Topics in the Management of Technology and Innovation A Synopsis of Major Findings

In addition, models of the innovation process were developed to support managerial actions.Whereas the theories by Schumpeter attempt to gain a fundamental insight into the nature and th

Management of Technology and Innovation:

A Synopsis of Major Findings

Topics in the Management of Technology and Innovation:

A Synopsis of Major Findings

Koenraad Debackere, Department of Applied Economics, K.U.Leuven

1 Introduction

Technology is a major stimulus for change in society We have come to look to technologicalinnovation to rescue us from the consequences of exhausting natural resources; to abate inflationthrough productivity increases; to eliminate famine; to cure cancer; and to maintain the competitiveposition of our nations’ industrial bases Indeed, technological change has become a major driver ofcompetition: it propels new firms to the forefront of the competitive arena while it destroys thecompetitive advantage of even well-entrenched firms Achievements such as electronic computers,test tube babies, supersonic aircraft, and manned space flights have bolstered our faith in technicaladvance We no longer ask if something is possible, but how soon it can be done and at what price.There is little doubt that the rapid technological progress we have witnessed during the lastdecades will come to an end soon Today, researchers around the globe are working intently ondeveloping ideas that may create new branches of technological practice and could ultimatelytransform industry in ways which are hard for most of us to imagine As a consequence, the ability ofmanagers and policy makers to comprehend the pace and the direction of technological advancementwill largely determine a firm’s or nation’s competitive performance in world markets into the nextcentury This is no small task, however Historical accounts of industrial evolution and innovation,such as with the development of semiconductors (Braun and Macdonald, 1978), videocassetterecorders (Rosenbloom and Cusumano, 1987), and personal computers (Smith and Alexander, 1988),show the immense difficulties some firms encounter when confronted by new technologies Theinertia, introduced by a firm’s existing technological base, often is a powerful barrier to internalisenew technological trajectories (Utterback, 1994) Hence, there is an obvious need to harness theprocess of technological innovation effectively To do this, technological innovations cannot beisolated from the complex economic, social and political systems within which they operate

As a consequence, an extensive research agenda into the nature of the technological innovationprocess started in the 1950s This brought a recognition that innovation is an activity which needscareful ‘managerial’ attention and actions But, before there can be effective management, theremust be a detailed understanding of the process of innovation, its characteristics and its specificproblems.

Therefore, the first part of this paper will focus on the major characteristics of the innovationprocess as they have emerged over the last three decades The models to be discussed are chosen forthe complementary insights they offer into the complex nature of the innovation process We startwith the theories on technical change developed by Schumpeter Although the study of technicaladvance as an economic phenomenon is a relatively recent event, it was Schumpeter who in three

books, The Theory of Economic Development (1934), Business Cycles (1939), and Capitalism,

Socialism and Democracy (1942), portrayed most fully the active role played by economic agents in

technical advance From these studies, technological innovation emerges as a non-linear, dynamic,interactive and complex process

In addition, models of the innovation process were developed to support managerial actions.Whereas the theories by Schumpeter attempt to gain a fundamental insight into the nature and thecauses of technological evolution, the three models discussed subsequently focus on the innovationprocess within the firm The first model, by Roberts and Frohman (1978), depicts technologicalinnovation as a process of uncertainty reduction This process necessitates three important activitieswithin the firm Ideas have to be generated Once generated, these ideas have to be turned into goodcurrency And, finally, appropriate organisational structures have to be implemented to manage thetransition from what first seems to be an abstract ‘idea’ into a ‘product’ desired by customers As aconsequence, an important focus of this model is on managing people and their innovative ideas.The second model explicitly makes the link between technological innovation on the one handand organisational strategy and structure on the other hand The Abernathy-Utterback model (1975 &1978) describes how product and process innovations evolve during the technological life cycle of a

‘productive unit’ or ‘business unit’ and, still more important, how competitive strategy, productionfacilities, and organisation structure ‘co-adapt’ during this evolution

Finally, the technological S-curve model (Roussel, 1984; Foster, 1986) enables managers tobetter estimate the strategic importance of the different technologies in a firm’s technology portfolio

To this end, the S-curve model is used to develop a technological typology A distinction is madebetween emerging, pacing, key, and base technologies The competitive implications of this typologyare highlighted A link is made with Wheelwright and Clark’s definitions of breakthrough, platformand derivative projects (1992)

Once we have obtained a basic understanding on the nature of the innovation process, we willturn our attention to the management of technological innovation This will be the theme of thesecond part of this paper From the previous discussions, we know that ‘managing’ the innovativecapabilities of the organisation involves different levels of attention: (1) attention to the relationshipbetween technology and strategy; (2) attention to an appropriate organisation structure in whichinnovative activity can flourish; and (3) attention to the management of innovative professionals It ishence necessary to discuss the management implications associated with each level of attention Theintegration of these three levels of attention leads to the development of a partnership model onorganising technology and innovation, as discussed by Roussel and his colleagues in their influential

book Third Generation R&D (1991).

Finally, the third part of the paper brings together the major issues raised in the previous sectionsand ends with a ‘checklist’ of focal activities that are to be considered when managing a firm’sinnovation efforts

Before embarking upon a detailed discussion of these topics, though, there is an obvious need toclarify two major concepts used throughout this paper, i.e what is meant by ‘technology’ and how do

we define ‘technological innovation’?

1.1 Technology: what’s in a name?

Throughout the decades of research on the management of technology and innovation, a host ofdefinitions has surfaced, attempting to describe what is meant by a ‘technology.’ According to the

Oxford Dictionary, technology is “the science of industrial arts.” This definition, despite its brevity, combines two concepts that are essential to fully grasp the meaning of ‘technology’: science and

arts Of course, we do not imply that technology is the same as science, not even that it always has to

be based on scientific principles or developments Indeed, examples exist where the technology was

in place before the underlying scientific principles were known or clarified One of the most notableexamples undoubtedly is the steam-engine It was developed before the science of thermodynamicshad originated However, the Oxford definition does imply that technology has an important

‘knowledge’ component Thus, a major input into technological activities is knowledge about why

things work the way they do This is the know-how versus the know-why question This knowledge

can be derived from scientific developments, but also, from previous technological experience

The Oxford definition also points to the fact that technology has to do with arts The products of human art are artefacts Artefacts are tangible products and processes created by human skill Thus,

contrary to scientific activity, the major output of technological activity is embodied in hardware, i.e.products and processes Technological output is tangible It is not mere knowledge Figure 1 (adapted

from Allen, 1977) highlights this important contrast between technological activity and scientificactivity The major inputs into any scientific activity are information and knowledge The majoroutputs of scientific activity are, once again, information and knowledge Oversimplified, scientistsread papers (knowledge input), they think and experiment, and they write papers (knowledge output).The major inputs into technological activity are also information- and knowledge-related However,the major outputs of technological activities are embodied in hardware, i.e products (which are moreand more frequently integrated with services, or vice versa) and processes.

