No title

1. Introduction

The distribution model in the diffusion of innovations: a comparison of different European countries

The importance of reliable predictions for the successful diffusion of a new product is a key factor in the strategic planning of a company and; as such, justifies the many contributions offered from different fields, such as industrial economy, strategic management and marketing. In fact, the diffusion process can well prevail over the innovation itself, since its economic and social impact is generated by the adopters of the innovation, which, on the other hand, becomes an important stimulus for new innovations. This has generated countless studies that have attempted to design a time model of the spread of an innovation at the aggregated level (Mahajan and Muller, 1979). The diffusion of an innovation is defined as the process by which it is transmitted, over time, throughout the members of a given social system, by means of certain communication channels (Rogers, 1983). Although diffusion is directly or indirectly affected by many different factors, such as the type of innovation that it happens to be, the communication channels employed, either inter-personal and/or mass-media, the type of social system involved, and the time, the very notion of diffusion has come to mean, essentially, a form of communication. This has meant that diffusion has come to be considered as the propagation of messages related to new ideas that lead to subsequent innovations (products, processes, technology, etc.), awaiting a change in the behaviour of the receiver, which will be evident in the adoption or the rejection of the innovation. The underlying behavioural theory suggests that a time-lag exists during the adoption period among the different members of a social system. In a first stage of the diffusion process, the new product is discovered and adopted by a small group of innovative consumers, who, with time, begin to influence others, the imitators. This social interaction between the adopting pioneers and the different potentials of the innovation explains the phase of rapid expansion in the diffusion process (Rogers, 1983). The modelling of diffusion to include a description of the adoption process of new

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz The authors Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz are both based at the University of Alicante, Alicante, Spain. Keywords Innovation, Film industry, Model, Diffusion Abstract The objective of this study was two fold: first, we wanted to test the goodness of fit of different diffusion models that include the variable ``distribution'', and then analyze the existence of the effects of ``country'' and ``time'' in the parameters of diffusion currently employed in different European countries. The literature on the matter suggests, on the one hand, that the diffusion models applied so far that employ multi-equation systems are not obviously better than other single-equation models that ignore marketing variables; and, on the other hand, that the socio-economic environment and the time-lag are both factors of the differences that the diffusion process offers among different geographical areas. As a novelty, our methodology proposes single-equation models that include the distribution, to which several indicators of the goodness of fit are applied, along with statistical tests on the differences in parameters among countries. The empirical application carried out in Spain, France and Italy demonstrates the superiority of the model proposed here, based on the Generalized Bass model. It also indicates a ``country'' effect, between Spain and France as well as between Italy and France. Electronic access The research register for this journal is available at http://www.mcbup.com/research_registers The current issue and full text archive of this journal is available at http://www.emerald-library.com/ft European Journal of Innovation Management Volume 4 . Number 1 . 2001 . pp. 6±19 # MCB University Press . ISSN 1460-1060

The authors wish to thank Professor Dipak C. Jain of Northwestern University for his help.

6

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

products has gradually been incorporated into the basic models (essentially timeseries), different factors to improve on the original model. They have now been so enlarged that some models incorporate, among other aspects, variables of marketing. In the particular case of the distribution of the product, the extensions propose modelling the impact that the number of distributors or intermediaries would have on the growth in sales. An examination of the literature shows, however, that there are important differences among the methods they propose. Jones and Ritz (1991), for example, integrate the two processes they consider implicit in reflecting the distribution within the adoption process of a product (the retailers and the consumers) into a system of equations. However, their proposed model, proves to be less accurate than other models that do not incorporate variables of marketing factors, such as the Bass model (Bass, 1969) and the NUI model (Easingwood et al., 1983). In contrast, Mesak (1996) offers a single-equation model that reflects the diffusion process only at the consumer's level, but by combining the effects of three different dimensions: price, advertising and distribution, his results seem to be more accurate than those of the original Bass model (1969), although they have not identified the individual effect of the distribution. Furthermore, the existing literature has focused on the space aspects of the diffusion process, and on the socio-economic factors that influence it. In fact, the social, economic, political, demographic and cultural differences that exist between different countries may well influence the form the adoption process takes in circulating among the majority of the consumers. In other words, with time, these factors might well condition the diffusion of certain new products or processes. The knowledge and understanding of this phenomenon must obviously be of paramount importance to the deciders, as it could give a clear indication as to how an innovation should be introduced into each country, which, in turn, would facilitate its commercialization in other countries or regions of similar characteristics. Researchers, such as Gatignon et al. (1989), Takada and Jain (1991), Helsen et al. (1993),

Redmond (1994) and Kumar et al. (1998), have all shown interest in the geographical aspect, examining the differences in the diffusion parameters for consumer durable products among different countries. They conclude that the differences in the adoption processes and, therefore, in the diffusion parameters may be explained by specific factors in each country that are beyond the companies' control: cosmopolitan populations, geographical mobility, the percentage of women in the workplace, cultural levels, prosperity and life-styles (i.e. the ``country'' effect), as well as the time-lag that always exists between the moment an innovation is introduced into its own pioneering country and when it is finally introduced into a late-coming country (i.e. the ``time'' effect). Faced with such dilemmas, we aimed at a twofold objective for our study. On the one hand, we wished to verify the goodness of fit for the different single-equation diffusion models that consider distribution to be the only marketing variable; and, on the other hand, we wanted to be able to demonstrate the existence of the effects of ``country'' and ``time'' in the diffusion parameters employed in different European countries. The empirical application was carried out for a group of films distributed in Spain, France and Italy, between 1997 and 1999. To achieve our proposed objectives, we found it convenient to employ the following sequence: in the section that follows, we review the existing diffusion models reported in the empirical literature on the matter. The third section justifies the analyzing method we propose here. The fourth section defines the sample, the collection of data, and the measuring of the different variables. The results obtained are described in the fifth section, and finally, our conclusions are presented.

2. Modelling the diffusion of innovations The original models of innovation diffusion attempt to describe the evolution over time of the demand's behaviour regarding a new product. Fourt and Woodlock (1960) explain it as a coefficient of external influences (or coefficient of innovation), created jointly by the individual's intrinsic tendency to adopt 7

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

innovations and partly by the influence of an external source (the mass media for example), while Mansfield (1961) proposes a coefficient of internal influences (or coefficient of imitation) as a result of their personal contact with previous adopters. Both elements are considered simultaneously in the model proposed by Bass (1969), a joint influence whose basic concept is that the conditional probability that an adoption will be made at time t, given that an adoption has not yet been made, is a linear function of the number of previous adopters and is expressed as f …t†=…1 ÿ F…t†† ˆ p ‡ qF…t†, where the random variable t denotes the moment of adoption of a new product by an individual (adopter), p and q are the coefficients of innovation and imitation respectively, f(t) is the adoption probability at moment t, and F(t) is the function of accumulated distribution. Denoting by M the potential market of adopters, by n(t) the nonaccumulated number of adopters at moment t (n(t) = Mf(t)), and by N(t) the accumulated number of adopters up to the moment t without including t (N(t) = MF(t)), the Bass model is expressed, in its common form, as:    q …1† n…t† ˆ p ‡ N…t† M ÿ N…t† M

and Yamada, 1988); or a combination of certain dimensions, such as price and advertising (Kalish, 1985; Thompson and Teng, 1984; Jain, 1992; Bass et al., 1994), as well as price, advertising and distribution (Mesak, 1996). Distribution is one of the aspects of a product's commercialization process to which the literature on diffusion has devoted very little attention, due mainly to the lack of available sources of information on the matter. Furthermore, in the only two papers that we have been able to find on the topic, their authors do not agree on the type of model that should be used to evaluate the impact that the distribution has on the diffusion process. Jones and Ritz (1991) take the three-stage diffusion process as their starting point ± the untapped market, the potential market and the current market ± described by Mahajan and Muller (1979), proposing an equation system in which two different diffusion processes interact, one through the intermediaries (producer to intermediaries), and the other among the consumers (intermediaries to consumers), since the consumers cannot adopt the product if the intermediaries do not offer it to them. The system presents the middleman diffusion process with a modified Bass model, and that of consumer diffusion with a constant transfer ratio. The above-mentioned authors conclude, however, that their model achieves a lower degree of fit than other models that only reflect a single consumer process (producer to consumers). However, Mesak (1996) employs the Bass mixed-influence model; in other words, he applies a single-equation model that reflects a single consumer diffusion process. The proposal introduces the pricing, the advertising and the distribution simultaneously, but it does not identify the specific effect of any dimension individually. The results achieved, however, show its clear advantage over the original Bass model (Bass, 1969). More recently, new research projects have been focusing on the differences in diffusion processes observed among certain countries or geographical areas. The influence of their different environments is highlighted by authors like Gatignon et al. (1989), Takada and Jain (1991) and Helsen et al. (1993), who all agree that an innovation spreads in

