history of computing that a group knew everythingabout the entire architecture of the computing environ-ment” (Hauser, personal interview). Acorn was able toattract talented software professionals to develop a rangeof software on an outsourced basis and its computerarchitecture enabled ‘add-on” peripherals from third par-ties. In April 1984, Acorn won the Queen’s Award forTechnology for the BBC Micro.

Acorn’s reputation was based on their innovativeR&D and high quality standards. New product designand development were a priority. Creativity was viewedas a strategic need, the view being that a leading companyhad to be ahead of its rivals on as many fronts as possi-ble. This was a period of excitement for the young team.“We were congratulating ourselves. We had got it right:quality, delivery—one success after another. We wereflying high. You really think you can do anything afterthat. It was: ‘Just give us a problem, we’ll solve it” (Her-mann Hauser, personal interview, 1987). In 1983 Acornwas floated on the stock market, though the foundersretained 90% of shares.

In the early 1980s the company had been devel-oping very new products that were later to becomefamiliar ideas. These included modems, telecommuni-cations products for satellite and cable broadcasting,interactive video and an alternative operating sys-tem to UNIX. But by 1983 many of these R&Dprojects were neither relevant to Acorn’s core busi-ness of microcomputers nor to identified market needs.Few developed into marketable products, and few ofthose produced sold well. Over the years, however,the propensity to experiment and innovate becamea defining feature of the company, creating compe-tences that Acorn employees used in creating spin-offcompanies.

Although microprocessors could be bought in, Acornhad experienced continual difficulties with obtainingchips of the standard they required. As scale-up tookplace, Acorn’s supplier Ferranti delivered plastic casingwhich failed to dissipate heat as the earlier prototypes in aceramic version had done. This resulted in a serious fault:the disappearance of the computer’s screen display.4 Inorder to prevent any further delays in the launch of newproducts, a group was set up in 1983 to look into chipdesign.

Acorn started to search for 16 or 32 bit micropro-cessors to replace the existing 8 bit MOS 6502 basedsystem. According to Acorn’s chief hardware designerChris Turner “16-bit and 32-bit microprocessors werebeginning to appear and the Lisa had been launchedwhich made us realise that we needed a windowssystem, so some of my more academically inclinedcolleagues, like Steve Furber, argued that we neededa new architecture”. Yet the existing 16- and 32-bitarchitectures did not fit Acorn’s vision. “We lookedat National’s 16032 and Motorola 680000”, recalledSteve Furber, “but they did not suit our design style.They had very complicated instruction sets givingpoor interrupt response. Basically they were too slow.”(Electronic Weekly, 19 April 1998).

As Andy Hopper explained, “Starting from scratchto design a 16-bit microprocessor would simply appearinsane to most of the people in the microprocessor busi-ness” at that time (personal interview). But a formercolleague in the US pointed them to the Berkeley RISC

4 Interview with Prof. Andy Hopper 11 April 2005. “In their earlyexperiments to introduce simplicity into instructions, Acorn developerscut the safety margin too thin, but over the next few years were able toget the new designs right.”



design.5 Using a reduced instruction set of this kindcould bypass many of the problems involved in standardchip design. In 1983 Steve Furber and Sophie Wilsonwent to the Western Design Centre in Phoenix where theywere working on a 16-bit version of the 6502. “We foundit to be a cottage industry working in a bungalow in a backstreet. That gave us confidence. Sophie started playingwith instruction set designs. Our mentality was: ‘let’shave a go at building a microprocessor”’ (The Guardian,08 March 2001). Hermann Hauser decided to back theproject. As he recalls “We decided to do a microproces-sor on our own so I gave my team the only two things wehad: no money and no people! The only could do it bykeeping it really simple” (personal interview). Furber setto work defining the architecture while Wilson developedthe instruction set.

Work started in 1983 and the first silicon was runon 26 April 1985. It had taken 18 months and five manyears. The Acorn RISC Machine (ARM) was designedby the same group of people who worked on the firstcomputer kit and the pattern was the same; it workedsufficiently well the first time to be debugged using thesystem itself: “The ARM chip performed well, given thatits power was limited because of a low number of tran-sistors. But this did mean it had extremely low-powerconsumption that could be adjusted up or down, and wasof very small size” (Sapsed, 1999, p. 73). The chip had30,000 transistors, the same number as a Z80 or the 6502that Acorn used in its BBC Micros, but it was twentytimes faster. It was also the world’s first RISC proces-sor. According to Hauser, “ARM was part of a policydecision that a computer should be designed on siliconrather than cobbled together out of third party compo-nents. This focus made Acorn one of a small, selectgroup of genuine computer companies (Apple being themost obvious) that own their own technology from theground” (personal interview). ARM’s first sample wasdelivered in 1985 and was manufactured by VLSI whichlater was also licensed to sell chips using Acorn’s newdesign.

The ARM microprocessor emerged in a period whenAcorn was experiencing financial difficulties. Howeverafter a 5-year period of expansion the founders were notalarmed by mutterings about retailers’ sales; they weresure these would pick up as Christmas approached. ByJanuary 1985, one third of the Christmas stock remainedunsold; sales were about 35% below the April fore-cast and the pressures on Acorn made it necessary to

5 The development of RISC in the US is documented by Khazamand Mowery (1994).

cut staff numbers from 450 to 250 by early in 1985.Acorn’s share values slumped. Suppliers’ demandsfor payment became pressing and by March 1985 awinding up petition was issued. Acorn had to ceasetrading.

