inovarea in sme

18
1 INNOVATION IN SME'S: THE MISSING LINK A. Le Bars, V. Mangematin, L. Nesta INRA/SERD Université Pierre Mendès France BP 47 X 38040 Grenoble Cedex 9 Tel: +33.(0)4.76.82.54.39 Fax: +33.(0)4.76.82.54.55 E-mail : [email protected] or [email protected] or [email protected] Abstract : The economics of innovation has addressed two main issues, respectively regarding the types of knowledge (scientific, technological) required in innovation, and the organisation of R&D. Drawing heavily on the central significance of R&D activities, such premises do not provide convincing explanations of other kinds of innovation that do not systematically draw on R&D efforts. We argue that innovation in SME's is critically related to individual members of the firm. Indeed, individuals embody two types of competencies, respectively component and architectural ones. While component competencies are linked to specific scientific or technological skills, architectural competencies deeply shape the innovation process by enhancing new combinations of existing component competencies. We further develop this framework by applying it to both high tech and low tech sectors. We show that architectural competencies are essential in the sense that they allow low tech SME's to innovate without performing internal R&D activities. While stars scientists remain central elements of architectural competencies in high tech industries, engineers appear as key drivers of new innovation projects in low tech sectors. Key words : SME, R&D, innovation, competencies, low and high technology. Paper presented at the Sixth Annual International Conference at the University of Twente, the Netherlands High-Technology Small Firms Conference, 4-5 june 1998

Upload: mhldcn

Post on 11-Jan-2016

2 views

Category:

Documents


0 download

DESCRIPTION

Innovation in SMEs

TRANSCRIPT

Page 1: Inovarea in Sme

1

INNOVATION IN SME'S: THE MISSING LINK

A. Le Bars, V. Mangematin, L. Nesta

INRA/SERDUniversité Pierre Mendès France

BP 47 X38040 Grenoble Cedex 9

Tel: +33.(0)4.76.82.54.39Fax: +33.(0)4.76.82.54.55

E-mail : [email protected] or [email protected] or [email protected]

Abstract : The economics of innovation has addressed two main issues, respectively regarding the types ofknowledge (scientific, technological) required in innovation, and the organisation of R&D. Drawing heavilyon the central significance of R&D activities, such premises do not provide convincing explanations of otherkinds of innovation that do not systematically draw on R&D efforts. We argue that innovation in SME's iscritically related to individual members of the firm. Indeed, individuals embody two types of competencies,respectively component and architectural ones. While component competencies are linked to specificscientific or technological skills, architectural competencies deeply shape the innovation process byenhancing new combinations of existing component competencies. We further develop this framework byapplying it to both high tech and low tech sectors. We show that architectural competencies are essential inthe sense that they allow low tech SME's to innovate without performing internal R&D activities. While starsscientists remain central elements of architectural competencies in high tech industries, engineers appear askey drivers of new innovation projects in low tech sectors.

Key words : SME, R&D, innovation, competencies, low and high technology.

Paper presented at the Sixth Annual International Conference at the University of Twente, the Netherlands

High-Technology Small Firms Conference, 4-5 june 1998

Page 2: Inovarea in Sme

2

1 INTRODUCTION

Scholars have devoted considerable attention towards understanding the dynamics of

innovation. Since Schumpeter's seminal research (1912 and 1942) which challenges the

role of the size of the firm on innovation dynamics, different contributions have been

investigating the relationship between size and innovation. But empirical results remain

inconclusive. Large firms exhibit greater innovative efforts and seem to produce more

innovative outputs, but beyond a certain extend, this level declines.

Economists emphasise the linkages between R&D and innovation, mainly for large firms

and high tech small and medium enterprises (SME's). If not sufficient, R&D activities

systematically appear to be a necessary condition. However, empirical data show that the

level of innovation in SME's is also high in sectors in which R&D is low. In the French

agro-food industry, 70% of the 2783 firms (which employ more than 10 persons) declare to

have achieved at least one innovation in the past three years (1993 survey). Such empirical

evidences deeply question the consensual linkages established between R&D activities and

innovation. Besides, the debate remains highly relevant for SME's that often do not have

R&D structures.

To understand the dynamics of innovation, economists have generally focused on the

organisation of R&D. In large firms, the main issue concerns both communication

channels between different functions and the centralisation vs. decentralisation of R&D

activities. In small firms, such issues are no more relevant, so that economists concentrate

on the role of R&D networks (mainly for high tech SME's). But economists do not pay

much attention to the role of the entrepreneur. In this respect, the seminal emphasis given

by Schumpeter in the role of individuals has largely been ignored in economics whereas

individuals have a great impact on the innovation process per se, especially in SME's.

Schumpeter's analysis suggests that the entrepreneur can best be described as a risk taker.

However, focusing exclusively on risk-taking behaviours downplays the dynamic

capabilities of organisation embodied in the entrepreneur. Interestingly, few recent

contributions (Acs and Gifford, 1996, Gifford, 1992a) develop a capability approach to the

role of the entrepreneur.