where ‰p ‡ …q=M†N…t†Š measures the diffusion effect[1] and ‰M ÿ N…t†Š the saturation effect. Although the Bass model has been considered an empirical marketing generalization (Mahajan et al., 1995), it has been criticized for its simplification of reality, as it does not consider other attributes that also exert some influence on the consumers. The models later incorporate different variables in a rather explicit way, especially those based on marketing decisions, such as distribution (Jones and Ritz, 1991), pricing (Robinson and Lakhani, 1975; Bass, 1980; Jeuland, 1981; Bass and Bultez, 1982; Jorgersen, 1983; Kalish, 1983; Kamakura and Balasubramanian, 1988; Horsky, 1990; Jain and Rao, 1990; Parker, 1992; and Mesak and Berg, 1995); advertising (Dodson and Muller, 1978; Horsky and Simon, 1983; Teng and Thompson, 1983; Simon and Sebastian, 1987; Dockner and Jorgensen, 1988; Chatterjee and Eliashberg, 1990; Jeuland, 1993), and the personal sale (Rao 8

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

In stage one, therefore, the estimates of three single-equation diffusion models, which all include the distribution variable, are compared jointly to the pioneer Bass model (1969), and their goodness of fit is evaluated with the coefficient of determination (R2) and the sum of squared residuals (SSR). We specifically consider the Bass (1969) three-parameter model, and three other proposals based on the work of Jain and Rao (1990), and on Bass et al. (1994) which relax some of the restrictive suppositions of the Bass model by introducing additional distribution parameters in their specifications. The first model to be applied is that of Bass (1969) (Model 1), which will be taken as a reference since it is considered as the most parsimonious aggregate diffusion model suggested in marketing literature (Parker, 1994), it is widely accepted (Mahajan et al., 1993; Sultan et al., 1990), and it usually provides a good fit to the adoption data (Mahajan et al., 1995). Taking equation (1) as our starting point, this model can be expressed as follows:

different ways among different cultures, depending on their socio-cultural and socio-economic environments (Redmond, 1994). On the other hand, and considering that time is one of the main factors in the diffusion process (Rogers, 1983), these marketing researchers focus on the geographical aspects and also examine how the time-lags that necessarily occur among the different moments of an innovation's introduction into different countries affect the diffusion phenomenon. There is no consensus among the authors, however, regarding the extent of such an influence, and no clear evidence exists as yet to suggest that the experience of the adopters in the pioneer country could help the potential adopters in other countries to be more assured about the success of the innovation in their own countries. Authors like Takada and Jain (1991), Mahajan and Muller (1994), and Kumar et al. (1998), observe that the longer it takes to introduce an innovation into a country, the quicker the ultimate adoption process will be (a positive relationship between the diffusion parameters and the time-lag). But Helsen et al. (1993), demonstrate quite the opposite (a negative relationship), implying that the fact that an innovation has been introduced in a ``pioneer'' country, produces a slowing effect on the diffusion processes in the belated countries. In summary then, our paper proposes, as a novelty, contrasting the goodness of fit of different diffusion models that introduce the distribution variable, using a single-equation model instead of the system of equations proposed by Jones and Ritz (1991), while we also verify the existence of ``country'' and ``time'' effects in the diffusion parameters of certain European countries.

f …t† ˆ ‰p ‡ qF…t†Š‰1 ÿ F…t†Š N…t† ˆ M:F…t†; F…t† ˆ

1 ÿ eÿ…p‡q†t : 1 ‡ pq eÿ…p‡q†t

…2† …3†

We make two alternative proposals (Models 2 and 3), that attempt to integrate the distribution into the diffusion. Both of these proposals are based on the work of Jain and Rao (1990), which relaxes the limitations of the Bass model by introducing marketing variables and dynamism in the potential market. The above-mentioned authors consider that, if ‰F…ti † ÿ F…tiÿ1 †Š is the probability that an individual (randomly chosen) adopts the new product within the time interval …tiÿ1 ; ti †, and if

3. Methodology

‰F…ti † ÿ F…tiÿ1 †Š ‰1 ÿ F…tiÿ1 †Š

The methodology we developed to achieve our objectives is composed of the following stages: verifying the goodness fit of different innovation diffusion models that include the distribution, and comparing the diffusion parameters for different European countries, so that we might be able to verify the hypotheses on the influence that a country's peculiar characteristics and the moment of an innovation's introduction have on the diffusion processes.

…4†

represents the conditional probability of and individual adopting within the interval …tiÿ1 ; ti †, although he has not yet adopted it at the moment tiÿ1 , then:  F…ti † ÿ F…tiÿ1 † Š ‡ uti …5† Sti ˆ …M ÿ Xtiÿ1 † 1 ÿ F…tiÿ1 † where Sti represents the sales during the interval …tiÿ1 ; ti †; Xti ÿ1 the total number of 9

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

adopters by the moment tiÿ1 ; and …M ÿ Xti ÿ1 † is the potential market at moment ti . Since a product cannot be bought unless it is made available to the consumer, the size of the potential market is directly influenced by the number of retailers who offer the new product to the public[2]. We therefore outline Proposal I, where the number of intermediaries who offer the product (I) affects the potential market (M) as follows: F…ti † ÿ F…tiÿ1 † Š ‡ vti …6† Sti ˆ …MIti ÿ Xtiÿ1 †‰ 1 ÿ F…tiÿ1 † and Proposal II, where the number of retailers who offer the product affects the effective potential market …M ÿ Xti ÿ1 †[3]: Sti ˆ …M ÿ Xtiÿ1 †It i ‰

F…ti † ÿ F…tiÿ1 † Š ‡ wti …7† 1 ÿ F…tiÿ1 †

 and being the intermediation parameters, and vti and wti the error terms, with an average of 0 and variances of 2v and 2w respectively. Finally, the fourth proposal of our study (Model 4) is derived from the Generalized Bass model (GBM) (Bass et al., 1994). The GBM is a flexible model[4], which is simple to apply and presents the Bass model (1969) as special case that occurs under given conditions, and incorporates the commercial variables of ``price'' and ``advertising'', and has the following structure: f …t† ˆ …a ‡ bF…t††…v…t††: ‰1 ÿ F…t†Š

…8†

Although the term v(t) reflects current marketing efforts, it can also reflect the retarded effect of the variables of decision. If the price and the advertising are considered at moment t ± P(t) and A(t), respectively ±, v(t) is a function of the percentages of change of such variables: v…t† ˆ 1 ‡ 1 ‰4P…t†= P…t ÿ 1†Š ‡ 2 ‰4A…t†=A…t ÿ 1†Š. For our particular case, the distribution in terms of retailers of the new product (I), the above expression would be expressed as: v…t† ˆ 1 ‡ ‰4I…t†=I…t ÿ 1†Š

moment of the innovation's introduction could have on the speed of its adoption in these different countries. In each moment, the geographical comparisons will be made in terms of the internal influence coefficient or imitation, since, on the one hand, the number of imitators of an innovation is generally much greater than the number of innovators and, on the other hand, the imitators are those which influence the diffusion curve most, through social interaction (Takada and Jain, 1991; Redmond, 1994). Regarding the influence of the specific characteristics of each country (the ``country'' effect), we verified whether or not the coefficient of internal influence q varies among the geographical areas we examined, (H1). This was done by an analysis of variance test for all of the countries considered, jointly, and by statistical tests for differences between pairs of countries. With regard to the effect that the moment of the innovation's introduction has on the speed of its adoption in different countries, (the ``time'' effect), H2 must be proven: that the delay in introducing an innovation into a market affects the speed of its adoption process. To verify this, the following regression model y ij k ˆ  ‡ xijk ‡ ijk

…10†

proposed by Takada and Jain (1991) is estimated, in which yijk and xijk are the differences in the values of the imitation coefficients and introduction years for a pair of two countries i and j for product k, respectively;  and are the coefficients of the regression; and ijk is the random disturbance.