No British company appeared willing to enter intoan arrangement acceptable to Acorn. Elserino Piol, aDirector at Olivetti, made an approach which was toresult in the Italian electronics company taking a 49%share in Acorn. By September 1985, Olivetti took own-ership of 79% of Acorn after sales’ targets were missed.Olivetti paid under £15 m for Acorn which had beenvalued at £100 m in the previous year. “In accordancewith Olivetti’s partnering policies, Acorn had originallybeen acquired to gain market share in the UK and astrong foothold in the education market. However, itcame as a surprise to Olivetti that Acorn’s labs con-tained a wealth of people, skills and on-going projectsthat turned out to be of strategic relevance” (Ciborra,1994).

The challenge was to make the transition fromAcorn as an independent company to Acorn as a sub-sidiary which could benefit from Olivetti’s internationalaccess and resources. But like most large organiza-tions, Olivetti was not a unified entity. In the UKbranch of Olivetti there was support for Acorn’s newdevelopments. In contrast, the more cautious opera-tional groups at Olivetti favoured standard, well-proventechnology. There were no effective processes for inte-grating Acorn’s many innovative projects into Olivetti’soperating branches. This caused friction and con-flict.

In the late 1980s, the possibility of spinning out themicroprocessor unit from ARM was under discussion,since the prospects for a RISC chip unit inside AcornComputers were poor. There was reluctance among man-ufacturers to source from a competitor company or onethat might terminate the license to protect their ownmarkets. Hermann Hauser and Elserino Piol, of Olivetti,were of the same mind as engineers engaged in advancedR&D at Acorn; they viewed a spin out of the micropro-cessor unit as desirable. This did not prove practicableuntil Apple Computers needed a microprocessor for thehandheld device, known as the Newton Notepad, that itwas developing. Apple Computers had identified Acornas a potential partner because they were impressed by theARM chip. But they did not want to depend on a com-petitor for the microprocessor for the Newton. Morever,Acorn, as a PC company, had little incentive to providewider support for their RISC technology. A joint venturewas proposed between Apple Computers and Acorn todevelop a microprocessor for the Newton notepad. VLSI



took part in the joint venture, with Acorn, Olivetti andApple Computers.6 Pressures at Acorn had left the mar-ket potential of the RISC microprocessor unrealised. Butthe new venture now had the freedom to leverage theextensive expertise it had inherited from Acorn. As man-aging director Sam Wauchope explained “It is a bit ofa wrench to separate what has been an integral part ofAcorn, but we have decided that ARM and Acorn are bestserved by the creation of a separate company. The dealopens up many possibilities in terms of product devel-opment which we probably would not have been ableto afford” (Electronic Engineering Times, 14 February2000).

5.2. Post 1990: inheritance and spin-off drivenspeciation

The recruitment of Robin Saxby as managing directorof the new company, was favoured by Hermann Hauser,a member of Acorn’s Board who had known him atMotorola. The goal of the new CEO was to “addressand attack the growing market for low-cost, low-power,high-performance 32-bit RISC chips” according to inter-nal documents. On the basis of his experience in theindustry, Saxby immediately sought to make the most ofthe company’s limited resources. Saxby had a degree inelectronic engineering and had spent 11 years as market-ing and engineering manager at Motorola. He a few yearsat a smaller company, European Silicon Structure (ES2)before taking over as CEO at ARM. He chose to pro-mote three of the 12 Acorn engineers who had left to formARM to positions of management in sales, marketing andengineering; they later became directors. Saxby believedit was quicker and easier to teach engineers to sell than toteach salesmen to understand ARM’s complex technol-ogy. ARM also shared the services of Acorn’s companysecretary and ES2’s financial controller to perform itsinitial licensing and financial work on a part time basis.In November 1991, ARM released ARM6, the first fam-ily of products after the spin-out. VLSI manufactured theprocessor, which was customized to users’ requirements.As recalled by Hauser: “By the time we spun out ARMthere was a very well established relationship with VLSI

6 Acorn’s original expertise and development team was valued at£1.5 million, Apple invested the same amount in cash and each com-pany took 30% of the company’s total shares. VLSI invested £250,000cash for a 5% holding, with the remainder of the shares set aside forfuture investors, including ARM’s employees. The majority of shareswere initially held by the partners’ shareholdings (Acorn and Applewith 46% and VLSI with 8%). Shares were set aside for further strategicalliances.

Technology, who manufactured the ARM. At that timewe also convinced Apple to use the ARM for the New-ton, so there were two built in customers for ARM whenwe spun it out (the other was Acorn). They made thechips and they sold them to Apple and to us” (personalinterview).

The expertise accumulated by the ARM team atAcorn, together with Saxby’s vision of defining theglobal standard for microprocessors for mobile devices,set ARM on a winning course at an early stage. “Saxby’sapproach of focusing on maximising the availableresources to deliver the technology and win business,rather than spending time raising finance, proved to beeffective as the small and fast moving ARM becameself sustaining by only its second year of operation,without the need for further financing” (ARM StrategicReport, 1999). Increasing numbers of customers neededlow-power consumption microprocessors which offeredsmall size, lower cost and lower power consumption,performance factors which were less relevant in the PCsector.

At this early stage, Saxby was confronted with a vari-ety of possibilities. One option was to merge the businesswith a semiconductor company and lead a new divisionthat would have financial resources. Another option wasto create a semiconductor company that would designand market chips but subcontract manufacturing. Alter-natively ARM could have partnered with Apple to driveall future product development. The option that wasadopted was to design a base technology and then licensethe intellectual property (ARM Annual Report, 2005).From Acorn, ARM had inherited a model based not onfabrication but on the contracting out of chips’ manu-facture. Saxby took the ‘fabless model’ a step furtherand decided that ARM would be a “chipless, chip com-pany,” their technology enabling microprocessor designby customers.