This paper analyses the capabilities mobilised in SME's in innovation. The first part of the

paper shows that there is a missing link in innovative process if R&D are considered as the

major source of innovation. Following Henderson and Clarck (1990), the second part of

this contribution proposes to consider two different capabilities: architectural competencies

Page 3: Inovarea in Sme

3

and component competencies. Thanks to empirical case study in agro-food industry, the

paper shows that R&D is not needed to develop architectural competencies for innovation.

2 R&D – INNOVATION: THE MISSING LINK

We combine two elements in order to enlighten the missing link. First, we show that

internal R&D is not systematically a necessary condition in the innovation process,

although Cohen and Levinthal have pointed out that internal R&D capabilities remain

necessary to take advantage of external R&D. We exhibit empirical evidences of

innovation without internal research. Second, studies concerning innovation have generally

been concerned with the size of the firm. Literature on the relationship between

organisation, size and innovation describes two issues: on the one hand, some papers

emphasise the decreasing returns of size for large firms. Up to a certain size, rigidities

become too strong and innovation is not encouraged any more (Kamien and Shwartz,

1982). On the other hand, statistical works (Lhuillery, 1997) show that the propensity to do

internal research increases with size. Small firms without any R&D capacities are not taken

into account in these studies. Altogether, the missing link here lies in the neglect of other

competencies that are needed in the innovation process. Importantly, it considerably

downplays the very role of individuals, given that in SME's, such competencies primarily

lies in human beings.

2.1 Innovation without research: benefit from external research

Most of the models developed in the Economics of technical change show that R&D is the

main source of innovation. However, the repartition of R&D expenditures is highly

asymmetric. In the French innovation survey in 1993, almost 70% of agro-food firms (with

more than ten persons) declare to have achieved at least one innovation while research

expenditures represent less than 1.7% of the added value.

Inter-sectoral comparison show that large firms have a competitive advantage to innovate

in specific sector like instrumentation, automotive industry, aircraft industry while SME's

have a higher rate of innovation in other sectors, especially low tech sectors (agro-food,

clothing industry, etc.) (Acs and Audretsch, 1988). Following Frascati manual's definition,

less than 3000 firms implement R&D activities in France. These are mainly large or high

tech firms (computer, software, biotechnology, etc.). But as established by Kleinknecht

(1987), there is an obvious lack of formal R&D in SME's, especially in low tech sectors.

Page 4: Inovarea in Sme

4

This leads us to wonder about alternative or complementary sources of innovation in these

SME's. In other words, the very sources innovation remain peculiarly mysterious if R&D

expenditures are considered as the main and only input in innovation.

Based on multi-sector analysis of US data, A. Jaffe (1989) suggests that the weak linkages

between innovation and R&D expenditures in SME's are due to the collective nature of

innovation. Small firms get more benefits from external R&D than large firms, be research

activities produced by public or private institutions. Link and Rees (1990) point out one

interesting aspect regarding large and small firm research behaviours.

'Although large firms are more active in university-based research per se, small firms

appear to be able to utilise their university based associations to leverage their internal

R&D to a greater degree than large firms', (Link and Rees, 1990, p 30).

Acs et al. (1994) show that the propensity of small firm to innovate is positively correlated

with research expenditures of neighbour universities. Interestingly, this correlation is

weaker for large firms. Their propensity to innovate is correlated with private research

centres expenditures wherever research centres are located. These empirical results are

quite stimulating. On the one hand, they shed light on the complementary nature of private

and public R&D expenditures with SME propensity to innovate. On the second hand, they

represent a first explanation of the distortion between R&D expenditures and the

propensity to innovate in SME's.

Nevertheless, these results remain unsatisfactory in several disputes. First, empirical data

from the French innovation survey in the agro-food industry show that the propensity to

innovate is not correlated with public or private expenditures of geographically neighbour

organisations. Indeed, geographic areas where the rate of innovation is the highest are not

the local areas where public research institute are prominent. Do SME's in agro-food

benefit from spill-overs from other sectors? Further research need to be undertaken yet to

confirm these preliminary results. Second, they assume the central hypothesis of R&D as a

main source of innovation. Thus, they do not take into account the absorptive capacities

(Cohen and Levinthal 1990, Rosenberg 1990) developed within the firm. Next section

introduces such issues.

2.2 Is it possible to benefit from external research without internal research?

Cohen and Levinthal (1989) show that R&D investments

'develop the firm's ability to identify, assimilate, and exploit knowledge from the

Page 5: Inovarea in Sme

5

environment' (p.569)

Examining what they call the absorptive capacity of the firm, they elucidate the two faces

of R&D investments. On the one hand, firms invest in R&D to generate innovations. On

the other hand, research activities contribute to the constitution and the broadening of the

firm's absorptive capacity. Defined as a set of knowledge and competencies, the firm's

knowledge-base remains a preliminary condition in the assimilation of spillovers from

public research institute and private R&D efforts. For Rosenberg (1990), fundamental

research inside the firm has strong complementarities with external R&D from either the

public or the private sector. Altogether, both Cohen and Levinthal and Rosenberg insist on

potential synergies between the firm's own knowledge base and external flows of scientific

and technical knowledge. However, absorptive capacity is considered as a by-product of

R&D investments (Cohen and Levinthal, 1990, p.129) which implies that it cannot be built

for its own sake.