4. The sample, data and the measuring of variables We shall now present the development of the methodological approach presented in the previous epigraph, in the specific case of films, which has proven to be an interesting product to analyze according to the objectives proposed by Jones and Ritz (1991). The film is a product that allows us to apply diffusion models of mixed influence (internal and external), since almost any individual is influenced as much by the comments the people closest to him/her (family, friends and work-mates) make about the films that are

…9†

denoting the intermediation parameter by . Its expected value is positive, since the increase in the number of retailers favours the diffusion process. In the second stage, the possible differences that might exist in the diffusion parameters among European countries that are geographically close to one another is examined, as well as the effect that the 10

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

currently being shown at the local cinemas (internal influence) as he/she is by the constant promotion that such films are given in the mass media (external influence). Our study was focused on three European countries ± Spain, France and Italy ± between September 1997 and March 1999, and we examined a total of 21 new films that made their premieÁre in Spain, eight of them in France, and nine in Italy (see in Table I). The films were chosen according to their duration in the cinemas, and a minimum of six weeks was considered to be the qualifying yardstick[5]. As might have been expected, the films we examined all had a rather short lifespan during their deÂbut, in all of the three countries studied, although their duration differed from one country to another. Owing to a lack of available information about certain films and/or countries, we have not been able to extend our study to other films and other geographical areas. The variable ``commercial distribution'' is defined as the number of retailers or cinema halls that project the films examined; obtained information of the magazine Variety,

a US publication that specializes in the film industry and provides weekly data on the number of cinemas at which a film is being shown and the box-office billings, according to geographical areas. Finally, the number of spectators has been estimated, based on the weekly revenues of the box-offices for each film, as well as the average price of an entrance ticket, obtained through the Institute of Cinema and Audiovisual Arts (ICAA) in Spain and from the different embassies.

5. Empirical results 5.1. Testing the goodness of fit of the diffusion models In this section, the mixed-influence models previously mentioned are considered for the 21 films that were selected in Spain (see Table II), with a view to discovering which of them provides the greatest level of accuracy. The non-linearity seen in the models has led us to use non-linear estimate procedures (NLS)[6] (Jain and Rao, 1990), starting with

Table I Films analyzed by country Code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Title

The Girl of Your Dreams The Mask of Zorro P. Tinto's Miracle Mulan There's Something about Mary Saving Private Ryan Six Days, Seven Nights Armageddon Deep Impact The Big Lewoski Torrente El Brazo Tonto De La Ley As Good as It Gets The Man in the Iron Mask The Full Monty Open Your Eyes Seven Years in Tibet Hercules The Truman Show Blade The Horse Whisperer The Jackal

Spain DeÂbut Weeksa

France Weeks DeÂbut

Italy Weeks

DeÂbut

15 10 7 7 10

15/11/98 29/11/98 20/12/98 22/11/98 08/11/98

± 7 ± 7 9

± 20/10/98 ± 01/12/98 17/11/98

±

±

± ± 8

± ± 22/09/98

8 8 8 7 7 15

20/09/98 16/08/98 19/07/98 17/05/98 17/05/98 15/03/98

8 ± 9 6 ± ±

06/10/98 ± 11/08/98 02/06/98 ± ±

7 ± ± 11 13 10

05/11/98 ± ± 21/05/98 07/05/98 23/07/98

7 9 19 8 7 7 7 7 7 7

01/03/98 12/04/98 12/09/97 21/12/97 07/12/97 23/11/97 01/11/98 11/09/98 04/09/98 25/01/98

± 6 13 ± 9 ± ± ± 10 ±

± 07/04/98 28/10/97 ± 02/12/97 ± ± ± 08/09/98 ±

± 7 10 ± ± ± 6 ± 6 ±

± 02/04/98 19/03/98 ± ± ± 12/11/98 ± 22/11/98 ±

Notes: a Weeks of duration; b Insufficient information or none available Source: Variety magazine (1997, 1998, 1999)

11

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

Table II Parameter estimates for Spain Film (code) 1

2

3

Weeks

Model

p

q

M (103)

15

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

0.0878*** 0.0869*** 0.0091 0.0867*** 0.1332*** 0.1327*** 0.1595** 0.1326*** 0.2389** 0.2325** Do not converge Do not converge 0.0951** 0.0856 0.0150 0.0909** 0.1063*** 0.0928** 0.0002 0.1009*** 0.1498*** 0.1432** 0.0655 0.1486** 0.1698*** 0.1602*** 0.1922 0.1680** 0.1746*** 0.1759** 0.2500** 0.1679** 0.2367*** 0.2325** Do not converge 0.2380** 0.1171** Do not converge 0.2061*** 0.1062** 0.0916*** 0.0867*** 0.2000** 0.0911*** 0.0357 Do not converge Do not converge 0.0389 0.2361*** 0.2283** 0.0470 0.2320**

0.0943 0.0883 0.0653 0.0915* 0.4197*** 0.4352** 0.4842** 0.4089*** 0.2479 0.6121**

2,854*** 2,573 2,829*** 2,902** 4,325*** 4,767** 4,353*** 4,396*** 1,082*** 8,176

±0.0197 ±0.0391 1.3464 94.11 0.6121**

0.3009 0.2920 0.2450 0.3475 0.2428 0.1345 0.1062 0.1978* 0.5071** 0.6730** 3.9825 0.4969** 0.4790** 0.6892** 2.2537 0.4750** 0.4542** 0.5218** 2.0569** 0.4789** 0.3974** 0.5272**

10

7

4

7

5

10

6

7

8

9

10

11

8

8

8

7

7

15

12

7

13

9

R2

SSR (106)

83.40 83.40 83.74 85.14 97.37 97.42 97.43 97.54 2,350 98.51

11,900 11,900 11,700 10,600 19,600 19,300 19,200 18,400

3,026** 23,068 2,538 2,985** 3,818*** 1,153 4,435 4,153*** 2,632*** 9,265 2,827*** 2,656*** 2,822*** 14,489 2,989*** 2,844** 2,583*** 5,672 2,715*** 2,608*** 1,905*** 4,529

52.10 ±0.3643 53.03 0.4030 59.11 ±1.2320 59.13 87.43 29,000 0.2602 84.43 1.1688 88.02 0.8375 86.26 91.83 22,800 ±0.2468 93.70 ±0.2281*** 98.75 0.7655 91.89 94.33 19,400 ±0.3184* 97.38 ±0.2060** 99.20 0.9800 94.40 95.64 12,100 ±0.1480 96.18 ±0.1907*** 99.70 1.5265 96.35 96.84 5,460 ±0.1735 98.26

18,200 17,900 15,560 15,600

0.4060 0.4586**

1,888** 562***

±1.1150 90.20

96.95 265

5,260

1.4472** 0.4340** 0.2559*** 0.3257*** 0.7192** 0.2542*** 0.1282

655*** 575*** 3,006*** 9,212 3,149*** 3,011*** 7,723

±0.2881 99.20 0.7828 93.96 89.96 23,100 ±0.2405 91.89 ±0.2365** 93.10 0.2945 89.97 69.42 4,670

30 225

0.1356 0.4023** 0.5593** 5.2210 0.3970

6,968 2,040*** 3,094** 2,080** 2,065**

±0.8934 69.93 94.12 18,700 ±0.0895 95.35 ±0.1940** 94.32 1.2120 97.67

12

d, g or b 0.0236 0.4443 0.7387

596

22,200 22,000 25,400 17,600 3,480 22,600 9,000 2,750 19,120 10,600 227 10,200 3,010

18,700 15,900 23,100

4,600 14,800 18,058 7,391 (continued)

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

Table II Film (code) 14

15

16

17

18

19

20

21

Weeks

Model

p

q

M (103)

19

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

0.0482*** 0.0373*** 0.0597*** 0.0320*** 0.1096** 0.0927 0.0248 0.0722** 0.2171*** 0.2120** Do not converge 0.2093** 0.0778** 0.1134 0.0290 0.0777* 0.2520** 0.2445** Do not converge 0.2510** 0.2165** 0.2121** Do not converge Do not converge 0.1458** 0.1327** 0.2085** 0.1470** 0.2298** 0.2237** Do not converge 0.2300**