Saxby decided to avoid the costs incurred in sales andmarketing. He reasoned that partnering with semicon-ductor companies was a better option, since they wouldbear those costs and deliver chips based on ARM cores.In 1992 GPS became the first independent licensee tomanufacture ARM chips, followed by Apple when theNewton was equipped with an ARM6 processor. In 1993Sharp became the third licensee to manufacture and mar-ket ARM processors. As experience increased, ARMoffered further services to its clients such us consultancy,feasibility studies, training and prototypes supply.

The learning and experience that ARM was able togain through these partnerships proved to be significant.When Texas Instruments (TI) took out a license in 1993,its aim was to win the mobile phone business of a rel-



atively unknown Finnish company, Nokia. Though TIhad considerable expertise in digital signal processing(DSP), it was relatively weak in the complementary tech-nology for central processing units (CPUs). What theyneeded was a CPU design that would work reliably in thebackground, use minimum power, and be well supportedwith design tools, models and applications. ARM’s CPUtechnology promised to provide that platform, so TIdecided to package the ARM design into its overall offerto Nokia.

TI, ARM and Nokia formed a consortium to seek anew solution. This provided an opportunity for ARMto gain a detailed understanding of Nokia’s prioritiesand needs. They found that memory size and the vol-ume of code were viewed by end users as the criticalimpediments. This gave the company a clear signal:the key to the business of end users was memoryreduction. ARM worked with TI to set and meet ambi-tious targets for power consumption and code size.The outcome was an innovation that became knownas the “Thumb” architectural extension which made itpossible to use ARM’s processor as a programmabletool for developing complex System on Chips (SoCs).Though the processor was not itself a differentiatingfeature of their product, customers could use the ARMchip to differentiate their product by developing com-plex SoCs. This experience taught ARM the value ofworking with partners and listening to their feedbackon its products. The company also realised that end-customers could provide information crucial to productdevelopment. The Thumb experience reinforced ARM’sapproach of working closely with market-leading OEMsas well as their direct licensees – the semiconduc-tor companies like Texas Instruments – leading torapid introduction of new product families and paral-lel entry into new markets (O’Keeffe and Williamson,2002).

As a result of this new approach, ARM’s emergingcapabilities enabled it to tap into extensive informationand competencies through numerous partnerships withclients. As one of ARM founders explained to us: “ARM,when it’s operating properly, should be able to assimilateinformation about what’s happening in the market muchmore quickly than most companies because bizarrely, itwill work with hundreds of competitors [. . .] not com-petitors to ARM, but competitors to each other and withtheir customers so (ARM) has an incredible intelligencemachine.” This intelligence gave ARM crucial indica-tions on the unfolding technological roadmap, helpingARM to foresee new product generations, decide whenand whether to shift to new areas, and solve differentuser problems.

Thus ARM was able to develop products incorporat-ing additional features and instruction set enhancementsappropriate to application needs. At the same time ARMmaintained a common, general purpose RISC instruc-tion set, which provided code compatibility (ARMAnnual Report, 2006). Architectural extensions wereintroduced in subsequent versions of the ARM architec-ture, building on the previous architectures, thus addingbackwards code compatibility of new processor coreswith older generations.7 As a result of this process,RISC chips that were designed after ARM was spun-out (ARM6, onwards) came to differ substantially fromthose that had been designed at Acorn, though there wasunderlying continuity. (The core remained largely thesame size throughout these changes. ARM2 had 30,000transistors, while the more powerful ARM6 increasedtransistors marginally, to 35,000.) ARM co-founder,Jamie Urquhart, who was VSLI Design Manager whenARM was formed, explained that the main changes inthe RISC chip involved altering the processor from a 26-bit address bus to a 32-bit address bus and convertingfrom a dynamic design to a static design. The formerchange made it easier to program for large applications,particularly when an operating system was used (Appleand Nokia both needed this change). The second changemade the design a little more complex and slightly largerbut easier to use. The dynamic version required that theclock running the processor never stopped to preventthe processor ‘forgetting’ its state. Making the proces-sor static meant this restriction was removed and so madeit easier to design the processor into complex ‘System onChips’ (SoCs). Acorn’s chip was designed as a compo-nent for their PC product. ARM in contrast offered a setof highly customized processors and supporting macro-cells suitable for the use in a wide range of applications.Table 2 provides a summary of differences in the RISCtechnology.

The technology inherited by ARM from Acorn’scomponent chip had the great benefit of requiringvery low-power consumption. The differences intro-duced at ARM reflected new technical and product

7 ARM first development was the next step from the ARM3 pro-cessor, which was named ARM6 and included full 32-bit addressingembeddedness support, one of many changes originally requested byApple in order to use the ARM in planned products. An improvedvideo controller, VIDC20, was also developed and a floating pointprocessor was also introduced. This was followed by a tight sequenceof product families matching new segments and applications includ-ing the ARM7 family, ARM9 family, ARM9E-S family, SecurCorefamily, StrongARM family, ARM10 family, ARM11 family, ARMCortex family, Intel-based products, Marvell Feroceon and QualcommScorpion processor cores.



Table 2Key technical changes between ARM 1/ARM3 at Acorn and ARM 6onwards at ARM Ltd.

Acorn ARM Ltd.