Other contributions have paid further attention to the constitution of the absorptive

capacity. Arora and Gambardella (1994) distinguish between the scientific and the

technological capabilities of a company. The former is required in the evaluation of

relevant inter-organisational alliances. It is both a means of knowledge diversification and

a phase of scientific specialisation. The latter, i.e. the technological capability, implements

such knowledge, thereafter leading to innovations. Convincingly, Arora and Gambardella

identify positive investments in R&D in the building of the firm's ability to exploit external

knowledge flows. Therefore, absorption mechanisms cannot be limited to passive attitudes

but include active processes of assimilation as well. In a similar vein, Mangematin and

Nesta (1998) show that the firm's absorptive capacity is linked to the characteristics of the

assimilated (or absorbed) knowledge. They analyse the relationship between three basic

concepts empirically defined: the fundamental or applied nature of knowledge, the tacit or

codified form of knowledge and the absorptive capacity of the firm. They show that a low

absorptive capacity inhibits co-operation in R&D. The collaboration concerns mostly

applied fields and needs informal interactions to support transfers (such as telephone calls,

informal interviews and meetings). A high absorptive capacity extends the assimilation to

all kinds of knowledge (applied, fundamental) through all types of vehicles (PhD students,

scientific papers, technical devices). Altogether, absorption mechanisms seem to diversify

as the firm’s absorption capacity increases.

In summary, innovation does exist without internal research activities. Economic theory

remains unclear on spillovers which low-tech industries benefit from. Empirical evidences

Page 6: Inovarea in Sme

6

show correlation between intensity of university research in a geographic area and

propensity to innovate whatever the economic sector. It has been shown that vectors of

absorption are different from firms with research activities and firms without. This

difference of vectors of absorption of knowledge leads to assume differences in

organisation. Mangematin and Nesta show that vehicles of knowledge transfer are more

informal (personal contacts, unpublished written notes, instruments) in low-tech sector

than in high tech sector. Even if goals are similar in low tech and high tech sectors, the

forms of produced knowledge is different: technical devices for low tech; articles, patents

and also technical devices and new material for high tech. Therefore, it seems relevant to

assume that internal organisation as well as networks in which employees of the firm are

involved are different. What role do organisational structures play in the propensity to

innovate? How are capabilities organised and developed ? And how does it affect the

propensity to innovate ?

2.3 Innovation and organisation

As it was already pointed out, investigations of the relationship between innovation and

size of the firms remain inconclusive. When the relation is empirically tested, there is little

support found for the hypothesis that the propensity of large firms to innovate is greater

than its of small firms. Cohen, Levin and Mowery (1987) find that the size of the firm (or

business units) explain a negligible fraction of the variance of R&D while industry

characteristics such as technological opportunity, Appropriability and demand conditions

are significant in explaining R&D intensity. They conclude that size can matter only in low

technological opportunity.

Acs and Audrescht (1990) define a clear research agenda when they underline that

economic theory should further address :

'how firms of varied size may be disparate innovative response to different economic

environments rather than focusing on which firm size is uniquely endowed to best promote

technological progress', p59.

Two types of works are related to this research agenda. On the one hand, Nelson and

Winter, Chandler and Nonaka (among others) pointed out the impact of internal

organisation to stimulate the creation of knowledge. These contributions mainly concern

large firms, and thus do not explicitly incorporate the role of the entrepreneur. On the other

hand, Gifford (1992b) and Acs and Gifford (1996) examine the optimal allocation of time

Page 7: Inovarea in Sme

7

for an entrepreneur between current projects and new projects in the innovation process. It

is shown that the optimal allocation of time limits the size of the firm and that the rate of

innovation depends not only on technological opportunities, firm size and monopoly

profits but also on the obsolescence degree of current innovation projects.

2.3.1 Organisation and innovation in large firm: How firms create and keep relevantknowledge

Innovation has largely been studied referring to large firms. In this perspectives one can

identify three types of approaches, respectively organisation theories, evolutionary

approaches and strategic management. Organisation theories (Cyert and March, 1963) have

advanced a picture of organisational structures by referring to the bounded character of

rationality. In this context, behaviours of individuals tend to conform to satisfacing patterns

of solutions. But facing random changes, i.e. production failures or the environment,

individuals search for solutions by a trial-error procedure. It follows that the organisational

structures can only be modified after external changes in the firm’s environment. Apart

from R&D efforts, changes can only occur in the form of behavioural modification. The

evolutionary theory of the firm as developed by Nelson and Winter (1982) draws largely

on such premises. They introduce the concept of routine to describe the regular pattern of

organisational activities. Routines are repositories of tacit knowledge and skills and are

selected over time following changes in the environment. However, search activities

implemented in the R&D function allows for the definition of new routines. In this context,

innovation is naturally derived from large corporate efforts for research activities.