0.2512*** 0.4155*** 0.8167*** 0.2804*** 0.5777** 0.7474 3.9842** 0.5190** 0.4902 0.7068

3,257*** 5,832*** 3,364*** 3,299*** 1,371*** 4,241 1,540*** 1,435*** 1,919*** 3,138

83.98 28,300 ±0.1513*** 89.76 ±0.2641*** 93.13 0.8784** 90.45 77.02 15,500 ±0.2453 78.06 ±0.2765 99.65 1.6688* 94.49 91.74 16,900 ±0.1126 92.87

0.5151 0.4848* 0.3220 0.2800 0.4677 0.4196 0.6621**

1,949** 2,155*** 497 2,033** 2,175** 1,556*** 3,132**

0.7579 63.22 0.2911 0.2710 0.0528 94.90 ±0.1475

92.30 19,700 68.95 68.38 63.38 6,940 97.91

15,800

0.4250** 0.5659** 0.6543

1,557** 1,724*** 2,284

0.1320 94.58 ±0.0594

94.91 10,300 94.98

6,926

0.6054** 0.7808** 1.3048** 0.5820** 0.4255** 0.6042**

1,069*** 2,442* 1,120*** 1,076*** 1,868*** 5,440

96.26 ±0.1810 ±0.1539** 0.2057 96.33 ±0.2224

1,510 98.38 99.57 96.39 6,160 98.34

0.4260**

1,868**

0.0050**

96.33

8

7

7

7

7

7

7

d, g or b

R2

SSR (106) 18,100 12,200 16,900 14,800 233 3,710 14,600

16,600 16,930 19,600 2,840

9,510

655 173 1,460 2,780 6,160

Notes: ***: prob.  0.001; **: prob.  0.05; *: prob.  0.09 Source: Authors' own estimates

66 per cent, 68 per cent, 50 per cent and 63 per cent of the films analyzed by Models 1, 2, 3 and 4 respectively. That is to say, the significance of the parameters surpasses 50 per cent in all of the models, except for the coefficient p in Model 3. These results confirm the importance of the two sources of communication, external and internal, in the diffusion of this product. In other words, when the Spanish public has to choose one of the films analyzed in the sample, they are influenced by both things, the comments of their acquaintances, who have already seen the films, and the commercial promotion done by the companies involved in their distribution.

the solutions to the differential equations for model 1 equation (2), and Model 4 equation (8), while Models 2 and 3 have been estimated directly from equations (6) and (7), respectively. This implies the estimating of three parameters (p, q, M) for all the models and, also, one of the parameters ; or for each of the models 2, 3 and 4 respectively. Regarding the significance of the parameters, the coefficient of external influence p is significant in more than 95 per cent, 84 per cent, 42 per cent and 94 per cent of the films examined by Models 1, 2, 3 and 4 respectively, while the coefficient of internal influence q is significant in more than 13

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

Regarding the evaluation of the goodness of fit, the coefficient of determination (R2) indicates that, in global terms, the models provide an acceptable level of precision with regard to the data. So that the R2 are higher than 90 per cent in more than 67 per cent of the cases (49 of a total of 73). No model, however, seems to be better than the others for all the films examined. Identical results are obtained when we examine the SSR. Focusing our scrutiny, however, on the 13 films, (1, 2, 4, 5, 6, 7, 8, 11, 13, 14, 15, 17 and 20), for which all the models have been able to be estimated, the combination of both indicators of goodness of fit shows that Models 2, 3 and 4 are slightly better than the original Bass model (Model 1), in presenting greater coefficients of determination and a smaller squared sum of errors. These results suggest that the inclusion of the number of cinema halls (variable distribution) in the proposed models improves their fit to the adoption data. In order to verify whether Model 1 also gives poorer results for other countries, the four models were later considered for the six films that were shown in all of the three countries ± Spain, France and Italy ± (see Table III). Observing the goodness of fit, it is obvious that, as in the previous case, all of the models offer an acceptable level of precision, since the coefficient of determination surpasses 90 per cent in more than 75.4 per cent (46 out of 61) of the cases. Again, none of the models seems to have performed better than the rest regarding any of the films analyzed. Similar results are also obtained with the SSR. When we combine both indicators, however, Model 4 gives the greatest coefficients of determination and the least sum of squared residuals in the majority of the cases, which is quite the opposite in the case of Model 1. Considering the superior performance of Model 4 in all three countries, the analyses in Epigraph 5.2 were taken as a starting point. For this particular case, the estimates of the coefficient of external influence p are significant for 16 of the 18 films analyzed (88.9 per cent), and those of internal influence q are significant for 13 (72.2 per cent), while those of the intermediation parameter  are only significant for six films (33.3 per cent). Their positive values are to be expected, since the diffusion process of these films was enhanced 14

by longer exposure to inter-personal and mass communication, and by the increase in the number of cinema halls. Although these results underline the great influence of internal and external sources, we have not obtained sufficient statistical evidence, however, to allow us to arrive at any firm conclusions about the influence of the intermediation coefficient on the diffusion. 5.2. Contrast of diffusion parameters among countries In this section, we analyze the existence of significant differences among the diffusion processes of the six films shown in the three European countries (Spain, France and Italy), and the effect of the moment when the novelty is introduced on the adoption speed, in the three different countries. To do so, the estimates for the coefficient of internal influence q (taken from Model 4) have been used for the three countries (see Table III). In fact, an examination of averages (F = 1.0181, prob. = 0.3850) demonstrates the similarity among the averages of the external influence parameters p, for all of the countries. A. The ``country'' effect To verify whether there were any specific factors about a country that could significantly influence the consumers' behaviour regarding the choice of the same group of films, we performed two different sorts of contrast, a global one for the three countries together, and separate ones for each pair of countries. At global level, the analysis of the variance gives a value for the statistical F of 13.966 (prob. = 0.0004), indicating that H1 (that the coefficients of internal influence vary among countries) is confirmed, with a confidence level of 99 per cent. The ``country'' effect has obviously been confirmed. In other words, the peculiar characteristics of each of the countries analyzed have caused differences among their consumers' behaviour patterns and, by extension, among the diffusion processes of the different films. To verify whether the differences in the mean values of the coefficient q is significant when the countries are examined two by two, the contrasts seen (see Table IV) indicate that they are significant between the pairs SpainFrance and France-Italy, with a confidence level of 95 per cent. The differences are not

15

7

20

10

13

6

6

8

9

WF

6

10

7

11

7

8

WI

0.1062

0.1978c

0.5071b

0.6730b

3.9825

0.4969b

b

0.1345

0.0928b

0.1498a

0.1432b

0.0655

0.1486b

a

0.0002

0.1009a

2

3

4

1

2

3

4

0.0680

0.1106c

±0.1070

0.4023

a

5.2210

0.3970

0.2512

0.2283b

0.2320b

a

0.0470

2

3

4

0.8167a

0.2804a

b

0.0373a

0.0597a

0.0320a

b

2

3

4

0.7808b

1.3048b

0.5820b

0.1458

0.1327b

0.2085b

0.1470b

1

2

3

4

0.6054

0.4155a

0.0482

1

a

0.5593b

0.2361

1

b

1,076a

1,120a

2,442c

1,069

a

3,299a

3,364a

5,832a

3,257

a

0.2057

±0.1539b

±0.1810

0.8784b

±0.2641a

±0.1513a

1.2120

±0.1940b

2,080b

2,065b

±0.0895

3,094b

2,040

a

96.39

99.57

98.38

96.26

90.45

93.13

89.76

83.98

97.67

94.32

95.35

94.12

96.95

1,460

173

655

1,510

16,900

12,200

18,100

28,300

7,391

18,058

14,800

18,700

5,260

0.1183

0.3511

0.9743

0.4079

0.4191

0.1762a

0.0920

0.1863a

0.1870

a

0.1106a

0.3575

b

0.0141

0.0100

0.0212

0.0289

0.1192b

0.5058

Do not converge

0.1371

a

0.2825b

0.4972

0.2843b

0.2841

b

±0.1790

Do not converge 0.2440

±1.1150

3,010 Do not converge

1,888b

98.26

0.4060

4,529

Do not converge

0.5272b

0.2380b

±0.0510

0.1162c

3

0.1690

0.2748a

Do not converge

0.2651a

0.2721a

0.1920

Do not converge

4

±0.1735

5,460

±0.1250

q

Do not converge

0.1500

p

0.2325b

96.84

22,600

3,480

17,600

22,800

25,400

22,000

22,200

29,000

SSR(106)