Address bus 26 bit 32 bitDesign Static DynamicInstruction execution 32 bit 16 bit (Thumb)a

a Thumb takes 32-bit instructions and compresses them to 16-bits.This innovation enables programs to be coded much more densely thanstandard RISC instruction sets (the code size is reduced by 30% at nocost to performance) and to cut some portions of the hardware consider-ably in size. Processors making it possible to take advantage of Thumbalso allow 32-bit instructions to run on the same processor. 16-bit and32-bit instructions can be mixed together and the hardware will be ableto decode and decompress at the same time without a performance hit,thus maintaining powerful computing capabilities. Cost is minimizedby having a simple, small structure with many configurations available.

requirements and the recognition of the need for IPprotection. The changes required to suit the new appli-cation domain facilitated programming and designingsystems on chips. A simultaneous change in ARM’sbusiness model had a major impact on the industry.ARM was no longer producing a chip for personalcomputers, but offering customers the capability todesign customized, low-power consumption chips forhighly integrated applications such as cell phones, per-sonal digital assistants, information appliances and otherembedded systems.

In 1993 ARM had 50 employees and £10 millionin revenues; 3.5 million chips had been produced withARM technology. At the end of the same year ARMlaunched ARM7 in California. At one 16th the size of theIntel 486, the reduced size enabled a price cut to under$50. More significantly, the reduced power consump-tion meant it could be run off a much smaller battery,strengthening its position as the leading microprocessorfor embedded applications. ARM’s growth continued tosoar during the second half of the 1990s, along withthe increase in licensing deals. At the end of 2001,with ARM chips in an estimated 77% of the embed-ded RISC processor market, they had become the defacto global standard. By 31 December 2006, ARM had£263.3 million in revenues, 1659 employees and ARM’stechnology has been licensed to 188 semiconductor com-panies, including many of the leading semiconductorcompanies worldwide.

The intellectual property inherited by ARM fromAcorn, the 32-bit Acorn RISC Machine, had languishedin VLSI Technology’s data book as a functional block forASIC design for several years without evoking interest inthe industry. The RISC chip, developed as process inno-vation at Acorn for their PC product, mutated into a new

platform technology, providing a software core on thebasis of which customers could customize chip designand apply it to a variety of high volume applicationsin the wireless, consumer electronics and networkingmarkets. This process was supported by the develop-ment of a highly distinctive IP centred business model.“The spin-off ejected ARM into a whole new world ofopportunities” (Saxby, personal interview).

Acorn Computers was disbanded in 1997 and part ofthe company moved into Element 14 (soon acquired byBroadcom), partly in order to realize the value of theshares in the ARM joint venture. This had an impactfurther afield. Apple Computers, whose share price wasat a low point in 1997, was able to mobilize resources torenew its capabilities from the sale of Apple’s shares inthe ARM joint venture, as Hermann Hauser pointed out(personal interview).

In what follows we present our interpretation of thecase and propose the concept of techno-organizationalspeciation to depict this instance of technologicalbranching enabled by the emergence of a new businessmodel.

6. Interpretation of the case evidence

The Acorn-ARM transition is made up of a varietyof strands that include a major change in applicationdomain, a new platform technology, the demise of theparent company and a new business model. The dimen-sions of speciation identified in evolutionary theory canencompass this diverse evidence in a coherent man-ner. It is always a challenge to assign actual instancesto conceptual categories. It is to deal with this issuethat different criteria for distinguishing between specieshave been proposed, including (1) inter-species incom-patibilities, (2) the selection environment of niches orecosystems, and (3) historical lineage. Dosi wrote that“The . . . idea of a technological paradigm should betaken as an approximation, adequate in some cases butless so in others. . .” (Dosi, 1984, p. 8). The same couldbe said of efforts to categorize ‘families’ of technologi-cal products and to deal with subsystems of technologyif there are ‘paradigm changes’ at some levels but not atothers in innovative architectures. Dosi points out thatthese challenges do not invalidate the use of conceptualdistinctions that clarify previously overlooked connec-tions between phenomena. Dosi’s rationale applies toidentifying a new species of technology.

Unlike the blind determinism of natural selection,processes of technology selection involve learning anddeliberate efforts by purposive actors to shape theiropportunity space. Nevertheless, concepts developed



to address the categorization problems of assigninginstances to species in the natural world also provideconceptual grounding for the idea of technological spe-ciation. As we go on to show, a new business design cancontribute to this process by providing distinctive techni-cal and organizational features. We refer to this processas techno-organizational speciation.

6.1. Technological incompatibilities and standards

Both biological and technological systems are basedon interaction, and consequently incompatibilities canoperate to eliminate an idiosyncratic competitor. As anew technology is altered to the point of incompatibil-ity with the originator technology, this brings to an endthe design replication process that keeps a technologyspecies extant (Rohlfs, 2001). Acorn’s technology hadbeen incompatible with what had become the dominantstandard in Acorn’s product market, the IBM PC andMicrosoft’s operating system. Bill Gates had offered tolicense MS-DOS to Acorn on favourable terms in the1980s but had Acorn adopted the Intel chip that sup-ported MS-DOS, this would have deprived Acorn ofcritical experience from the use of their own RISC chipin their products, experience which was to prove invalu-able to ARM and attract Apple Computers to the jointventure.8 But the incompatibility of Acorn’s productswith the IBM’s PC and Microsoft’s DOS operating sys-tem led to unfavourable network effects as these tookover as the industry standard, eroding Acorn’s mar-ket share (Garnsey et al., 2006). By the time Acornwas disbanded, ARM’s inherited RISC technology hadbranched away and no longer supported personal com-puter products.