— Insert Figure 1 about here —

Thus, technological activity can be defined as the processes by which knowledge (scientific and

experiential) is transformed into artefacts, i.e products and processes As a consequence,

technological activity is characterised by both a less-tangible knowledge component (i.e the side of the equation) and a tangible product or process component (i.e the output-side of theequation) Having defined ‘technology’, we still have to clarify the concept of ‘technologicalinnovation.’

input-1.2 Technological innovation

Technological innovation is the successful commercial exploitation of inventions as they become

embodied into new products and processes The emphasis thus is on exploiting the results of

technological activity There are, of course, different opinions of what constitutes a new product orprocess In the most general and pure sense, the product or process developed is new to the world.This need not be the case, however Certain experts go as far as considering any product or process

an ‘innovation’ as long as it is perceived as new to the organisations involved, even though it mayappear to others to be an ‘imitation’ of something that exists elsewhere (Van de Ven, 1986)

Whereas the invention process may be hard to manage, the management of the innovation process(as a systematic approach to exploit inventions and reduce them to practice in a successful manner)has been well-embedded both in theory and practice over the last decades

Before turning to these managerial issues, though, it is necessary to address the followingquestion: What are the characteristics and the complexities involved along the innovation trajectory?

2 The process of technological innovation

In this section three different approaches to unravel the characteristics of the innovation processare discussed (the models by Roberts (1978)and by Abernathy & Utterback (1975), and the S-curve

by Roussel and Foster (1984 & 1986)) The ‘economic’ origins of innovation theory are highlightedfirst This economic debate has focused on the relationship between market structure and innovativeactivity

2.1 The economic debate: market structure, technology-push and market-pull

Schumpeter was the first to fully portray the active role played by economic agents in technicaladvance Schumpeter’s different books, though, reveal the many subtleties involved in explaining theorigins of technological innovations It is important to grasp those subtleties since they are essential

to understand the more managerial oriented models of the innovation process to be discussed in the

next sections In his first two books (The Theory of Economic Development, 1934 & Business Cycles,

1939), the entrepreneur plays a central role The entrepreneur is defined as the person who creates

new combinations He sees how to fulfil currently unsatisfied needs or he perceives more efficientways of doing what is already done These acts may, though need not, involve the presence ofinventions In some cases, it may only involve a new application of an existing technology As aconsequence, the act of invention and the act of entrepreneurship are separate: the inventor need notnecessarily be the entrepreneur and vice versa However, the entrepreneur plays a central role since

he is the one who turns the invention into exploitation

Given the importance attributed to the ‘entrepreneur,’ this theory has often been calledSchumpeter’s theory of ‘heroic entrepreneurship’ or ‘creative destruction.’ Indeed, the logic of thetheory is as follows (see Figure 2) There exists a pool of inventions related in an unspecified way tothe state-of-the-art developments in scientific and technological knowledge The importantobservation now is that this pool of inventions is largely exogenous to existing firms and marketstructures Thus, they are unrelated to any specific and quantifiable type of market demand Ofcourse, this does not mean that they may not be influenced by an anticipated demand or shortage Weall know that, in an abstract manner, human needs are infinite However, in the realm ofSchumpeter’s theory of heroic entrepreneurship, there is no direct coupling between a measuredmarket need (as we would detect from extensive market research, for example) on the one hand, andthe efforts invested and the directions chosen in the pool of inventions, on the other hand

The essential link between the ‘pool’ of inventions and the ‘market’ is made through the person

of the entrepreneur He is aware of the potential of certain inventions, and as a consequence,becomes prepared to take the risk and the commitment necessary to turn these inventions into

innovations Thus, innovating is more than inventing As defined previously, it is the (successful)

commercial exploitation of inventions Schumpeter then remarks that such a hazardous activity

would not be undertaken by an ordinary capitalist economic agent (such as an existing firm) Only

the entrepreneur has the vision, the drive and the commitment to survive the turbulence and theuncertainty involved If he succeeds, though, the rewards are enormous The entrepreneur will realiseexceptional (be it temporary) monopoly profits and he may be able to fundamentally alter existingmarket structures.

— Insert Figure 2 about here —Examples of the successes of ‘heroic entrepreneurs’ abound For instance, the advent of TexasInstruments as a major electronics firm can be seen as the result of heroic acts of technologicalentrepreneurship The company did not actually invent the transistor, though it made judicious use ofthe new technology to create products that would meet hitherto unfulfilled customer needs Otherexamples of heroic technical entrepreneurship are Edwin Land and the development of the Polaroidcamera and Joe Wilson who turned the Haloid Company, a small photographic paper and supplyfirm, into today’s giant Rank Xerox through his vision and ideas about a revolutionary copyingprocess More recent examples of ‘heroic entrepreneurship’ can be found in the formation of newbiotechnology firms such as Plant Genetic Systems, Genentech, Amgen, etc They all symbolise theentrepreneurial vision that attempts to turn ‘knowledge’ into ‘commercial exploitation.’ In doing so,

those firms are at the origin of what Schumpeter called “the eternal gale of creative destruction.”