2

1,905

91.89

98.75

93.70

91.83

86.26

88.02

84.43

87.43

R2

0.2367

a

a

0.7655

±0.2281

2,656a

2,827

±0.2468

a

0.8375

1.1688

0.2602

d, g or b

Spain

9,265

2,632a

4,153a

4,435

1,153

3,818a

M (103)

1

0.3974

0.2428a

0.1063a

q

1

Model p

2,494a

2,436b

Source: Authors' own estimates

2,169

2,405

a

2,113a

2,106a

2,060

a

1,813b

1,795b

1,633

1,756

a

1,528

1,915

3,119a

2,042

3,153a

2,719

3,292

M (103)

0.5695

0.1130

0.0190

0.9981b

±0.2474c

1.1687

±0.1022

0.0125

±3.7090

±0.4854

0.0750

±7.2480

d, g or b

France

97.39

96.78

96.76

96.76

92.79

81.15

87.33

98.45

98.19

97.99

97.98

59.00

18.45

99.39

99.04

99.34

88.74

85.31

R2

3,250

4,013

4,040

4,040

6,870

18,000

12,100

2,640

3,080

3,440

3,450

27,923

55,540

2,710

2,420

2,910

11,533

15,047

0.6267b

0.4758b

0.4387a

0.6988a

0.3635

0.5491

0.6084

0.4569

0.4907

0.1344b

b

0.4145b

0.1865b

0.4243c

0.3613

b

0.5228

0.4201b

Do not converge

0.1481b

0.1483

b

0.1422a

0.0079

0.1719a

0.1717

a

0.1551b

a

0.3932a

Do not converge

0.1602b

0.1596

b

0.4225a

b

0.4330b

Do not converge

0.4123a

0.4451

a

0.2391b

Do not converge

0.2400a

0.2459a

0.1685a

0.8323

b

b

0.2746

0.5336b

0.4559a

q

0.1747a

0.1174a

SSR(106) p

803a

617

743

a

1,476a

1,472a

1,672b

1,454a

1,504b

1,136

1,369

a

903a

1,007a

888a

1,805a

2,363b

1,758a

1,679a

1,676

a

1,957b

1,652a

M (103)

Italy

c

0.7745b

0.0460

0.6565b

0.6223c

±0.0370

1.5202

0.0440

0.9957a

±0.0374b

0.5752

±0.0593

0.3413b

±0.1154

±0.0417

d, g or b

99.77

95.83

95.74

97.27

95.22

92.73

92.59

90.66

88.22

88.01

99.83

99.74

79.22

99.30

99.29

98.82

99.77

99.40

99.29

99.07

R2

18

331

339

2,990

5,240

7,970

8,120

9,120

11,500

11,700

249

389

1,150

1,340

1,370

2,270

264

708

831

1,090

SSR(106)

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

Notes: WS: weeks of duration in Spain; WF: weeks of duration in France; WI: weeks of duration in Italy; a : prob.  0.001; b: prob.  0.05; c: prob.  0.09.

19

7

9

14

8

6

9

10

5

13

WS

Film

Table III Parameter estimates for Spain, France and Italy The distribution model in the diffusion of innovations European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

variables of marketing decisions, and the conviction that both the socio-economic environment and the unavoidable time-lag are decisive factors in the differences seen among the diffusion processes of different geographical areas, have inspired us to analyze these phenomena in the context of films, in three European countries, between 1997 and 1999. As a novelty, the methodology employed is based on different single-equation models that include the distribution of the product, following the proposals of Jain and Rao (1990) and those of Bass et al. (1994). The goodness of fit of these models was compared to that of the original Bass (1969) model. The use of different contrasts and statistical techniques then allowed us to detect and examine the differences that exist among the diffusion parameters of different countries, as well as to verify whether the time-lag in the introduction of new films into different countries had any significant effect on the speed of their diffusion processes. The empirical application, which was carried out in Spain, France and Italy, allows us to conclude that our Model 4, derived from the GBM, seems to be superior to the others for the analysis of our sample of film premieres. In that model it is obvious that the influence of external sources and the experience of previous adopters are basic factors in reducing uncertainty in new adopters with regard to the innovation. The mere knowledge of the existence of an innovation does not seem to be sufficient for an individual to become a new adopter of the innovation. We have not obtained enough statistical evidence, however, to conclude that there is any influence exerted by the intermediation coefficient on the diffusion process. On the other hand, significant differences have been detected in consumer preferences between Spain and France and between Italy and France, although no differences have been noted between Spain and Italy. The slight cultural, economic or social differences that might exist between these two Mediterranean countries, Spain and Italy, do not seem to be great enough to cause any significant differences between the internal influence coefficients of their diffusion processes. Finally, our chosen model has not detected sufficient statistical evidence to arrive at any conclusions

Table IV Constrast of significance of the difference in averages by pairs of countries

qÃÅSpain ± qÃÅFrance

qÃÅSpain ± qÃÅItaly

qÃÅFrance ± qÃÅItaly

±0.0437

±0.3846*

0.3409* Note: *: prob  0.05 Source: Author's own estimates

significant, however, in the case of Spain-Italy. In other words, the tests do not detect any important cultural, economic and/or social differences between these two very similar Mediterranean countries, Spain and Italy, that could have caused different preferences for the group of films analyzed. However, the geographical closeness between France and her two European neighbours (Spain and Italy) is not enough to eliminate the differences in the intrinsic characteristics between them. The data clearly indicate that a ``country'' effect does exist between France and Spain, and between France and Italy, but not so between Spain and Italy. B. The ``time'' effect To verify H2, (the effect of the moment of introduction on the adoption ratio), the regression model detailed in equation (10) (part 3) is used. The parameter of this expression allows us to verify whether the temporary lag in the introduction of a new film either speeds up or slows down the diffusion process in the different countries. The estimation of this regression using least ordinary squares gives a coefficient of 0.0041 (standard error = 0.0115; prob. = 0.7279), which is not statistically significant. This result indicates that there is no sufficient statistical basis to either support or reject the hypothesis that the time lag in the introduction of new films into the countries examined here affects the speed of their diffusion processes. In fact, our results do not support any of the hypotheses defended by Takada and Jain (1991), Mahajan and Muller (1994), Kumar et al. (1998), or by Helsen et al. (1993).

6. Conclusions The implication that the type of distribution modelling that has been carried out, so far, in diffusion models that employ systems of equations, does not prove to be superior to other single-equation models that ignore 16

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

about the hypothesis that the time-lag in the introduction of new films into the different countries causes either an acceleration or a deceleration in their diffusion processes.

which the ordinary least squares do not, and the error term reflects the net effect of the sample errors as a result of the exclusion of variables and employing erroneous specifications for the density function. The standard errors of the parameters are more realistic than those obtained by Maximum Likelihood. The limitations of the model (as with Maximum Likelihood) are: its execution requires search routines, and the algorithm might not converge.