Hauser later acknowledged that youth and inex-perience had limited understanding of the businessenvironment faced by Acorn and recognition of the ben-efits of open innovation for hardware as well as software:“We thought our wonderful proprietary technology gaveus an advantage over competitors, so we bagged it.”That is, they refused to license out Acorn’s technol-ogy. “We should have really analysed how we comparedto competing computers around the world. This wouldhave shown us we had a real lead in terms of speed,price, operating system and expansion slots compared to

8 According to Hermann Hauser (personal communication): “TheBill Gates story is always misquoted. Acorn would have declined muchfaster had we adopted MS-DOS because this would have meant us touse an Intel microprocessor. It was our use of the ARM that allowedAcorn to survive for 10 years against Intel and Microsoft, because wehad better price/performance.”

our nearest rivals Apple II and Sinclair Spectrum . . . Weshould have gone around the world persuading people toadopt Acorn’s products as the industry standard. Then weshould have licensed our hardware and software to any-one that wanted them” (The Guardian, 2001). Whetheror not such an open strategy would have succeeded forAcorn, it became the business model adopted by ARM.The RISC chip design technology licensed by ARM wasnot for PCs. Instead it offered compatibility with a rangeof embedded systems as a new technology standard withprospects for much wider diffusion into multiple marketsthan could be achieved by Acorn’s proprietary technol-ogy.

6.2. Speciation in new selection environmentsthrough niche construction

The second perspective on speciation focuses onnew species that arise in new niches and ecosystems.ARM succeeded in no small part through niche con-struction involving interaction with users (Schot andGeels, 2007). How ARM positioned itself within newbusiness niches was revealed by the account of col-laboration with Texas Instruments and Nokia. Closeinteraction with co-producers became as a central featureof ARM’s business model. So extensive were the nichescreated in different sectors through ARM’s variouspartnership arrangements (mobile devices, multi-media,automotive, etc.) that ARM gave rise to what has beendescribed as a business ecosystem of users and sup-pliers (Moore, 1993, 1996), based on interactive andcompatible technologies. This was a result of intendedeffort as well as exogenous constraints. ARM’s latestprocessor had been destined for the Newton Notepaddeveloped by Apple, but this product was not adoptedby consumers. “We have to find a way to turn ourenemies into friends” Robin Saxby had declared. Inachieving this conversion, managers at ARM positionedthe venture to grow in expanding markets for embed-ded devices, a much more munificent environment thanthe intensely competitive market for PCs and hand heldcomputers.

To increase the chances of favourable selection, ARMgarnered insights and knowledge through partnershipsand licensing agreements worldwide. Unlike the usualbilateral relationships with lead clients, ARM developedclose relationships not only with the big chip manufactur-ers, but also with their partners’ customers. Learning ofthis kind engendered a structure of coordination practices(Kogut, 2000) that translated into effective knowledgesharing (Dyer and Singh, 1998). The role of teacherand student shaped ARM’s partnering relationships and



thereby created barriers to potential competitors (Arinoet al., 2001). This was how ARM moved from construct-ing favourable niches to creating a business ecosystemfor the users of their microprocessing design tools, aidedby the massive growth in demand for embedded devices,especially mobile phones. Companies like Nokia wereable to use ARM’s processor to develop complex systemson chips run on the client company’s chosen soft-ware. “ARM created an ecosystem for companies thatneeded chips for embedded devices” Urquhart explained(personal interview). In contrast with Acorn, whoseincreasingly marginal technological design lost groundin the PC business ecosystem, the global standard cre-ated by ARM in microprocessor design for embeddeddevices grew to be a dominant standard and compatibil-ity with a wide range of applications promoted its furtherexpansion.

6.2. Historical lineage and techno-organizationalspeciation

Historical lineages and their continuities have beenthe focus for a third approach to speciation. Centralto evolutionary processes are interactors that repli-cate the design they embody, or alter it by chance orintention (Beinhocker, 2006). Through the reproductionof designs, lineages of replicable knowledge developamong technologies and among organizations and arediscontinued if a knowledge base falls into disuse.9

In the case history examined here, there was criticalcontinuity in the inherited technological design but acomplete break in business model design between parentand progeny. Technological differences were at first rel-atively minor as between the ARM chip used by Acornand that launched by ARM, initially involving the pro-cessors’ bus address design and clock. But these and thefurther changes undertaken were crucial for adoption innew markets.

ARM’s new business model was of particular impor-tance because prior to ARM’s success it was doubtedwhether licensing a technology could bring in sufficientrevenues to support a new company. ARM’s businessmodel involved the company in co-production activi-ties with customers and the ability to maintain relationalcontinuity by building and sharing a development roadmap: “It’s not just a licensing business, it’s a roadmap. It’s investing in a future. And that comes back to

9 Organizations can be viewed as interactors that carry and acti-vate knowledge embodied in specific technologies while organisms areinteractors (phenotypes) that carry and express genetic information.

the shared destiny point. Again, it is partnership . . . aslong as ARM can keep evolving and in a way bringvalue to its customers, it earns its roadmap and so ithas value” (Urquhart, personal interview). A businessmodel favouring a distinctive pattern of alliances withlead clients (von Hippel, 1986) has significant strategicadvantages for resource-constrained new ventures in net-worked industries which must rapidly reach a thresholdin scale and market share to avoid unfavourable networkeffects (Lorenzoni and Baden-Fuller, 1995).

Certainly there was lineage continuity between thetwo companies. Indeed ARM was favoured with auniquely valuable set of endowments from Acorn: (1) acore technology that was proven and protectable, remov-ing the need for speculative, time consuming R&D; (2) ageneric technology in the form of platform microproces-sor; (3) team continuity in the 12 engineers from Acornwho founded ARM10; (4) immediate custom and rev-enues from Acorn Computers. Nevertheless, in supportof the idea that practices in the company of origin canexert an unfavourable influence on the spin-off, ARMdeliberately broke with the way many things had beendone at Acorn. Here ARM benefited from the input ofmanagers recruited for experience and skills achievedelsewhere.