In his third book, Capitalism, Socialism and Democracy (1942), Schumpeter’s focus on technical

progress took on new directions (see Figure 3) Instead of focusing solely on the ‘heroicentrepreneur,’ Schumpeter now incorporates the importance of scientific and technological activitiesconducted by (mostly large) firms In this additional model of the innovation process, the couplingbetween science, technology, innovative investment and the market, which was tenuous at best in thefirst model (see Figure 2), is much more intimate and continuous Successful innovations generateprofits leading to increased in-house innovative activity and R&D investments

— Insert Figure 3 about here —

As a consequence, the heroic entrepreneur is not the only central agent linking invention to itssubsequent exploitation Whereas science and technology are largely exogenous in the first modeldepicted in Figure 2, they are at least partly endogenised in the model described in Figure 3 Thus,the link between invention and exploitation is internalised within existing economic agents, i.e thefirm (and preferably the large firm, as Schumpeter hypothesised) As a consequence, the role of theheroic entrepreneur who couples invention and exploitation, is complemented by intrapreneurialmodes of invention exploitation Still more important, Schumpeter’s paradigm on the economics oftechnical advance inspired the hypotheses that innovative activity would be proficient in (1) largefirms and (2) in monopolistic industries

Large firms were deemed more innovative than small firms because they can finance a largerresearch and development staff, leading to economies of scale in R&D; because large firms are betterable to exploit unforeseen innovations given their more diversified product lines; and, becauseindivisibility in cost-reducing innovations makes them more profitable for large firms.

In the same vein, it was hypothesised that innovation would be greater in monopolistic industriesthan in competitive ones because a firm with monopoly power can prevent imitation and thereby cancapture more profit from an innovation; and, because a firm with monopoly profits is better able tofinance research and development (Kamien and Schwartz, 1982)

Although the hypotheses on the relationship (1) between firm size and innovative activity as well

as (2) between monopoly power and innovative activity have only received limited support, themodels described in Figures 2 and 3 further lead to the origins of a debate that has engaged students

of the innovation process for quite some time, namely: what is the causal direction of the relationshipbetween technological research and the market? In other words, is technological research the initiator

of innovations that lead to the creation of new markets (i.e the ‘technology-push’ hypothesis)? Or,

on the contrary, is it the market that initiates innovations (i.e the ‘market-pull’ hypothesis)?

Although the question on causality may seem superfluous, it has nevertheless importantconsequences, not in the least at the macro-level of economic policy-making Indeed, if one adheres

to the technology-push hypothesis, then one will recur to a supply-side oriented (neo-classic) economic policy with respect to technological innovation Enough money has to be invested inresearch facilities and R&D programs, and markets will ultimately follow suit Oversimplified, atechnology-push oriented policy will allocate considerable sums of money to R&D, hoping thatheroic entrepreneurs will tap the pool of knowledge thus generated and create new products andprocesses that ultimately serve markets

macro-On the other hand, a market-pull policy will stimulate innovation through creating a demand fornew products or processes This demand will trigger innovative behaviour For instance, in order tostimulate innovations in telecommunication technology, a market-pull oriented policy might operatethrough the creation of a national demand for a new telecommunication network This demand wouldthen spur the innovative behaviour of the organisations participating in the national program Atechnology-push oriented policy, on the other hand, might operate through formulating R&Dprograms in telecommunications technology

A scrutiny of government policy frameworks aimed at stimulating technical innovation showsthat, in fact, both technology-push and market-pull orientations are relevant Innovations have theirorigins both in the market and in the creation of new technological capabilities For sure, researchhas shown that market-induced innovations tend to have a higher probability of commercial successthan innovations that originate from a technological capability and that are isolated from a market-

selection environment However, this relationship between market-relatedness and commercialsuccess is moderated by the fact that market-induced innovations tend to be more incremental andthus less radical than innovations having their origins in the research laboratory (Rothwell et al.,1977).

Moreover, it appears that technological innovation certainly is not a linear, sequential process asmight be (incorrectly) deduced from Figures 2 and 3 Instead, it is a complex, multi-stage, cross-functional, and multi-disciplinary process in which both supply-side and demand-side arguments arerelevant and should be carefully considered This is all the more true when studying technologicalinnovation at the firm-level Here the managerial models of technological innovation becomerelevant

2.2 Managerial models of technological innovation

2.2.1 The process of technological innovation: a general model

The first managerially relevant model to be discussed is the one proposed by Roberts andFrohman (1978) This model (see Figure 4) emphasises three key generalisations First of all, ideasand opportunities for innovation originate both from the supply-side (‘technology’) and the market-side (‘market’) Thus, both the ‘technology-push’ and the ‘market-pull’ dimensions are highlyrelevant Second, the process of technological innovation is a multi-stage or -phase process.Significant variations in the primary task as well as in the managerial issues and effectivemanagement practices occur across these different stages Third, in the model, six stages arepresented The exact number of stages or phases is, of course, somewhat arbitrary What is key,though, is that each phase of activity is dominated by the search for answers to different managerialimperatives Finally, each phase requires clear go/no go decision points and phase-reviews

— Insert Figure 4 about here —

At the outset (stages 1 and 2), emphasis is on finding a motivating idea, a notion of possibledirection for technical endeavour Thus, ideas have to be generated and, still more important, oncegenerated, attention has to be paid to those same ideas Good managerial practice at these earlystages frequently involves loose control, the pursuit of parallel and diverse approaches, fosteringconflict or at least contentiousness, and stimulating a variety of inputs Small amounts of moneyshould be rather freely available to enable the assessment and evaluation of the ideas generated Amajor mistake is to set up rigid formal processes for approval of the small sums needed to try out anidea But, most of all, an organisational environment and culture has to be developed that tolerates

new ideas and allows proper attention to be paid to them 3M’s statement “Thou shall not kill an

idea” clearly reflects what is meant by this attitude In addition, a tolerance for new ideas also

implies a tolerance for failure As has been often documented in the innovation management

literature, false foundations can prove to be challenging new starts This, though, is only possible

when ‘failures’ are tolerated

As we move further along the different stages, the managerial issues and the actions requiredchange dramatically During stages 4 and 5, for example, the task involves in-depth specification andmanufacturing engineering of ideas that are being reduced to an acceptable working prototype Themanagerial issues evolve towards co-ordinating a number of scientists and engineers of differentdisciplinary backgrounds to achieve, within previously estimated development budgets andschedules, a predefined technical output ready for manufacture in large volumes; reliable and atcompetitive production costs Effective managerial practice will thus involve tight control,elimination of duplication, strong financial criteria and formal evaluations of resource use, evensomewhat rigid adherence to planning Thus, during the later stages of the innovation process, themanagerial actions required are rather opposite to the ones advocated during the first stages of theprocess Whereas the first stages focus on ‘managing ideas,’ the later stages focus on themanagement of the ‘part-whole relationships’ required to turn these ideas into a tangible product orprocess Part-whole management indeed requires the co-ordination of the efforts of people withdifferent disciplinary backgrounds, working together toward achieving the new product or processgoals