Notes 1 The coefficient of diffusion may be interpreted as a coefficient of conversion or mechanism that changes the potential adopters into effective adopters. In other words, it indicates the rate at which the adoption of the innovation takes place and is, therefore, an index of the response from the non-adopters. Its value is a function of all the elements that influence the diffusion process, including the innovation under analysis, the communication channels used and the social system in which the innovation is diffused. If these elements are established beforehand, this coefficient becomes a proportional constant. 2 The early models considered a social system of a constant, finite and known size, or easily verifiable one. These models were static and did not allow for changes in the size during the diffusion process. The size of a social system may indeed vary with time, thanks to numerous factors, both exogenous (economic, social or technological) and endogenous (advertising campaigns or changes in distribution channels). Among the authors that have relaxed this restriction and have proposed dynamizing the population of prospective adopters, the following stand out: Chow (1967), Dodson and Muller (1978), Lackman (1978), Mahajan and Peterson (1978), Mahajan et al. (1979), Sharif and Ramanathan (1981), Jorgensen (1983), Kalish (1985), Polo (1996), Kamakura and Balasubramanian (1988), Jain and Rao (1990), Horsky (1990), Jones and Ritz (1991) and Parker (1992). The accuracy of a dynamic diffusion model depends, to a great extent, on identifying the variable, or variables, that affect the potential market, and determining just how they affect it. 3 The term I ti has been included in equation (7) in a multiplying way so that it will also affect the i †ÿF…tiÿ1 † adoption ratio F…t1ÿF…t (Jain and Rao, 1990). iÿ1 †

References Bass, F.M. (1969), ``A new product growth for model consumer durables'', Marketing Science, No. 15, pp. 215-27. Bass, F.M. (1980), ``The relations between diffusion rates, experience curves and demand elasticities for consumer durables technological innovation'', Journal of Business, No. 53, pp. 51-67. Bass, F.M. and Bultez, A. (1982), ``A note on optimal strategic pricing of technological innovations'', Marketing Science, Vol. 1 No. 4, pp. 371-8. Bass, F.M., Krishnan, T.V. and Jain, D. (1994), ``Why the Bass model fits without decision variables'', Marketing Science, Vol. 13 No. 3, pp. 203-23. Bewley, R. and Fiebig, D. (1988), ``A flexible logistic growth with applications in telecommunications'', International Journal of Forecasting, No. 4, pp. 177-92. Chatterjee, R. and Eliashberg, J. (1990), ``The innovation diffusion process in heterogeneous population: a micromodeling approach'', Management Science, No. 36, pp. 1057-79. Chow, G. (1967), ``Technological change and the demand for computers'', American Economic Review, No. 57, pp. 1117-30. Dockner, E. and Jorgensen, S. (1988), ``Optimal advertising policies for diffusion models of new product innovations in monopolistic situation'', Management Science, No. 34, pp. 119-30. Dodson, J. and Muller, E. (1978), ``Models of new products diffusion through advertising and worth-of-mouth'', Management Science, No. 24, pp. 1568-78. Easingwood, C. (1987), ``Early product life-cycle forms for infrequently purchased major products'', International Journal of Research in Marketing, No. 4, pp. 3-9. Easingwood, C. (1988), ``Product life-cycle patterns for new industrial products'', R&D Management, No. 18, pp. 23-32. Easingwood, C. (1989), ``An analogical approach to the long-term forecasting of major new product sales'', International Journal of Forecasting, No. 5, pp. 69-82. Easingwood, C., Mahajan, V. and Muller, E. (1981), ``A nonsymmetric responding logistic model for technological substitution'', Technological Forecasting and Social Change, No. 20, pp. 199-213. Easingwood, C., Mahajan, V. and Muller, E. (1983), ``A nonuniform influence innovation diffusion model of new product acceptance'', Management Science, No. 2, pp. 273-96.

4 The flexible models suggest structures that offer a wider range of diffusion patterns to better reflect the underlying function in the sample data. This avoids the imposition of any given function and allows a wider margin for the dependent variable in the model, such as diffusion curves that are not symmetrical and flexible inflexion points. The most outstanding studies published on the topic are those of Floyd (1968), Sharif and Kabir (1976), Jeuland (1981), Easingwood et al. (1981; 1983), Easingwood (1987; 1988; 1989), Bewley and Fiebig (1988) and Bass et al. (1994). 5 Jones and Ritz (1991) discard films that run for under five weeks. 6 Non-linear estimation offers two advantages (Srinivasan and Mason, 1986): it allows us to obtain the standard errors for the parameters estimated,

17

The distribution model in the diffusion of innovations

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Floyd, A. (1968), ``A methodology for trend-forecasting of figures of merit'', in Bright, J. (Ed.), Technological Forecasting for Industry and Government: Methods and Applications, Prentice-Hall, Englewood Cliffs, NJ, pp. 95-109. Fourt, L.A. and Woodlock, J.W. (1960), ``Early prediction of market success for new grocery products'', Journal of Marketing, No. 25, pp. 31-8. Gatignon, H., Eliashberg, J. and Robertson, T.S. (1989), ``Modelling multinational diffusion patterns: an efficient methodology'', Marketing Science, No. 8, pp. 231-47. Helsen K., Jedidi, K. and DeSarbo, W.S. (1993), ``A new approach to country segmentation utilizing multinational diffusion patterns'', Journal of Marketing, No. 57, pp. 60-71. Horsky, D. (1990), ``A diffusion model incorporating product benefits, price, income and information'', Marketing Science, No. 9, pp. 342-65. Horsky, D. and Simon, L. (1983), ``Advertising and the diffusion of new product'', Marketing Science, No. 1, pp. 1-18. Jain, D. (1992), ``Marketing mix effects on the diffusion of innovations'', working paper, Northwestern University, Kellogg's Graduate School of Management. Jain, D. and Rao, R.C. (1990), ``Effect of price on the demand for durables'', Journal of Business and Economic Statistics, No. 8, pp. 163-70. Jeuland, A. (1981), ``Parsimonious models of diffusion of innovation: Part a, derivations and comparison'', working paper, Graduate School of Business, University of Chicago. Jeuland, A. (1993), ``The Bass model with population heterogeneity (a lesson in parsimony)'', working paper, University of Chicago. Jones, J.M. and Ritz, C.J. (1991), ``Incorporating distribution into new products diffusion models'', International Journal of Research in Marketing, No. 8, pp. 91-112. Jorgensen, S. (1983), ``Optimal control of a diffusion model of new products acceptance with price-dependent total market potential'', Optimal Control Applications and Methods, No. 4, pp. 269-76. Kalish, S. (1983), ``Monopolist pricing with dynamic demand and production cost'', Marketing Science, No. 2, pp. 135-60. Kalish, S. (1985), ``A new product adoption model with pricing, advertising and uncertainty'', Management Science, No. 31, pp. 1569-85. Kamakura, W. and Balasubramanian, S. (1988), ``Long-term view of the diffusion of durables: a study of role of price and adoption influence processes via test of nested models'', International Journal of Research in Marketing, No. 5, pp. 1-13. Kumar, V., Ganesh, J. and Echambadi, R. (1998), ``Cross-national diffusion research: what do we know and how certain are we?'', Journal Product Innovation Management, No. 15, pp. 255-68. Lackman, C. (1978), ``Gompertz curve forecasting: a new product application'', Journal of Marketing Research Society, No. 20, pp. 45-7.

Mahajan, V. and Muller, E. (1979), ``Innovation diffusion and new product growth models in marketing'', Journal of Marketing, No. 43, pp. 55-68. Mahajan, V. and Muller, E. (1994), ``Innovation diffusion and borderless global market: will the 1992 unification of the European Community accelerate diffusion of ideas, products and technologies?'', Technological Forecasting and Social Change, Vol. 45 No. 3, pp. 221-35. Mahajan, V. and Peterson, R. (1978), ``Innovation diffusion in a dynamic potential adopter population'', Management Science, No. 24, pp. 1589-97. Mahajan, V., Muller, E. and Bass, F.M. (1993), ``New-product diffusion models'', in Eliashberg, J. and Lilien, G.L. (Eds), Handbooks in Operations Research and Management Science, Elsevier Science Publishers, New York, NY. Mahajan, V., Muller, E. and Bass, F.M. (1995), ``Diffusion of new products: empirical generalizations and managerial uses'', Marketing Science, Vol. 14 No. 3, pp. G79-G88. Mahajan, V., Peterson, R., Jain, A. and Malhotra, N. (1979), ``A new product growth model with a dynamic market potential'', Long Range Planning, No. 2, pp. 51-8. Mansfield, E. (1961), ``Technical change and the rate of imitation'', Econometrica, No. 29, pp. 741-66. Mesak, H. (1996), ``Incorporating price, advertising and distribution in diffusion models of innovation: some theorical and empirical results expectations in diffusion models'', Computers and Operations Research, Vol. 23 No. 10, pp. 1007-23. Mesak, H. and Berg, W. (1995), ``Incorporating price and replacement purchases in new products diffusion models for consumer durables'', Decision Sciences, No. 26, pp. 425-49. Parker, P. (1992), ``Price elasticity dynamics over the adoption life cycle'', Journal of Marketing Research, No. 29, pp. 358-67. Parker, P. (1994), ``Aggregate diffusion forecasting models in marketing: a critical review'', International Journal of Forecasting, No. 10, pp. 353-80. Polo, Y. (1996), ``Modelo de crecimiento de nuevos productos con mercado potencial dinaÂmico'', InvestigacioÂn y Marketing, No. 24, pp. 29-35. Rao, A. and Yamada, M. (1988), ``Forecasting with a repeat purchase diffusion model'', Management Science, No. 34, pp. 734-52. Redmond, W. (1994), ``Diffusion at sub-national levels: a regional analysis of new product growth'', Journal of Product Innovation Management, No. 11, pp. 201-12. Robinson, B. and Lakhani, C. (1975), ``Dynamic price models for new product planning'', Management Science, No. 10, pp. 1113-22. Rogers, E.M. (1983), Diffusion of Innovations, The Free Press, New York, NY. Sharif, M. and Kabir, C. (1976), ``A generalized model for forecasting technological substitution'', Technological of Forecasting and Social Change, No. 8, pp. 353-64. Sharif, M. and Ramanathan, K. (1981), ``Binomial innovation diffusion models with dynamic potential adopter population'', Technological of Forecasting and Social Change, No. 20, pp. 63-87.