These outsiders abandoned Acorn’s practice of avoid-ing the expense of patenting. Part of the strategy as thecompany moved forward was to develop a roadmap thatincluded patentable IP and programs were put in place toreward engineers for the creation of relevant, patentableideas. Acorn’s aim to turn a proprietary standard intoan industry standard without licensing or partnershipswas abandoned. ARM’s new managers fostered a muchmore commercial culture while building on Acorn’scommitment to technological excellence. Thus a com-ponent technology from an unsuccessful product wasinherited from the parent company and transformedinto a new technological species, enabled by distinctiveorganizational arrangements and new forms of businesspartnering. This change was not the result of a revolu-tionary upheaval in technology but emerged graduallyfrom its antecedents and as a result of different selec-tion pressures operating in new niche domains (Basalla,1988).

10 Virata, a successful spin-off from another spin-off from Acorn,Olivetti Research, also met the conditions of having a proven, pro-tected and generic technology applied in a high growth market (forInternet connection), a different domain from that originally envis-aged. Licensing models without such positive attributes are less likelyto succeed.



7. Conclusions

The concept of speciation has helped us to interpretthe case evidence on innovation and to identify strongand recurrent causal currents amid the eddies of detail.Three perspectives on speciation derived from evolution-ary theory were found to illuminate case evidence. Thestudy focussed (1) on boundary constraints on techno-logical replication; (2) on new variants arising in newniches and ecosystems, and (3) on common lineages ofdesign and replication. Empirical evidence thereby inter-preted provides theoretical grounding for the concept oftechno-organizational speciation. This requires analysisat the level of the organization’s business model. Theimportance of business model innovations is increas-ingly recognised (Chesbrough and Rosenbloom, 2002;Beinhocker, 2006). But the concept of business modelhas not yet been assimilated as a construct in evolution-ary theory which still depends on the notion of routines(Nelson and Winter, 1982). A business model amountsto more than a collection of routines. A business modelcan be thought of as a design that specifies how a firm isconnected to others in its ecosystem in order to create andcapture value. A business model is a design which can beoperationalized in practice and which may be wroughtas a response to experience, as occurred at ARM, or bedeveloped ex-ante. The links between business modeland co-evolutionary processes are apparent when thefirm’s business model is conceptualised as a design forcreating value for customers and capturing value fromits economic activities (Lim et al., 2006). The businessmodel, endorsed by the firm’s Board and operationalizedby its management, guides the allocation of resourcesand specifies the ensemble of routines needed for valuecreation for customers and value capture for the firmand its investors. Sustained value creation and capture(Bowman and Ambrosini, 2000) are the goals of busi-ness organizations. ARM’s business model secured valuefor customers through co-production and other formsof collaboration with suppliers and resource providers.The firm’s strategic positioning in the fragmented semi-conductor value chain gave it a system integration role,as an IP and system support provider, that enhancedvalue capture for the firm and its investors. Techno-organizational speciation occurs when a firm reordersthe way in which it creates and captures value, that is,alters its business model, in launching a new type oftechnology.

The key reason for the branching and diversifyingof the technology revealed here was the branching anddiversity of the firms that were the agents of change.This idea is intuitive when firms are seen as incuba-

tors and carriers of technologies (Rosenberg, 1994) asthey move technologies into new domains of applicationthrough entrepreneurial spin-off. But the idea is not obvi-ous when speciation studies focus on incumbent firms.Spin-off studies, on the other hand, have not addressedthe issue directly. These two strands of research, on spe-ciation in incumbent firms and on spin-off, when broughttogether, illuminate the speciation process and show howit involves continuity alongside discontinuity.

Discontinuity is seen when a new company finds itsown direction, informed by knowledge and experiencefrom elsewhere, free to adapt to the new environmentand explore market prospects (Chesbrough, 2003). Thisinvolves abandoning inappropriate practices in the com-pany of origin (de Holan and Phillips, 2004; Nystromand Starbuck, 1984). On the other hand, continuity helpsto overcome the liabilities of newness, as identified byStinchcombe (1965). This has implications for tech-nology strategy. While advice on corporate venturingrecommends launching emerging and unproven tech-nologies in a separate venture (Christensen, 1997), ourevidence demonstrates a different set of benefits fromspinning out a proven technology, nurtured within a par-ent company, so that the venture can launch its variantof the technology in new domains without the risks andcosts of speculative, pre-competitive R&D. The liabili-ties of newness can be greatly reduced when a proventechnology is applied in a new domain (Garud andNayyar, 1994). This is a promising theme for furtherinquiry.

The Acorn-ARM story sheds light on the moregeneral problem of how to preserve and safeguardknowledge with high technological and wealth creationpotential, the value of which cannot be fully antici-pated ex-ante (Garud and Nayyar, 1994). By the timethe ARM spin-off took place, Acorn had already takena downward path which would eventually lead it to itsdissolution. Without the successful spin-off of ARM, avaluable legacy of know-how that generated wealth andtechnological progress might have been lost. The roleof progeny in preserving important knowledge threat-ened by the closure of parent firms is an area for furtherinquiry that has important policy implications. As notedby Hoetker and Agarwal (2007, p. 462) “To the degreethat knowledge languishes after the demise of the inno-vating firm, other industry participants and society atlarge lose the benefit of a potential source of techno-logical progress”. By the same token, an apparentlyunsuccessful company may make an important contri-bution to the economy through its progeny. What isnoteworthy in this case was the way competencies devel-oped at Acorn were preserved and enhanced even after



its demise. This raises a new research agenda aroundfirm reproduction and the preservation and transfer ofknowledge after firm closure.