In order to overcome this difficulty, Bateson (1972) and Argyris and Schön (1978)

developed the concept of organisational learning, distinguishing between single and double

loop learning. The former refers to minor modifications in the operational conduct of

production, while the latter concerns changes in the firm’s strategy, potentially leading to

organisational re-structuring. Nevertheless, double loop learning, i.e. the questioning and

re-definition of the firm’s strategic issues, can be very difficult to implement by

organisation themselves, apart from existing development program. Therefore, the creation

of new knowledge, potentially leading to innovations,

'assumes implicitly that someone (-) knows objectively the right time and method for

putting double-loop learning into practice'. (Nonaka, 1994).

Nonaka and Takeuchi’s landmark contribution (1995) has explored the very nature of

organisational knowledge creation by giving emphasises on the role of individuals played

Page 8: Inovarea in Sme

8

within the firm. They advanced the notion of the spiral of knowledge creation, where tacit

knowledge becomes explicitly expressed in objectives and organisational procedures by

means of individuals’ interactions. The authors refer to the knowledge creating crew,

where ‘knowledge practitioners (front line employees), knowledge engineers (middle

managers) and knowledge officials (top managers) interact on a day-to-day basis, or in

project teams. In this sense, the knowledge creating crew is composed of intrapeneurs that

reshape the knowledge base of the firm.

While in broad agreement with Nonaka and Takeuchi, we argue that models of

organisational knowledge creation remain concerned with large firms. Most contributions

emphasise the role of organisational structures in the process of knowledge creation.

However, the work of Nonaka and Takeuchi constitute a first step towards an

understanding of knowledge creation, potentially leading to innovation, based on

individuals.

2.3.2. The Entrepreneur as a manager of projects

The model proposed by Gifford (1992a) analyses the allocation of the entrepreneur

attention between current projects and new projects. It analyses how the limited attention

affects the entrepreneur's ability to maintain current project through product improvement

and to innovate new products. This model completes Penrose's work (1959) who views

entrepreneurial abilities as the recognition of new profit opportunities and the acquisition

of the needed resources to implement a new innovation project.

In the literature on innovation, the model of allocation of entrepreneurial attention is most

closely related to the models of decision theory. Gifford addresses the question of whether

the firm should consider the attempt of undertaking a research project at all and suppress

the characteristics of the research process itself. Constrains are given by the limited time of

the entrepreneur.

The allocation of attention provides an explanation of the varying growth rates of the firm

by modelling directly the process of growth. The model predicts a bound on firm size that

prevents loss of control. Gifford argues that the size of the firm is correlated to the degree

of obsolescence of technology. If the degree of obsolescence is sufficiently low, she

demonstrates

'a discrepancy between the largest attainable firm size under the growth process implied by

the optimal policy and a larger more valuable firm size that is unattainable through this

optimal process' (p286)

Page 9: Inovarea in Sme

9

She remarks that these results are consistent with those obtained by Lunn (1986) who

underlines that innovative effort is more product oriented than process oriented in market

with high technological opportunities.

Gifford's model of allocation of time emphasises a particular type of capabilities of

entrepreneur, notably in the evaluation of project, in the restoration of profitability of

existing projects and in the drawing of new projects. Entrepreneur capabilities are not

analysed per se, but a specific role of combination of resources is recognised to the

entrepreneur. Interestingly, the introduction of entrepreneurial abilities implicitly suggests

that the innovation process is primarily initiated by entrepreneurs and run by managerial

abilities. Rather than R&D expenditures, both abilities are confronted to the allocation of

time one may dedicate to projects.

Nevertheless, the Economics of innovation derives innovation from the creation of new

knowledge held by R&D efforts even if a specific contribution non-based on R&D is

recognised to the entrepreneur. In this perspective, innovation is not the mere fruit of an

isolated researcher, but draws heavily on scientific collaborations, collective brainstorming

and interactions among researchers. While adhering to such results, we are left with a

striking paradox. While claiming the collective character of knowledge creation, authors

mostly concentrate on the disciplinary content of new knowledge, leaving aside the very

management of the innovation process. Though innovation is not held by one single

individual, it only results from new specialised knowledge.

We argue that such paradox is due to a homogeneous conception of innovation. But

innovation can be of different nature. Recalling Henderson and Clark (1990), we focus

specifically on component vs. architectural innovations. The former is defined as

innovation by introducing a new component in a product or in a production process.

Interestingly, the latter comes from a new combination of existing components. Therefore,

innovation can be derived from existing knowledge, leaving the R&D issue to secondary

considerations.

3 COMPETENCIES FOR INNOVATION: ENTREPRENEUR AND RESEARCHERS

Analysing the innovation process requires more than the mere investigation of R&D

efforts. We follow Henderson and Cockburn (1994) who have identified different types of

competencies in the innovation process, respectively component and architectural

competencies. Because they do not serve the same purposes, both types draw on different

Page 10: Inovarea in Sme

10

resources.

3.1 Component competencies and architectural competencies

In a recent contribution, Henderson and Cockburn (1994) propose a twofold vision of

organisational capabilities that impulse the innovation process. Echoing the distinction

between architectural and component innovation made by Henderson and Clark (1990),

organisational competencies comprise both component and architectural competencies. The

former is defined as

'the possession of skill or assets specific to a particular local activities within the firm'

(p.64).

In other words, they consist of specialised areas of expertise in one or more given disciplines.