As is obvious from Figure 4, innovation occurs through efforts carried out primarily within anorganisational context, but involving heavy interaction with the external technological as well asmarket environments Proactive search for technical and market inputs, as well as receptivity toinformation sensed from external sources, are critical aspects of technology-based innovation Manystudies of effective innovations have indeed shown the presence of significant contributions ofexternal technology (e.g via contacts with universities) and have found success heavily dependentupon awareness of customer needs and competitor activity

Another remark is warranted with respect to the model in Figure 4 For ease of presentation, allstages are shown at equidistant intervals inappropriately suggesting perhaps the similarity of thesephases from a time duration and/or resource consumption perspective Stage 5, commercialdevelopment, for instance, usually takes as long as the several earlier stages combined and requiresmore resources than most of the other stages together This is why tight financial control becomesnecessary at this stage A typical cash flow diagram for new product or process development isshown in Figure 5 (source: Twiss, 1992) In addition, current practices in the area of concurrentengineering have emphasised the possibility to engage in overlapping problem-solving activities,pointing to the different phases just discussed running (partly) in parallel and simultaneous

— Insert Figure 5 about here —

From Figure 5 it is obvious that the upstream ‘research-oriented’ activities (idea formulation andproblem solving) only consume minor expenditures compared to prototyping, manufacturing start-up,and market launch Of course, the relative importance of these items further depends upon the nature

of the industry In aerospace, for example, the construction and testing of prototypes usually requiresmajor expenditures, while investments in manufacturing or marketing are small or non-existent Inchemical industries, on the contrary, investments in efficient production facilities may be the majoritem Consumer industries, on the other hand, may incur extremely important marketing costs inlaunching the new product

Finally, although only a limited amount of feedback loops is shown in Figure 4, they inevitablyexist causing reiteration among the different stages in the model For example, the problem-solvingprocesses in stage 3 often generate useful insights for alternative idea formulations (stage 2) Also,during stage 6 (transfer to manufacturing) new problems may surface requiring a reiteration to theproblem-solving stage (stage 3) Thus, the real process of technical innovation involves flows backand forth over time among differing primary activities, internal and external to the innovatingorganisation, with major variations arising throughout the process with regard to specific tasks,managerial issues and managerial answers Still more important, as mentioned above, the practice ofconcurrent engineering and parallel development has convinced managers of the possibilities ofhaving several innovation tasks running in parallel as well as of the potential of engaging inoverlapping problem-solving activities (Deschamps and Nayak, 1995)

2.2.2 A dynamic model of the innovation process

The above model gives a general description of the innovation process and the different stagesinvolved Another extremely influential model was developed by Abernathy and Utterback (1975 &1978) It takes us one step further It describes how the nature of a company’s innovation activity(and its response to innovative ideas) changes as it grows and matures It investigates how the types

of innovations attempted by productive units change as these units evolve Thus, the model takes adynamic stance towards the innovation process It relates patterns of innovation within a productiveunit to that unit’s competitive strategy, production capabilities, and organisational characteristics.The model implies that a productive unit’s capacity for and methods of innovation depend critically

on its stage of evolution from a small technology-based enterprise to a major high-volume producer(see Figure 6; source: Abernathy, 1978)

— Insert Figure 6 about here —The model thus focuses on innovation patterns within a productive unit For a simple firm or afirm devoted to a single product, the productive unit and the firm would be one and the same In the

case of a diversified firm, a productive unit would usually report to a single operating manager andnormally be a single operating division Thus, the productive unit consists of both a manufacturingunit and the product line produced targeted toward specific markets or market segments Forexample, an engine plant and the line of engines it produces is one productive unit An assemblyplant and the particular car it produces is another As a consequence, a productive unit can becompared to the business unit form of organisation as it has been well-documented in the literature

on organisational design and development

In the model, two major stages along the unit’s innovation trajectory are discerned During thefirst phase, the product design is subject to major changes, product characteristics areunderdetermined, product innovations are numerous However, their emphasis is on improving thefunctional performance of the product rather than reducing product costs This phase is therefore

called the fluid stage As the product is still in flux, production systems have to be highly flexible but

at the same time, they necessarily are inefficient In sum, during the fluid phase, the competitiveemphasis of the productive unit is on functional product design performance Recent insights from

the management control and accounting literature, focusing on the notion of ‘target costing,’ further

show that during these early product definition phases, about 85 percent of the product’s final andtotal cost structure is being determined

Many of the product innovations introduced during this first phase are to a large extent stimulated

by information on users’ needs and users’ technical inputs about how they want the product to

perform von Hippel (1977) has identified these users as ‘lead users.’ They are the users who are ‘at

the front of the trend.’ Not only do they face the need for the new product or process well in advance

of the bulk of the marketplace But, they also are willing to experiment with the new product orprocess to obtain maximal satisfaction and solutions to their needs As a consequence, they represent

an important input to the productive unit which is still experimenting with the product in order toestablish a ‘stable’ product design This stable design or standard, once achieved, is called adominant design in the Abernathy-Utterback model (cf infra)

Thus, during the fluid stage, the predominant mode of innovation are frequent and major changes

to product designs The product line then necessarily is highly diverse, often including customdesigns The production process is flexible and inefficient since major product changes have to beaccommodated easily Production equipment is general-purpose and requires highly skilled labour.Materials requirements are limited to generally available materials, as long as the productcharacteristics have not stabilised (i.e as long as a dominant design is absent) The productionfacilities will be small-scale Organisational control is informal and entrepreneurial