We have found that consequences and mechanismsof speciation include the following. (1) Techno-organizational speciation has lasting consequences whenit launches a technology that becomes a dominantstandard compatible with multiple applications. (2)This can be achieved through multiple niche cre-ation, from which develops a new business ecosystem.(3) There is a complex interplay between inheritedknowledge and new learning, manifest in a firm’s tech-nological and organizational lineage. By such routes,inter-generational transmission among firms is a crucialsource of the branching and diversifying of tech-nologies that shape industrial structure and economicchange.

Acknowledgments

We are grateful to all the participants in the eventsdescribed here for generously contributing their time andknowledge. We acknowledge financial assistance fromthe UK EPSRC-ESRC-AIM IPGC (Innovation and Pro-ductivity Grand Challenge), from the EU-Marie CurieIntra European Fellowship Programme (grant no. MEIF-CT-2005-011465), and from the FIRB funding scheme ofthe Italian Ministry of University and Research (MIUR).We also acknowledge helpful comments from Gino Cat-tani, Vincent Managematin and participants of the 21stEGOS Colloquium 30 June–2 July 2005 Berlin. Anyerrors are our responsibility.

References

Adner, R., Levinthal, D.A., 2002. The emergence of emerging tech-nologies. California Management Review 45 (1), 50–66.

Agarwal, R., Echambadi, R., Franco, A., Sarkar, M.B., 2004.Knowledge transfer through inheritance, spin-off generation,development and survival. Academy of Management Journal 47(4), 501–522.

Arino, A., De La Torre, J., Ring, P.S., 2001. Relational quality: man-aging trust in corporate alliances. California Management Review44, 109–131.

ARM, 1999. Strategic Report. ARM.ARM, 2005. Annual Report. ARM.ARM, 2006. Annual Report. ARM.Arthur, W.B., 1990. Positive feedbacks in the economy. Scientific

American 80, 92–99.Atack, A., van Someren, A., 1993. The ARM RISC Chip—A Program-

mer’s Guide. Addison-Wesley, Wokingham, UK.Athreye, S., 2004. Agglomeration and growth: a study of the Cam-

bridge hi-tech cluster. In: Bresnahan, T., Gambardella, A. (Eds.),Building High Tech Clusters: Silicon Valley and beyond. Cam-bridge University Press.

Ayala, F.J., Fitch, W.M., 1997. Genetics and the origin of species: anintroduction. Proceedings of the National Academy of Science 94,7691–7697.

Basalla, G., 1988. The Evolution of Technology. Cambridge UniversityPress, New York.

Beinhocker, E., 2006. The Origin of Wealth, Evolution, Complexityand the Radical Remaking of Economics. Harvard Business SchoolPress.

Bower, J.L., 1970. Managing the Resource Allocation Process: A Studyof Corporate Planning and Investment. Harvard Business SchoolPress, Boston, MA.

Bowman, C., Ambrosini, V., 2000. Value creation versus value capture:towards a coherent definition of value in strategy. British Journalof Management 11, 1–15.

Bruner, J., 1990. Acts of Meaning. Harvard University Press, Cam-bridge, Mass.

Burgelman, R.A., 1994. Fading memories: a process theory of strategicbusiness exit in dynamic environments. Administrative ScienceQuarterly 39 (1), 24–56.

Cattani, G., 2006. Technological pre-adaptation, speciation, andemergence of new technologies: how Corning invented and devel-oped fiber optics. Industrial and Corporate Change 15 (2), 285–318.

Cattani, G., Levinthal, D., 2005. Reconciling phyletic gradualism withpunctuated equilibrium: an integrated view of evolutionary change.In: Proceedings of the Academy of Management Conference, Hon-olulu, Hawaii, 5–10 August.

Chesbrough, H., 2003. The governance and performance of Xerox’stechnology spin-off companies. Research Policy 32 (3), 403–421.

Chesbrough, H., Rosenbloom, R.S., 2002. The role of the businessmodel in capturing value from innovation. Industrial and CorporateChange 11 (3), 529–555.

Christensen, C.M., 1997. The Innovator’s Dilemma: When New Tech-nologies Cause Great Firms to Fail Boston. Harvard BusinessSchool Press.

Ciborra, C., 1994. A Platform for surprises: the organization of globaltechnology strategy at Olivetti. In: Baskerville, R., Ngwenyama,O., Smithson, S., DeGross, J. (Eds.), Transforming Organiza-tions with Information Technology. North-Holland, Amsterdam,pp. 97–111.

Claridge, M., Dawah, H., Wilson, R. (Eds.), 1997. Species: The Unitsof Biodiversity. Chapman and Hall, London.

Cooper, A.C., Gimeno-Gascon, F.J., 1992. Entrepreneurs, processes offounding, and new-firm performance. In: Sexton, D.L., Kasarda,J.D. (Eds.), The State of Art of Entrepreneurship. Pws-Kent,Boston, MA.

Darwin, F., 1887. The Life and Letters of Charles Darwin, vol. 2. JohnMurray, London, p. 88, including an autobiographical chapter.

de Holan, M.P., Phillips, N., 2004. Remembrance of things past? Thedynamics of organizational forgetting. Management Science 50(11), 1603–1613.

Dosi, G., 1984. Technological paradigms and technological trajecto-ries. In: Freeman, C. (Ed.), Long Waves in the Word Economy.Francis Pinter, London.

Dupre, J., 2001. In defense of classification. Studies in the History andPhilosophy of Biology and the Biomedical Sciences 32, 203–219.

Dyer, J.H., Singh, H., 1998. The relational view: cooperative strat-egy and sources of competitive interorganizational advantage.Academy of Management Review 23 (4), 660–679.