But investing pharmaceutical large firms, component competencies also concern expertise in

a given disease, thus gathering individuals from different disciplinary traditions. Through

time, the group develops tacit knowledge and competencies that become dedicated to both the

team and the concerned disease. However, and as stated above, component competencies are

to be implemented in order to perform production. Architectural competencies are then

defined as the ability to use and develop these component competencies. In this respect, they

consist of the integration of different sets of knowledge and competencies in the firm’s

production process.

Component competencies are based on technological or scientific competencies. They are

defined as special capabilities in a specific field. The strategic advantage of the firm which

gets this component competency is based on technological or scientific advances following

advances from research efforts.

Our concern is essentially devoted to the treatment of architectural competencies.

Interestingly, the authors provide further details by emphasising its dual nature. First,

architectural knowledge is defined as knowledge about

'the ways in which the components (of a product) are integrated and linked together in a

coherent whole' (Henderson and Clarck, 1990, p.11).

Such knowledge therefore enhances new combinations between components by means of

which new products can be designed and developed. The idea is indeed very similar to that

of Kogut and Zander when they define combinative capabilities as the ability

'to generate new applications from existing knowledge' (Kogut and Zander, 1992, p.391).

Thus, innovation does not necessarily emerge from heavy R&D efforts, but may be the

fruit of unexplored organisational arrangements of a given product. However, a second

item is required for knowledge to become an architectural competence. Knowledge need to

be implemented in organisational settings that implies the participation of individuals

requires communication channels that shape the development of new competencies. In this

Page 11: Inovarea in Sme

11

respect, Henderson and Cockburn refer to culture, norms and values as key elements that

shape in time the organisational structure of the firm.

The dual nature of architectural competencies is indeed very promising. While

technological combinative capabilities are key drivers of architectural innovations, the

effective integration of the existing technology draws on organisational structures to

implement idea. Besides, the dichotomy between architectural and component

competencies seems to be consistent with the definitions of Christensen and Carlsson in

the sense that competencies (or capacities in Christensen terminology) comprise both

technical/component and managerial/architectural dimensions. Grasping the intangibility

of such features (i.e. culture, norms, and value) leads to inherent difficulties. However,

Henderson and Cockburn propose several variables to measure the architectural

competencies of the firms. The first variable (PROPUB) deals with publications: do

publications play a key role in promotion within the firm? The second one (CROSS)

concerns the information flow across the boundaries of the firm. And the two last variables

relate to the management. DICTATOR represents the fact that a single individual makes

resource decisions, whereas with GLOBAL research management is an integrated whole.

The results show that pharmaceutical firms that encourage extensive flows of information

across the boundaries of the firm and across scientific disciplines and therapeutic classes

within the firm are more productive in drug discovery. Though not more efficient in the

short term, they prove higher profitability in the long run.

Henderson and Cockburn's contribution enlightens two kinds of competencies that firms

use at different levels:

(1) component competencies results from a specific effort of the firm in a specialised

technological or scientific field. Output can be measured by expenditures (number of

projects, number of researcher, R&D expenditures, etc.) devoted to applied R&D in a

specific field. It can also be approximated by the purchase of external technology

(patents, exclusive rights, etc.),

(2) architectural competencies result from the circulation of information inside and outside

the firm (PROPUB and CROSS) and from the capacity of managers to be charismatic,

i.e. to lead employees to share organisational goals (DICTATOR and GLOBAL).

Notably, this study statistically shows the significance of architectural competencies for

innovation in the pharmaceutical industry. But implemented in a science based sector, this

model need be further developed by exploring other sectors.

Insert 1 : Innovation in the food industry is not based on research

Case 1:Two individuals observing the food market have noticed that fresh department was developing

Page 12: Inovarea in Sme

12

rapidly. Particularly they have remarked that fresh paste for pie had a growing market share. Theyhad the idea to develop a complementary product for fresh paste: a fresh garniture. First, theyconfirm their market intuition with a market study. Having been realised by a firm specialised inmarketing, the study has shown that there is a potential market for fresh garniture. Second, theycreated a firm to develop, industrialise and market the product innovation. Third, they have startedthe development of the formulation not in a R&D laboratory but in their kitchen! The main problemwas related to eggs texture. They resolved it in their kitchen and with the help of an externallaboratory for analysing formulas. Finally, they contacted partners to accomplish the tasks for whichthey did not have the required competencies, such as marketing competencies for packaging design,commercial documents, etc. This project has been a success not because of a strength in R&D. It hadsucceeded because innovators had initially a global view of the innovation process, includingpotential market, formula development, industrial processes, packaging and commercialcommunication. They have managed to co-ordinate the different stages of the innovation process,without neglecting stages for which they did not have the required competencies. Instead, theydelegated to adequate partners. They developed competencies in order to select the competentpartner and manage the partenariale relation. The architectural competencies of the firm includes theco-ordination of the different stages of the innovation process and the integration of diversefunctional competencies both internal and external.