The second major phase to be discerned in the dynamic model is characterised by a stabilisedproduct concept (a dominant design has emerged) The product is standardised, changes are highlyincremental and production systems are rigid but efficient Therefore, this phase has been called the

specific stage During the specific stage, competitive emphasis is on cost reduction Innovations level

off (both from a product and from a production process perspective) and are stimulated by thepressures to reduce costs and to improve quality As well product as process innovations thus becomehighly incremental, with cumulative improvements in productivity and quality The product line hasstabilised significantly Product variety and flexibility occur within determined boundaries Whereasduring the fluid stage, products often contain custom-designed features, the specific stage ischaracterised by mostly undifferentiated standard products

To achieve these cost and quality imperatives, the production process is streamlined It isincreasingly efficient, capital-intensive, and rigid The cost of change to the process is large Majorchanges in product characteristics would almost certainly involve major investments to turn aroundthe existing production facilities Specific-stage production processes are characterised by special-purpose machines, mostly automatic with labour tasks mainly monitoring and control Specialisedmaterials will be demanded If these materials are not readily available from reliable suppliers,vertical integration will be extensive Production plants are large-scale, often dedicated to specificproducts (or product families) Economies of scale are important Organisational control is tight.Emphasis is on structure, goals, and rules

As a consequence, the fluid and specific stages are quite opposite with respect to the managerialissues and responses they raise The stage linking the fluid to the specific patterns in the model is

called the transition stage During the transition stage, a dominant design emerges The dominant design marks a reduction in product innovations It marks the advent of a ‘product standard.’ For

example, the DC-3 aircraft is an example of a dominant design The DC-3 was a cumulating of priorinnovations It was not the largest, or fastest, or longest-range aircraft It was, however, the mosteconomical large, fast plane able to fly long distances All the design features introduced in the DC-3had been proven in prior aircraft It was the combination of those prior innovations into one stable,

‘standard’ product that made the DC-3 a unique design No major innovations were introduced intocommercial aircraft design from 1936 onward until the development of the turbojet enabled a newgeneration of aircraft in the 1950s Instead, many incremental refinements were made to the DC-3concept During the period of these incremental changes, airline operating costs per passenger-miledropped an additional 50 percent Similar design milestones have been identified in numerousproduct lines, for example: the internal combustion engine, the standardised diesel locomotive,sealed refrigeration units for home refrigerators and freezers (see also Abernathy & Utterback, 1975and Abernathy, 1978), and more recently the personal computer or the Internet protocol TCP/IP.Thus, during the transition stage, competitive emphasis is on product variation rather than onfunctional performance (fluid stage) or cost reduction (specific pattern) Innovation is stimulated byexpanding the company’s internal technical capabilities The rate of product innovations slows downand takes on another character The advent of a stable, dominant design enables significant

production volumes It has the effect of enforcing standardisation so that production economies can

be sought This implies that subsequent product innovations, whenever they occur, will be ratherincremental

As far as the production process is concerned, major innovations are realised during the transitionperiod, though (see also the changes in both curves in Figure 6) The rising volume, enabled by thestabilised product design, calls for a more rigid production process with this rigidity being added inmajor steps Sub-processes become increasingly automated, ‘islands of automation’ are created.Specialised materials may be demanded from some suppliers, the plant is a mixture of general-purpose machinery combined with specialised (automated) sections Organisational control begins totighten through the creation of liaison functions, the formation of project groups and task forces.Thus, the dynamic pattern underlying innovation processes within a productive unit has importantmanagerial implications As a unit moves towards large-scale production, the goals of its innovationprocess change from ill-defined and uncertain targets to well-articulated design objectives In theearly stages, there is a proliferation of product performance requirements and design criteria whichare difficult to quantify ‘Lead users’ provide important inputs to reduce this performanceuncertainty During this initial stage, market needs are ill-defined and market uncertainty is high.The relevant technologies are only marginally understood As the enterprise develops, though, bothtechnological and market uncertainties are reduced Larger investments in innovative research,process engineering, and production facilities can be justified

As the innovation patterns evolve over time, the organisation as such also has to adapt Theorganisation’s methods of co-ordination and control change with the increasing standardisation of itsproducts and production processes The enormous amounts of uncertainty during the first stage of itsdevelopment put a premium on the capacity of the organisation to process the information necessaryfor uncertainty reduction This can be done by establishing vertical and lateral information systems

as well as project management structures Later, these may be complemented with more formalmechanisms such as task force groups and management control systems such as job procedures andjob descriptions

As the organisation moves into the specific stage, though, products are standardised and change is

at best incremental This is reflected in the control and co-ordination systems in place in the

organisation: formalisation and routinisation become prevalent They both tend to reduce the need

for information processing

The dynamic model by Abernathy and Utterback has provided many useful insights into thecharacteristics and the complexities of the innovation process Whereas the model by Roberts andFrohman presents a way to look at the different stages of a particular innovation effort (mostly using

a project structure as a vehicle for execution), including the relevant managerial issues and solutions,

the model by Abernathy and Utterback explicitly brings a multitude of organisational characteristicsand their dynamics (described at the level of a productive unit) into play As a consequence, it yieldsinsights complementary to the ones derived from the Roberts and Frohman model.

At the same time, though, we have to conclude this discussion of the Abernathy-Utterback modelwith some cautionary comments First of all, although the model seems rather generalisable acrossindustries, it need not necessarily be applicable to all possible types of productive units that exist.The model is easily applicable to industries such as the automotive industry, the electronics industry,and the mechanical industry It becomes different, though, when studying bulk chemicals, forinstance In this latter industry, product innovations are minimal or non-existent What matters isbuilding an extremely efficient and reliable process able to manufacture the standard ‘commodity’product This remark also holds for other ‘process industries.’ Thus, the Abernathy-Utterback model

is less appropriate to describe the dynamic innovation patterns as they occur in process industries.The same is true for certain service industries If we take as example a hamburger restaurant chain,such as Burger King or MacDonald’s, we will have obvious difficulties to observe thecomplementary product and process innovation patterns and their evolution over time, as they aredepicted in Figure 6 Once again, the product innovation curve is not present

The second remark on the model is still more fundamental Oversimplified, it boils down to thefact that the Abernathy-Utterback model, as it has just been described, implies a fundamentaldilemma: innovativeness and productivity cannot be achieved simultaneously During the fluid stage,the enterprise displays a high degree of innovative behaviour However, because of the extremedegrees of flexibility required to be innovative, the organisation just cannot be efficient and highlyproductive The opposite is true during the specific stage Here, the organisation puts a premium onproductivity As a corollary, innovative activity is at best incremental, if it has not totallydisappeared