18

The distribution model in the diffusion of innovations

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 6±19

Enar RuõÂz Conde and Francisco Jose MaÂs RuõÂz

Simon, H. and Sebastian, K. (1987), ``Diffusion and advertising: the German telephone company'', Management Science, No. 33, pp. 451-66. Srinivasan, V. and Mason, C. (1986), ``Nonlinear least squares estimation of new product diffusion models'', Marketing Science, No. 5, pp. 169-78. Sultan, F., Farley, J. and Lehmann, D. (1990), ``A meta-analisys of applications of diffusion models'', Journal of Marketing Research, No. 27, pp. 375-88.

Takada, H. and Jain, D. (1991), ``Cross-national analysis of diffusion of consumer durable goods in Pacific Rim countries'', Journal of Marketing, No. 55, pp. 48-54. Teng, J. and Thompson, R. (1983), ``Oligopoly models for optimal advertising when production costs obey a learning curve'', Management Science, No. 29, pp. 1087-101. Thompson, R. and Teng, J. (1984), ``Optimal pricing and advertising policies for new product oligopoly models'', Marketing Science, Vol. 3 No. 2, pp. 148-68.

19

Introduction

Innovation as newness: what is new, how new, and new to whom?

During the last decade we have observed an explosive attention, both in the popular press (e.g. Young, 1994) and among academics (e.g. Drazin and Schoonhoven, 1996; Kanter, 1985), on innovation as a means to create and maintain sustainable competitive advantages. Innovation is considered a fundamental component of entrepreneurship (e.g. Covin and Miles, in press) and a key element of business success (e.g. Nonaka and Takeuchi, 1995). This is becoming even more evident as we move into a post-capitalist, knowledgebased society (Drucker, 1993). Jacobson (1992) argues that continuous changes in the state of knowledge produce new disequilibrium situations and, therefore, new profit opportunities or ``gaps''. The rate of change is also increasing due in part to exponential advancements in technology, frequent shifts in the nature of customer demand, and increased global competition. D'Aveni (1994) categorizes the situation in its extreme form as ``hyper-competition'' and, as we move into a more knowledge-based society, an increasing number of industries and firms are likely to face such hypercompetitive conditions. Hence, the unending and increasing stream of knowledge that keeps marketplaces in perpetual motion will require companies to focus even harder on being innovative in order to create and sustain competitive advantages. The growing importance of innovation to entrepreneurship is reflected in a dramatic increase in literature that addresses the role and nature of innovation (Drazin and Schoonhoven, 1996; Drucker, 1985). In spite of this increase and the resulting vibrancy within the field, prior research has not yielded a widely-held consensus regarding how to define innovation. Additionally, without a good working definition, we still lack good measures of innovation. Kotabe and Swan (1995) argue that one of the greatest obstacles to understanding innovation has been the lack of a meaningful measure. Without adequate measures, theory development is impeded and it becomes difficult to suggest appropriate interventions for firms seeking to pursue

Jon-Arild Johannessen Bjùrn Olsen and G.T. Lumpkin

The authors Jon-Arild Johannessen is a Professor at the Norwegian School of Management, Oslo, Norway. Bjùrn Olsen is an Associate Professor at the Bodù Graduate School of Business, Norway. G.T. Lumpkin is Assistant Professor at the University of Illinois at Chicago, Chicago, Illinois, USA. Keywords Innovation, Measurement, Entrepreneurialism, Norway Abstract Innovation implies newness. To define and measure innovation better, we investigated three dimensions of newness: what is new, how new, and new to whom? Drawing on prior research by Schumpeter and Kirzner, we developed a scale that addresses six areas of innovative activity: new products, new services, new methods of production, opening new markets, new sources of supply, and new ways of organizing. Using factor analysis on data from two separate field studies ± 684 firms from eight industries and 200 information technology firms ± we found that innovation as newness represents a unidimensional construct, distinguished only by the degree of radicalness. Electronic access The research register for this journal is available at http://www.mcbup.com/research_registers The current issue and full text archive of this journal is available at http://www.emerald-library.com/ft

The authors thank Rod Shrader for his helpful comments on an earlier draft and gratefully acknowledge the funding that was provided for this research by the Norwegian Research Council (The FAKTA-programme).

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . pp. 20±31 # MCB University Press . ISSN 1460-1060

20

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

Literature review

innovations. To address these issues, the overarching research question considered in the present study is, what is innovation and how should it be operationalized? As a starting point, we note that nearly every definition of innovation focuses on the concept of newness. Slappendel (1996) argues that the perception of newness is essential to the concept of innovation as it serves to differentiate innovation from change. The newness theme is especially important to understanding the link between innovation and entrepreneurship as suggested by prior studies that emphasize its pivotal role in new venture creation and management: ``new business startup'' (Vesper, 1988), ``new entry'' (Lumpkin and Dess, 1996), ``new organizations'' (Gartner, 1988) and ``organizational renewal'' (Stevenson and Jarillo, 1990). Thus, we suggest that, in order to isolate a useful definition and measure of innovation, we need to address three newness-related questions: what is new, how new, and new to whom? With these innovation concepts in mind, we developed a study that investigated six different types of innovative activity: (1) new products; (2) new services; (3) new methods of production; (4) opening new markets; (5) new sources of supply; and (6) new ways of organizing.

The innovation literature can be categorized into four different approaches or orientations: (1) individual-oriented; (2) structure-oriented; (3) interactive-oriented; and (4) systems of innovation-oriented. The individual-oriented perspective emphasizes the role of individual factors such as age, educational level, gender, cognitive style and creativity (e.g. Scott and Bruce, 1994). Influential theoretical sources are mainly found in the notion of the rational actor and in limited rationality as a determinant of innovation (Pettigrew, 1985; Cyert and March, 1963). The structural perspective focuses on organizational characteristics. Influential theoretical sources here are structural functionalism and contingency theory, i.e. how organizational structure constrains or propels innovation. A strongly emphasized area in this perspective is the relationship between the organization and the environment (Slappendel, 1996). The interactive perspective has recently received increased attention (e.g. Van de Ven et al., 1989; Van de Ven and Rogers, 1988). The focus in this perspective is on how action influences structure, and vice versa in the innovation process (Van de Ven and Poole, 1988; Pettigrew, 1985; Walton, 1987), and some importance has been attached to the political context of innovation (Child and Smith, 1987). A fourth research school which has also received increased attention in recent years is the study of how national and regional innovation systems influence innovation activity in companies (Nelson and Winter, 1982; Lundvall and Johnson, 1994; Edquist, 1997). The main focus is on the organization in the environment, interactive learning, knowledge creation, the practical use of knowledge and the distribution of knowledge. In particular, the knowledge infrastructure and the organization of networks between companies and knowledge institutions, suppliers, customers and other entities are emphasized in this perspective. Each of these orientations may be useful for addressing issues of the definition and measurement of innovation. But the picture that emerges from these diverse approaches underscores the point that a multitude of factors are interacting to induce innovation in economic life. The various perspectives are