Case 2:A small firm producing cheese permanently creates new products. At the beginning, the CEO meet aresearcher of the INRA that has invented a new process for milk treatment. He had the idea to usethis process for cheese production. They developed the technology in a close partenarialerelationship. Now the firm possesses a technical competency and uses it in order to enlarge hisproduct range.Recently they have developed a new product based on this technology. This innovation seemsmainly based on technical competencies. However, a deep analyse of the innovation process showthat architectural competencies greatly participate to the success of the innovation. First, theyapplied the technology to produce the new product. Preliminary results were not satisfying. Theyhad to adapt the technology especially for the new product. As they did not have the requiredcompetencies internally to perform the adaptation, they have worked with the INRA laboratory. Thiscollaboration was due to the close relationship they kept for a long time. They managed to adapt thetechnology but they still had to find a solution concerning the structure of the product. The solutioncame from both internal competencies in production process and external competencies of themachine supplier. Moreover, during the entire innovation process the firm has tested his product ona panel of customers in order to adapt his esthetical aspect and gustative qualities. In this example,technical competencies play an important role but is clearly insufficient to innovate. Successfulinnovation require architectural competencies to co-ordinate the internal functions and the externalrelationships in a coherent whole.

3.2 Beyond Henderson and Cockburn model

Henderson and Cockburn deal with architectural competencies devoted to the management

of R&D activity. They do not take into account the different functions of the firm involved

in the innovation process such as marketing, design development, production process, etc.

3.2.1. From competencies to organise R&D to competencies for innovation

Referring the 'chain-linked model' (Kline and Rosenberg, 1986) we argue that innovating is

mainly a problem of co-ordination between technological, marketing, and organisational

constrains. The authors emphasise on the feed back links that occur between the different

stages of the innovation process. That leads to the redefinition of previous design,

integrating information coming from downstream stages. Thus, the process of innovation

Page 13: Inovarea in Sme

13

supposes the co-ordination of different functions held by the firm, enlarging the locus of

innovation to other activities than the mere function of R&D. According to Kline and

Rosenberg, when a problem arises, members of the firm first search for a solution in the

existing and accumulated knowledge base. Only if no solution can be found within the

knowledge base, the firm undertakes a research activity. Then, limiting their study to

architectural competencies devoted to research activity, Henderson and Cockburn do not

consider the innovation process as a whole, i.e. including all the functions. We propose to

enlarge the analysis of architectural competencies to the entire innovation process. That is

the architectural competencies used to manage each function of the firm and also the one

used to integrate this function in a coherent whole to innovate.

Approaches based on a complete census of resources are difficult to draw on. Indeed,

resources mobilised in the innovation process are heterogeneous and an exhaustive list of

resources may lead to endless enumeration. Instead, we choose to study architectural

competencies by characterising configuration of resources mobilised to innovate at

different stages of the innovation process.

Insert 2 : Methodological proposal to analyse the competencies for innovation

Empirical studies on innovation in high tech sectors underline the R&D competencies that aredeveloped in the innovation process. We propose to focus on innovation processes of low techfirms in order to emphasise the architectural raher than component competencies. Indeed,innovations draw largely on existing knowledge combined in innovative, i.e. new ways. Thus, wehypothesise that in low tech sectors, such as food industry, innovation is mainly based onarchitectural competencies.One may distinguish between three types of architectural competencies carried out in innovation.First, food firms seem to develop a capacity in combining existing technologies for new purposes.Second, innovation is successful when they manage to co-ordinate the different functions withinthe firm, such as marketing, production, finance … Finally, the competencies to manage therelationship with external partners play an important role in the innovation process.Because of a lack of data on innovating SME's in the food industry, we build an a priori typologyof innovation in the food industry. Interviews of ten experts of the sector have permitted to identifythree main types of innovation:• Firm's creation based on an innovation• Innovation originated by the SME• Innovation impulsed by external actors such as equipment suppliers, department stores, or

laws.The sample of 20 innovation processes is selected in order to cover the different types ofinnovation. For each case, we interview thoroughly three to four actors of the innovation processin the firm. During the interviews, we identify the main problems the firm confronts : (i) the originof the problems; (ii) key resources mobilised to solve the problems; (iii) configurations ofresources adopted; (iv) persons involved; (v) the co-ordination of individuals and ideas to solve theproblem.The aim of the empirical study is to identify the configurations of resources and competenciesfirms adopt to innovate, ie the architectural competencies for innovation.

Page 14: Inovarea in Sme

14

3.2.2. From high tech to low tech industry

Henderson and Cockburn are concerned with science based sectors. But innovation does

indeed occur in low tech industries that draw on R&D processes to a considerable lesser

extent. By applying our study to low tech industry, we thus enlighten architectural

competencies mobilised in the very process of innovation.

We argue that component and architectural competencies differ according to the

technological intensity of the industry. Competencies can best be defined as the

mobilisation of knowledge in a given context. As Foss puts it :

'By 'competence', we understand a typically idiosyncratic knowledge capital that allows its

holder to perform activities (-), and typically to do this more efficiently than others.' (p.1)

Such a definition is not necessarily dedicated to component competencies, but also

comprises architectural competencies. But a deeper view on competencies allows one to

distinguish between component and architectural competencies. While the former supposes

the mobilisation of both assets (technical devices, instruments, etc.) and knowledge

embodied in individuals, the latter more generally refers to diversified knowledge

mobilised in the process of innovation. This dichotomy is, however, of a polar nature

because it emphasises the differences between component and architectural competencies,

leaving aside the interactions that occur between component and architectural

competencies. However, these interactions are of great significance, notably in the

different types of knowledge that support innovation. Do scientific knowledge lead to

architectural competencies ? Are technological competencies exclusively devoted to

component competencies ?