The changing environment of the firm, with its emphasis on lead time reductions, efficientproduct development, and time-based competition, necessitates the firm to achieve both objectives(productivity and innovativeness) more or less simultaneously This is no small task and theAbernathy-Utterback model, through its description of the internal dynamics of innovation patternswithin productive units, points to the many hurdles and to the organisational inertia to be overcome

to develop this simultaneity Fortunately, technological evolution in and off itself can support thetrend towards greater simultaneity More specifically, the emergence of new technologies (such ascomputer-aided-design, computer-aided-manufacturing, flexible manufacturing systems, andcomputer-integrated-manufacturing), if used judiciously, precisely aim at bridging this difficult gapbetween ‘innovativeness’ and ‘productivity.’ Hence, the dominant logic as presented by Abernathyand Utterback need not be as irreversible anymore as it was suggested in the early 1980s However,the natural tendency for the firm’s innovative activity to become more focused, less original and

more formalised as the business matures of course remains a relevant threat to any innovationmanagement programme.

The two models of the innovation process we have discussed sofar, have yielded many significantinsights to better understand the managerial issues involved Each model has proven to be extremelyuseful at a particular level of action The Roberts-Frohman model is appropriate to understand theinnovation process at the project level of action In other words, it is valuable to understand the manyissues involved in managing innovation projects The Abernathy-Utterback model is more useful tounderstand the life cycle of a productive unit and to suggest appropriate managerial actions at theorganisational (and even strategic) level However, we still lack an important perspective in order toachieve a truly holistic view of the innovation process

Indeed, most organisations are not committed to one specific technology Instead, they manage aportfolio of different technologies, some of which are more important to the company’s competitiveposition than others Thus, we need a model that allows us to prioritise different technologies infunction of their impact on the company’s competitive position This can be achieved through theconcept of the technological S-curve as described by Roussel (1984) and Foster (1986)

2.2.3 The technological S-curve

Just like living organisms, technologies have life cycles, from birth to old age Indeed, analysis ofhistorical data from a considerable number of phenomena shows that technological progress is notrandom and discontinuous Instead, it follows a regular pattern when a selected attribute, such asfunctional performance (e.g number of MIPS for computer CPUs, aircraft speed, efficiency ofinternal combustion engines), a technical parameter (e.g tensile strength to density ratio for a newmaterial), or economic performance (e.g operating cost per passenger-mile for aircraft) is plottedagainst time Typically, an S-shaped pattern emerges (see Figure 7) These S-curve patterns can then

be applied to a more qualitative analysis of the different technologies represented in the company’stechnology portfolio

At birth, a new technology is called embryonic or emerging Emerging technologies are defined

as new technologies with a potential impact on industry structure The exact boundaries of thisimpact are unknown, though Both technical and market uncertainty are extremely high It may stilltake one-to-two decades before tangible commercial products emerge In the beginning of the 1990s,technologies such as micro-machinery, optical computing or molecular farming (a biotechnologyapplication where plants and animals are conditioned to produce selected molecules) could beconsidered ‘emerging.’ The technological problems that remain to be solved are huge, marketuncertainty is very high The route from vision to industrial reality is still cloudy The emerging state

is one of substantial techno-scientific tumult and controversy However, if it can capture the interest

and commitment of enough bright minds all over the world, it will become the favourite pastime ofmany research laboratories.

— Insert Figure 7 about here —

Technologies in the next stage of the life cycle are called pacing technologies Pacing

technologies are technologies in an early development stage with a demonstrated potential forchanging the basis of competition By the end of the 1980s, prominent examples of pacingtechnologies were: neural network technology (a computing technology different from traditionalalgorithm-based von Neumann computing) and computer integrated manufacturing technology Itmay still take several years before the competitive impact of a pacing technology is fully realised.Important leaps in the technology’s performance still have to be realised before it reaches its full-blown commercial potential as embodied in a product or process

Next, key technologies are those technologies that have the greatest impact on competitive performance at a particular time Finally, base technologies are the very rocks on which the company

rests They are essential, but they are no longer critical to the basis of competition typically becausethey are widely available to competitors throughout the industry An often used example are CRT-displays Every computer company has to have them However, a company like IBM is ratherunlikely to realise a competitive advantage through its CRT-technology Instead, such an advantagemay be better realised through the company’s competencies in the field of micro-processortechnology or system architecture This technology would then be a key technology to the computerfirm Another example of a key technology is genetic engineering applied to plant protection Usingthis technique, companies have succeeded in making hybrid seeds that are bacteria resistant Theseseeds can only be used once, though, since they loose their bacteria-resistant properties after oneharvest period Thus, a company who has mastered plant protection technology through geneticmanipulation can acquire a substantial competitive advantage on the basis of its technologicalcompetencies

The S-curve pattern of technological evolution further implies that there are performance limits toany technological system In other words, performance increases cannot continue indefinitely Themodel points to the presence of diminishing marginal returns on R&D investments as the technologymatures As the technology nears its natural performance limit (e.g the efficiency of electric powerplants), the increase in performance for every man-year of effort invested to improve the technologydiminishes This process is often described as maturation and ageing of the technology As thematuration process continues, the technology becomes increasingly vulnerable and exposed to

substitutes or technological discontinuities that ‘destroy and replace old practice.’ In this context,

Abernathy and Clark (1985) refer to the competence-destroying or disruptive character (as well with

respect to technological competencies as with respect to market competencies) of revolutionary and

architectural innovations What is the impact of this model of technological evolution for the

practice of innovation management?