Electronic Engineering Times, 2000. ARM RISC leads the 32-bitEmbedded Market, Electronic Engineering Times, February 14.

Electronic Weekly, April 29, 1998.



Fleck, V., Garnsey, E., 1987. Managing growth at Acorn Computers.Journal of General Management 13 (3), 4–23.

Frenken, K., Nuvolari, A., 2004. The early development of the steamengine: an evolutionary interpretation using complexity theory.Industrial and Corporate Change 13 (2), 419–450.

Garnsey, E., Fleck, V., 1988. Acorn Computers and technology pol-icy. International Journal of Technology Management 2 (3–4),554–566.

Garnsey, E.W., Heffernan, P., Ford, S., 2006. In: Garnsey, E.W.,McGlade, J. (Eds.), Diversity and Uniformity in the Evolution ofEarly Information and Communication Technologies, Complex-ity and Co-evolution: Continuity and Change in Socio-economicSystems. Edward Elgar, Cheltenham.

Garud, R., Nayyar, P.R., 1994. Transformative capacity: continualstructuring by inter-temporal technology transfer. Strategic Man-agement Journal 15 (5), 365–385.

Golden, B., 1992. The past is the past—or is it? The use of retrospectiveaccounts as indicators of past strategy. Academy of ManagementJournal 35, 848–860.

Hoetker, G., Agarwal, R., 2007. Death hurts, but it isn’t fatal: thepost-exit diffusion of knowledge created by innovative companies.Academy of Management Journal 50 (2), 446–467.

Khazam, J., Mowery, D., 1994. The commercialization of RISC: strate-gies for the creation of dominant designs. Research Policy 23,89–102.

Klepper, S., 2001. Employee startups in high tech industries. Industrialand Corporate Change 10, 639–674.

Klepper, S., Sleeper, S., 2005. Entry by spinoffs. Management Science51 (8), 1291–1306.

Kogut, B., 2000. The network as knowledge: generative rules and theemergence of structure. Strategic Management Journal 21 (SpecialIssue on Strategic Networks), 405–425.

Levinthal, D.A., 1998. The slow pace of rapid technological change:gradualism and punctuation in technological change. Industrial andCorporate Change 7 (2), 217–247.

Lim, L., Garnsey, E., Gregory, M., 2006. Product and process devel-opment in the biopharmaceutical sector: a new perspective ondevelopment. R&D Management 36 (1), 27–36.

Lindholm-Dahlstrand, A., 1997. Growth and inventiveness intechnology-based spin-off firms. Research Policy 26, 331–344.

Lorenzoni, G., Baden-Fuller, C., 1995. Creating a strategic center tomanage a web of partners. California Management Review 37 (3),146–163.

Moore, J., 1993. Predators and prey: a new ecology of competition.Harvard Business Review May–June, 75–86.

Moore, J., 1996. The Death of Competition: Leadership and Strategyin the Age of Business Ecosystems. John Wiley, Chichester.

Mustar, P., Renault, M., Colombo, M.G., Piva, E., Fontes, M., Lockett,A., Wright, M., Clarysse, B., Moray, N., 2006. Conceptualizingthe heterogeneity of research-basedspin-offs: a multi dimensionaltaxonomy. Research Policy 35, 289–308.

Nelson, R., 2006. Evolutionary social science and universal Darwin-ism. Journal of Evolutionary Economics 16 (5), 491–510.

Nelson, R., Winter, S., 1982. An Evolutionary Theory of EconomicChange. Belknap Press, London.

Nystrom, P.C., Starbuck, H.W., 1984. To avoid organizational crises,unlearn. Organizational Dynamics 12 (4), 53–65.

O’Keeffe, E., Williamson, P., 2002. ARM Holdings Plc, European CaseClearing House, Case No. 02/9999.

Panhuis, T.M., Butlin, R., Zuk, M., Tregenza, T., 2001. Sexual selectionand speciation. Trends Ecology & Evolution 16, 364–371.

Penrose, E., 1959. The Theory of the Growth of the Firm. Wiley, NewYork.

Penrose, E., 1960. The growth of the firm—a case study: the HerculesPowder Company. Business History Review 34 (1), 1–23.

Piol, E., 2005. Il sogno di un’impresa. Dall’Olivetti al venture capital:una vita nell’Information technology, Il Sole 24 Ore.

Rohlfs, J., 2001. Bandwagon Effects in High-Technology Industries.MIT Press MA, London.

Rosenberg, N., 1994. Exploring the Black Box: Technology, Eco-nomics, and History. Cambridge University Press, Cambridge.

Sapsed, J., 1999. Restricted Vision: Strategizing under Uncertainty.Unpublished D.Phil. thesis. University of Sussex.

Schein, E., 2003. DEC is Dead, Long Live DEC. Berrett-KoehlerPublishers.

Schot, J.W., Geels, F.W., 2007. Niches in evolutionary theories oftechnical change: a critical survey of the literature. Journal ofEvolutionary Economics 17 (5), 605–622.

Shane, S., Stuart, T., 2002. Organizational endowments and the per-formance of university start-ups. Management Science 48, 154–170.

Stinchcombe, A.L., 1965. Social structure and organizations. In:March, J. (Ed.), Handbook of Organizations. Rand-McNally,Chicago, IL, pp. 142–193.

The Guardian, 2001. From the BBC Micro, little Acorns grew, March8.

von Hippel, E., 1986. Lead users: a source of novel product concepts.Management Science 32 (7), 791–805.

Yin, R.K., 1994. Case Study Research: Design and Methods, seconded. Sage, Thousand Oaks, CA.

Acorn ARM published Research Policy 2008

Documents

Transcript of Acorn ARM published Research Policy 2008