We build two different typologies for respectively high and low technology firms. That is,

competencies differ with the level of R&D. The nature of the mobilised knowledge in high

technology is science-based while it remains more technology-based in low technology.

Assets in high tech are dedicated to research activities whereas in low tech, these are

dedicated to the production process. Arguably, individuals also differ with respect to the

very content of their job, e.g. engineers vs. researchers. However, we can state that

component competencies are in both cases based on specific knowledge whereas

architectural competencies are based on diversified knowledge.

Table 1 and 2 present the nature of component and architectural competencies in both low

tech and high tech industries.

Page 15: Inovarea in Sme

15

Table 1 : Component and architectural competencies in high technology SME

COMPETENCIES COMPONENT ARCHITECTURALASSETS Technological devices and dedicated

equipment to research activity-

KNOWLEDGE Specific knowledgeCodified knowledge in specificdisciplinesTacit knowledge in running experiments

Diversified disciplinary knowledgeGeneral and abstract knowledge inseveral disciplinesTacit functional knowledgeTacit knowledge in projectmanagement (including networking)

INDIVIDUALS ResearchersEngineers

Star researchersManagers

MEASUREMENTS Patents and publication in a specificscientific area

Circulation of information inside andoutside the firm

Table 2 : Component and architectural competencies in low technology SME

COMPETENCIES COMPONENT ARCHITECTURALASSETS Machinery and equipment dedicated to

production process-

KNOWLEDGE Specific knowledgeTechnological knowledgeTacit knowledge in running production(including production failures)

Diversified functional knowledgeKnowledge about other technologicaldevices that prove potentialapplicability (relationship withequipment suppliers and laboratoriesdeveloping technologies)

INDIVIDUALS EngineersTechnicians

EngineersManagers

MEASUREMENTS Patents in a specific technological areaScientific advances in market

Combination of existing technologiesto other purposes;Ability of the manager to improuveproduction process

Three important points are worth noticing :

� In low tech sectors as well as high tech, both component and architectural are needed to

innovate. The nature of component and architectural competencies and the degree of its

mobilisation are different. In high tech sectors, the creation and acquisition of

respectively new and external knowledge remains the key factor. Thus, the firm will

manage its R&D process carefully. As Arora and Gambardella point out (1994), while

some resources are devoted to the evaluation of knowledge (screening from the

environment and the scanning of projects), the main effort is devoted to develop new

technology based on scientific advances. In low industry, the screening of environment

is not an organised activity because the sources of improvement are too heterogeneous

to be systematically scanned. Equipment suppliers as well as intermediary labs are in

relation with the firm to discuss incremental innovation. Original combinations are

mainly based on trial-error procedures or on casual findings.

� In high sector, the component competencies can be identified separately to the

Page 16: Inovarea in Sme

16

production. Indeed, R&D activities allows for scientific advances, thereafter leading to

new component competencies while production deals with the very implementation of

these competencies. In low tech sectors, innovation is often based on knocked up, i.e.

improvised, spontaneous new technical devices, so that the component competencies

are not clearly identified from the production process. They are often mixed with

maintenance or quality functions or design and methods office.

� In low tech industry, the technology mobilised in the production process is simple

enough to allow engineers or managers of the firm to be involved in the formulation of

new technological combination, to set up new products and to improve production

processes. In high tech industry, architectural competencies are based on the ability to

combine different fields of science. That is the reason why firms in high tech sector are

hiring scientific 'stars', who are supposed to be able to link different scientific fields

(Zucker, et al., 1995; Zucker, et al., 1994).

4 CONCLUSION : THE PROMINENT ROLE OF INDIVIDUALS

Following Henderson and Cockburn (1994) we have shown that innovation requires two

types of competencies. Component competencies consist of specialised areas of expertise

in one or more given disciplines. They are based on technological or scientific

competencies. Architectural competencies are first composed of architectural knowledge

that enhances new combinations between components by means of which new products

can be designed. Second, knowledge need to be implemented in organisational settings that

implies the participation of individuals, requires communication channels that shape the

development of new competencies. Nevertheless, the Economics of innovation have

mainly emphasised on R&D and on component competencies. Because of the important

weight of R&D activities to innovate in high tech industries, architectural competencies are

rarely underlined. To analyse thoroughly this kind of competencies, we propose to focus

our study on low technology industries, where innovation is not science-based. We

hypothesise that architectural competencies are key component of the innovation process

both in high and low tech sectors. However, these competencies are embedded in different

types of knowledge and individuals in high or low tech industries.

This emphasises the role of individuals in the innovation process. Arguably, it implicitly

assumes that individuals do affect the knowledge base of the firm but also the conduct of

its nurturing and development. That is, in SME's, entrepreneurs not only initiate new

innovation projects, but they also shape the knowledge base of the firm by embodying

Page 17: Inovarea in Sme

17

several types of knowledge. Such remarks question the links between the individual and

organisational knowledge bases on the one hand and the initiation of new innovation

projects on the other hand.