First of all, it implies that the strategic mission of R&D within any company is to exploit thepotential for improvement in competitive position in technologies that are important to the business.This means that the technology portfolio has to be connected to the business portfolio of the firm.This can be achieved through the development of technology roadmaps that link technologydevelopment to product line development (see Figure 8) Among the technologies in the portfolio,the ones that deserve foremost attention are, of course, the key technologies, then pacingtechnologies, and, always, competence in base technologies

 Insert Figure 8 about here Second, the firm has to acquire a detailed insight into the nature of its technology portfolio so as

to categorise its different technologies and to attach the appropriate priorities to each of them.Management has to assess its competitive position with respect to the different technologies in theportfolio In order to structure the management of the corporate technology portfolio, manyinnovation-intensive companies have appointed CTOs or Chief Technological Officers who areresponsible for portfolio management and monitoring An important aspect of managing thetechnology portfolio consists of managing the make-or-by (preferably the make-and-buy) decision.Third, the maturity of technologies in the portfolio provide insights into the potential for futureadvances And, this maturity distribution does not only point to diminishing returns on investments inperformance increases as the technology evolves along its S-shaped life cycle curve Still moreimportant, indeed, it points to the occurrence of discontinuities (see also Figure 7) A technologicaldiscontinuity is a technology which substitutes for the more mature technology, quickly surpassingthe ‘old’ technology’s performance potential A treacherous aspect of those technical discontinuities,though, is that, at the outset, their performance almost always appears to be inferior to theperformance of the more ‘mature’ technology they will ultimately replace For instance, decades ago,mechanical typewriters were a key technology, immensely superior to the pencil they replaced As

we all know, mechanical typewriters aged and they were superseded by electric typewriters Today,the electric typewriter is facing the same fate with the advent of the word processor

To conclude, the S-curve pattern of technological growth shows companies that they need tomanage their technology portfolio carefully Wheelwright and Clark (1992) further extend portfoliomanagement by introducing, besides the concept of a technology portfolio, the notion of a projectportfolio In doing so, they explicitly link technical change to product and process output In otherwords, whereas the technology portfolio is rather input-oriented (pointing to the diverse base of thefirm’s technological capabilities), we find that the project portfolio approach, as advocated byWheelwright and Clark (1992), is output-oriented (pointing to the varying impact the various

innovation efforts may have on the firm’s technology-product-market combinations) In Figure 9, theportfolio map proposed by Wheelwright and Clark is shown.

 Insert Figure 9 about here Research and advanced development projects or activities are taken apart in this portfolio model.This is because their uncertain and unpredictable nature causes extreme difficulties to foreseespecific outputs and results within a pre-defined time-frame and budget constraint Hence the plea toconsider them as a separate, long-term investment whose progress cannot be measured against well-defined and pre-determined criteria and standards This does, of course, not mean that the quality ofthe effort cannot be measured and monitored Only, in terms of future business performance,outcome predictability is low

From a strategic perspective, though, the distribution of the firm’s innovation efforts into

breakthrough, platform and derivative projects is crucial As is obvious from Figure 9, breakthrough

projects imply fundamental changes both from a product and from a process perspective New coreproducts and new core processes are created They support the long-term competitive position of the

company Platform projects are at the origins of the creation of new product families They

symbolise the degree of product-market differentiation and diversification the company aims at As a

consequence, platform projects are mostly medium-term oriented Derivative projects, finally, point

to incremental changes (both from a product perspective and a process perspective) that furtherenhance the performance (in terms of cost and/or functionality) of the firm’s existing platforms Bytheir very nature, they are short-term oriented

It is obvious that the bulk of the firm’s innovation efforts should go into the execution of platformprojects as they stand for the medium-term survival of the company Typically, experience suggeststhat 50-to-60 percent of the firm’s innovation efforts should be devoted to the creation of newplatforms Derivative projects are important since they sustain existing market relationships.However, portfolio management should be aware of the dangers entailed by placing too muchemphasis on derivative project activities since they quickly degenerate into imitative behaviour (asthey are often driven by short-term customer requests)

Technology and innovation thus become strategic issues that cannot be left into the hands of afew corporate scientists and technologists As a consequence, the three models just discussed allpoint to the necessity to ‘manage’ the innovation process carefully This then is the subject of thenext part of this paper

3 The management of technological innovation

In this section, the major issues related to the management of technological innovation will bediscussed Three dimensions are used to guide this discussion These can be succinctly summarised

as strategy, structure, and staff We already touched upon them during the description of the differentmodels of the innovation process The managerial implications of each dimension are now discussed

in greater detail

3.1 Technology and innovation strategy

During the discussion of the technological S-curve, we already suggested the concept of atechnology portfolio, consisting of a mixture of technologies with differing competitive impact.From the previous discussions, technological innovation appears to be a highly uncertain andcomplex process A firm cannot predict accurately the outcome of its own innovative efforts or those

of its competitors, so that the hazards and risks which is faces if it attempts any major technicalchange are very great Yet not to innovate ultimately means to die Still more, in times of turbulent

market environments, innovation provides a powerful instrument to prevent a status-quo or

steady-state from settling in A well-articulated innovation strategy is a starting-point to continuouslyquestion one’s assumptions about products, markets and technologies

3.1.1 A typology of innovation strategies

One way to approach the strategic management of innovation is to look at the various strategiesavailable to a firm when confronted with technical change Any classification of strategies isnecessarily somewhat arbitrary and, to a certain extent, violates the infinite variety of strategicsolutions in the real world However, the ‘ideal types’ to be discussed offer useful starting points toany strategic planning and strategic implementation effort Freeman (1982) proposes a typologyconsisting of six generic ‘ideal types.’

An ‘offensive’ innovation strategy is one designed to achieve technical and market leadership by

being ahead of competitors in the introduction of new products Since scientific and technologicaldevelopments nowadays are quickly accessible to other firms, such a strategy necessitates strongrelationships with leading parts of the world’s science-technology complex, or strong independentR&D, or the ability to react very quick on new possibilities, or some combination of these strengths.Strong relationships with the scientific-technological environment involves the recruitment of keyindividuals, good personal linkages with leading scientists and technologists (for instance, throughthe presence of ‘technological gatekeepers,’ cf infra), consulting arrangements, contract research,

and good information systems Unfortunately, the information and knowledge necessary for anyinnovation is unlikely to come from a single source, and even if it does so, is unlikely to be in aformat readily applicable to products or processes Therefore, ‘offensive’ innovators will attachcrucial importance to their internal R&D department The R&D department must generate thatscientific and technical information which is not available from outside and it must take theinnovation to the point at which normal production operations can be launched.

Consequently, ‘offensive’ innovators will have to be highly ‘research-intensive’ firms They willusually have important in-house fundamental research capabilities, as well as strong developmentcompetencies They will attach extreme importance to patents, to the acquisition of scientific andtechnical information, and to the education and training of their technologists They will also developthe necessary long-range planning techniques and tools to keep abreast of their competitors

Examples of offensive innovators are: DuPont during the development of nylon and Corfam,RCA during the development of television and colour television, and ICI during the development ofTerylene It took ten to twenty years from the inception of the research before most of theseinnovations showed any profits Many never did so

Only a small minority of companies in the world are willing or able to follow ‘offensive’innovation strategies Even companies that follow an offensive innovation strategy are only able to

do so consistently over limited time periods In any case, they will often have a portfolio oftechnologies at different stages of their S-curve, some of them emerging, others occupying a keyposition and still others that are maturing or even ageing

According to Freeman (1982), the vast majority of firms, including some of those that were once

‘offensive’ innovators, will follow a different strategy: ‘defensive,’ ‘imitative,’ ‘dependent,’

‘traditional,’ or ‘opportunist.’ A summary of the requirements for each of the six technologystrategies in terms of scientific and technical efforts by the firm is shown in Figure 10 (source:Freeman, 1982)

 Insert Figure 10 about here 

A ‘defensive’ strategy does not imply absence of R&D On the contrary, it may be equally

research-intensive as an offensive strategy The difference lies in the nature and the timing of theinnovations The ‘defensive’ innovators do not wish to be the first in the world, but neither do theywish to be left behind They may prefer to avoid the risks of being the first to innovate and they mayattempt to learn from the mistakes of first movers and from their opening up of the market Forexample, in the case of the videocassette recorder (Rosenbloom and Cusumano, 1987), Sony learned

a lot from the deficiencies of the Ampex machines (Ampex being the first mover) As is obviousfrom Figure 10, the ‘defensive’ innovator attaches as much importance to development and design

work as the offensive innovator does He puts less emphasis on research activities, though; whilepatents, education, and the acquisition of scientific and technical information remain of utmostimportance The emphasis on the different capabilities as shown in Figure 10 is obvious: the

‘defensive’ innovator must be capable at least to catch-up with the game, if not to ‘leap-frog.’ A

‘defensive’ strategy is more characteristic of firms in smaller industrialised countries

‘Defensive’ innovators do not usually aim to produce an exact copy or imitation of the products

or processes introduced by early innovators Instead, they tend to take advantage of early mistakes toimprove upon the design and to differentiate their products by minor technical improvements To do

so, they must have the necessary technical strength They will prefer to build an independent patentposition rather than taking licenses

‘Imitative firms,’ on the contrary, will not aspire to ‘leap-frog’ or even to ‘keep up with the

game.’ The ‘imitative’ firm is happy to follow way behind the leaders in established technologies,often a long way behind As a consequence, the technical competencies of the imitator will be much

less advanced than the ones of offensive and defensive firms (see Figure 10) A ‘dependent’ strategy

involves the acceptance of a subordinate role in relation to other stronger firms The ‘dependent’ firmdoes not attempt to initiate or even imitate technical changes in its product, except as a result ofspecific requests from its customers or its parent Typically, it has lost all initiative in product designand has no R&D facilities (see Figure 10) The pure ‘dependent’ firm can be considered a department

of a larger firm

The ‘traditional’ firm differs from the dependent firm in the nature of its product The product

supplied by the ‘traditional’ firm changes little, if at all The product supplied by a dependent firm,

on the other hand, may change a lot Though, these changes are responses to initiatives andspecifications from outside The ‘traditional’ firm sees no reason to change its product because themarket does not demand a change, and the competition does not compel the firm to do so.Consequently, this type of firm’s technical capabilities are absolutely minimal (see Figure 10).Freeman calls them ‘the peasants of industry.’

Finally, a number of firms follow a strategy that can be qualified as ‘opportunist’ or ‘niche.’

Entrepreneurs identify some new opportunity in a rapidly changing market environment, which maynot require any in-house R&D, or complex design, but will enable them to prosper by finding animportant ‘niche,’ and providing a product or service which consumers value, but nobody else hasthought to provide Consequently, ‘opportunist’ firms do not need strong internal R&D capabilities.However, contrary to the ‘traditional’ firms, they need a strong awareness of their scientific andtechnological environment (see Figure 10) Thus, the acquisition of scientific and technicalinformation is a central focus of any opportunist strategy

It is obvious that the above typology consists of six ideal types They may be rather difficult toobserve in their pure form in reality However, the major managerial imperatives of each genericstrategy are clear.

3.1.2 Formulation and implementation of a technology and innovation strategy

The strategic management of technology includes both a planning and an implementationcomponent Strategic planning involves the formulation of the organisation’s goals and objectives,and the policies needed to achieve those objectives, including identification of the organisation’sprimary resources and priorities

The technological S-curve implies that a company should carefully evaluate the position of thedifferent technologies in its portfolio along their growth curve or life cycle As already describedpreviously, key technologies are of utmost importance in supporting the competitive advantage of thefirm Base technologies, on the contrary, are a necessary prerequisite to the survival of the firm.However, in and off themselves they will not procure the firm a competitive advantage Thus, inevaluating its firm’s technology portfolio, management should consider the following questions:

• What are the different technologies in the portfolio?

• Which of these are emerging, pacing, key, and base technologies?

• What is our position with respect to each technology in the portfolio viz our major

competitors? Are we leading or are we lagging behind? And, if we are lagging behind, is their

a chance to bridge the gap?

• What technologies might be included in the portfolio? In which technologies should we increase/decrease our investments given their (potential) attractiveness to our business?

• How does this technology portfolio translate into a project portfolio, distinguishing among breakthrough, platform and derivative activities?

• Once the above questions have been answered, the firm should focus on the ‘make’ or ‘buy’ decision In other words, what technologies in the portfolio are going to be developed

internally and what technologies will be acquired from external sources?

As further suggested by the Abernathy-Utterback model, different strategic imperatives areassociated with each stage of technological development The earliest stage tends to feature frequentmajor product innovations, heavily populated and characterised by small entrepreneurialorganisations, often closely tied to user needs This is clearly illustrated in the case of newbiotechnology firms, most of which directly originated from university laboratories or from otheryoung small firms This has lead inevitably to the explosion of alliances between large companies