REFERENCES

Books:

Acs, Z., and D. Audretsch, (1990) Innovation and small firms, MIT Press, Cambridge MA.

Argyris, C. and Schon, D. A. (1978) Organizational Learning : A Theory of Action, AddisonWesley Reading, Cambridge MA.

Cyert, M. and March, J. (1963) A Behavioral Theory of the Firm, Englewood Cliffs, NewJersey, Prentice Hall.

Kamien, M. and Shwartz, N. (1982) Market structure and innovation, Cambridge UniversityPress, Cambridge Massachusetts.

Penrose, E. (1959) The theory of the growth of the firm, Basic Blackwell, Oxford.

Nelson, R. R., and S. G. Winter, (1982) An Evolutionary Theory of Economic Change, HavardUniversity Press.

Nonaka, I. and Takeuchi, H. (1995) The Knowlegde Creating Company : How JapaneseCompanies Create the Dynamics of Innovation, Oxford University Press, New York.

Chapter in a book:

Christensen, J.F. (1996), 'Analysing the technology base of the firm : a multi-dimensionalresource and competence perspective', in Foss, N.J. and Knudsen, C. (eds.) Towards a CompetenceTheory of the Firm, Routledge, London.

Foss, N.J. (1996), 'Introduction : The emerging competence perspective', in Foss, N.J. andKnudsen, C. (eds.) Towards a Competence Theory of the Firm, Routledge, London.

Kline, S. J., and Rosenberg, N. (1986), 'An overview of innovation', in Landau, R. andRosenberg. N. (eds), The Positive Sum Strategy, Academy of Engineering Press.

Article in journal:

Acs, Z., et. al. (1994) R&D spillovers and the recipient firm size, Review of economics andstatistics, 76(2), pp.336-340.

Acs, Z. and Gifford, S. (1996) Innovation of Entrepreneurial Firms, Small Business Economics,8, pp.203-218.

Acs, Z. and Audretsch, D. (1988) Innovation in Large and Small Firms: An Empirical Analysis,American Economic Review, 78(4), pp.678-90.

Arora, A., and Gambardella, A. (1994) Evaluating technological information and utilizing it,Journal of Economic Behavior and Organization, 24, pp.91-114.

Page 18: Inovarea in Sme

18

Cohen, W. M., et. al. (1987) Firm size and R&D intensity: a Re-examination, Journal ofIndustrial Economics, 35,pp.543-565.

Cohen, W. M. and Levinthal, D. A. (1989) Innovation and Learning : The Two Faces of R&D,Economic Journal, 99, pp.569-596.

Cohen, W. M. and Levinthal, D. A. (1990) Absorbtive capacity, a new perspective of learningand innovation, Administrative Science Quaterly, 35, pp.128-152.

Gifford, S. (1992a) Heterogeneous Ability, Career Choice and Firm Size, Small BusinessEconomics, 5, pp.249-259.

Gifford, S. (1992b) Innovation, firm size and growth in a centralized organization, RANDjournal of economics, 23(2), pp.284-298.

Henderson, R. and Clarck, K. (1990) Architectural innovation : the reconfiguration of existingproduct technologies and the failure of established firms, Administrative Science Quaterly, 35,pp.9-30.

Henderson, R. and Cockburn, I. (1994) Measuring competence? Exploring firm effects inpharmaceutical research, Strategic Management Journal, 15, pp.63-84.

Jaffe, A. (1989) Real effects of academic research, American Economic Review, 79(5), pp.958-970.

Kleinknecht, A. (1987) Measuring R&D in small firms : How much are we missing ?, Journalof industrial economics, 36, pp.253-256.

Kogut, B., and U. Zander, (1992) Knowledge of the firm, combinative capabilities, and thereplication of technology, Organization Science, 3, 3 : 383-397.

Lhuillery, S., (1997), An Empirical Analysis of R&D Transactions : surrogate TechnologicalNetworks and firms Performance, Economy, Innovation and New Technology, 4 : 301-319.

Link, A. and Rees, J. (1990) Firm size, university based research and the return to R&D, Smallbusiness economics, 2(1), pp.25-32.

Lunn, J. (1986) An empirical Analysis of Process and product Patenting : A SimultaneousEquation Framework, Journal of Industrial Economics, 34, pp.319-330.

Mangematin, V. and Nesta, L. (1998) What kind of knowledge can a firm absorb?, InternationalJournal of Technology Management, forthcoming .

Nonaka, I. (1994) A Dynamic Theory of Organizational Knowledge Creation, OrganizationScience, 5(1), pp.14-37.

Rosenberg, N. (1990) Why companies do basic research with their own money ?, ResearchPolicy, 19, pp.165-174.

Research papers:

Zucker, L. et al. (1995) Collaboration structure and information dilemmas in biotechnology:Organisational boundaries as trust production, working paper, NBER.

Zucker, L., M. R. Darby, and J. Armstrong, (1994), Intellectual Capital and the Firm: TheTechnology of Geographically Localized Knowledge Spillovers, working paper, NBER