The purpose of the study was to explore how perceptions of innovative activity in these six areas might contribute to a meaningful definition of innovation and inform us about effective ways to measure innovation. Two different mailed surveys were conducted among Norwegian firms ± a ``general'' study to which 696 CEOs from eight industry groups responded, and a ``knowledge sector'' study that yielded 200 CEO respondents from the information technology sector (IT-sector). The remainder of this analysis is divided into four parts. Next, we discuss prior theory and literature that has guided the study of innovation and contributed to our understanding of the salient issues. Then, we describe the sample and methods used to investigate these issues. Third, we present our results and, finally, we discuss the implications of our findings for current and future innovation and entrepreneurship research. 21

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

indicative of the lack of common definitions and measures of innovation. The resulting inconsistency makes it difficult to conduct comparative studies because it is often unclear whether researchers are studying the same phenomena. Thus, in the subsections below, we draw on important insights from each of these four schools to focus on the elements of innovation that are common across approaches. Innovation as newness Most of the widely-used definitions of innovation focus on novelty and newness. For example, the European Commission Green paper on innovation defines innovation rather broadly as a synonym for ``the successful production, assimilation and exploitation of novelty in the economic and social spheres'' (European Commission, 1995, p. 9). Nohria and Gulati (1996) defined innovation to include any policy, structure, method or process, or any product or market opportunity that the manager of an innovating unit perceives to be new. Damanpour defined innovation as ``the generation, development, and adaption of novel ideas on the part of the firm'' (1991, p. 556), and Zaltman et al. defined it as ``any idea, practice, or material artifact perceived to be new by the relevant unit of adoption'' (1973, p. 10). Although newness is a theme in all of these definitions, they do not agree on three basic questions about the nature of newness: what is new, how new, and new to whom? Several of the definitions suggest a theme of ``successful adoption'', for example, but are vague in terms of what is adopted and what constitutes a success. Specifying what is new is important for distinguishing innovation from mere change (Slappendel, 1996) because all innovation presupposes change, but not all change presupposes innovation. Additionally, none of the above definitions addresses the issue ``how new?'', that is, the degree or extent of newness that constitutes an innovation. Finally, the issue of new to whom? is also unresolved in the above definitions. Nohria and Gulati's (1996) definition seems confusing because it is unclear whether the newness of an innovation applies to the manager of an innovating unit or to the innovating unit itself. Damanpour's (1991) emphasis on newness to the firm seems to exclude the kind of innovation that 22

might be associated with individuals or emerge from systems of innovation outside the firm. As a starting point, this study embraces Zaltman et al.'s (1973) definition of innovation as ``any idea, practice, or material artifact perceived to be new by the relevant unit of adoption'' to guide our examination of what is new, how new, and new to whom? Next, we will discuss each of these subquestions in more detail. What is new? Evidence of vagueness in specifying what about innovation is new can be found by analyzing how innovation has been operationalized in prior studies. A European example illustrates this well. In 1991, the European Commission stated the following: ``economic performance depends upon the progressive introduction over time of innovations in products and processes . . . '' (European Commission, 1991, p. 8). This notion was elaborated in the European Commission Green Paper on innovation which emphasized ``the successful production, assimilation and exploitation of novelty in the economic and social spheres'' (1995, p. 10). When it came to operationalizing the construct, however, the Green Paper used proxie's as measures of innovative activity rather than explicitly addressing what is new. These proxie's include, among other measures, total expenditure on R&D, proportion of R&D scientists and engineers, and number of patents. Similar measures can be found in other innovation research: Daft and Becker (1978) analyzed the number of innovations adopted within a given period of time, Blau and McKinley (1979) investigated the number of patents, Miller (1987) measured the relative amount spent on R&D, and Miller and Friesen (1978) used the number of new product and service introductions. The measures in these earlier studies often had limited face validity and tended to foster a narrow view of innovation. Such operationalizations are rather weak indicators of what is new and they generate several levels of problems for research. First, these measures indicate a general lack of consistency between definition and measurement. Second, a heavy focus on R&D suggests a linear approach to the innovation process, although most contemporary

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

meaning of an industry by creating new technological regimes or paradigms. The invention of the combustion engine and IBM's introduction of the DOS operating system are examples of such revolutionary innovations. Lawless and Anderson (1996) argued that most punctuated equilibrium models are explaining broad patterns of change on a historical time scale. However, within these paradigms, considerable innovative activity may take place. Henderson and Clark (1990), for example, argued for the importance of architectural innovations, i.e. the reconfiguration of existing products. Drazin and Schoonhoven (1996) noted that the emergence of a dominant design leads to additional innovation, bringing new approaches and technologies in its wake. For instance, in the IT-sector, the pace of innovation has been extremely rapid within existing technological regimes. The distinction between radical and incremental innovations is also often highlighted in studies of innovativeness. Hage (1980) argued that innovations vary along a continuum from incremental to radical. The term ``radical'' has been associated with revolutionary innovations, whereas ``incremental'' is associated with innovations within a paradigm (Dosi, 1982; Dewar and Dutton, 1986). However, the distinctions noted above suggest that the terms radical and incremental could also be used in a within-paradigm context. Damanpour (1996), for example, used the term radical innovations to characterize innovations that produce fundamental changes in the activities of an organization and large departures from existing practices, whereas the term incremental innovations was used to depict innovations that represent a lesser degree of departure from existing practices. In both cases, the terms apply to within-organization innovations. From this very brief review of the literature, we observed that the issue of ``how new'' is closely linked to the question, ``new to whom?'' That is, in order to operationalize the distinction between incremental and radical innovations, we must also determine the relevant unit of analysis. It is that issue that we turn to next.

research emphasizes circular processes (e.g. Nelson and Winter, 1982; Edquist, 1997). Third, by focusing on the proportion of scientists and engineers, they leave out other members of the organization who may be equally important to the innovative activity within a firm (Johannessen and Hauan, 1994). Fourth, by using patents as measures of innovative activity, they ignore those who argue that patents are often not commercialized (Manu and Sriram, 1996), and that innovations may take other forms than only those that it is possible to patent. It may also be argued that all innovations are not patented. Hence, the operationalizations and measurement of innovation in prior research provide little guidance to the question ``What is new?'' Exceptions, however, do exist. Some researchers have used methods that are consistent with Zaltman et al.'s (1973) notion of perceptions of newness. McGrath et al. (1996) operationalized innovation by having participants address to what extent 15 different project characteristics were new to the firm at the moment. The characteristics ranged from new products to the skill of the project team. Damanpour (1996) operationalized innovation broadly to encompass a range of types, including new products or services, new organizational structures or administrative systems, new process technologies or new plans or programs pertaining to organizational members. We follow in the footsteps of McGrath et al. (1996) and Damanpour (1996), who themselves relied on important work by Schumpeter (1934; 1939; 1942) and Kirzner (1976; 1985), to operationalize what is new in a fashion that addresses a range of innovative activities across broadly-defined ``relevant units of adoption''. How new? A review of the four orientations in the innovation literature reveals that several different approaches have been used to address the issue of how new, that is, the degree of newness that constitutes an innovation. The literature has devoted considerable attention to debating the issue of revolutionary innovations (Gersick, 1991). Revolutionary innovations, often reflected in punctuated equilibrium models (Tushman and Romanelli, 1985), describe situations where discontinuities totally redefine the

New to whom? Prior innovation research suggests that the extent of newness of an innovation may be 23

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

``knowledge-sector'' study, data were collected from the information technology sector (the IT-sector). Both studies included both manufacturing and service firms. The general study was part of a larger study on critical innovation factors in Norwegian firms. The study was initiated and sponsored by the Norwegian Ministry of Local Government and Labor. A field survey was designed and survey data were collected from eight different industry groups including both manufacturing and service firms. The firms were selected from a Norwegian database (Bedriftsdatabasen) consisting of all incorporated firms in Norway. Within each group, we randomly selected 750 firms, using a total of 40 two-, three-, and five-digit NACE codes (Nomenclature geÂneÂrale des ActiviteÂs eÂconomiques dans les CommunauteÂs EuropeÂennes): 111, 18, 192, 193, 20, 21, 23, 24, 28, 630, 28, 752, 297, 30, 31, 32, 33, 34, 300, 35, 361, 364, 365, 36, 631, 36, 633, 40, 45, 45, 442, 524, 55, 633, 72, 73, 741, 746, 747, 85. Since not all of the eight categories contained as many as 750 firms, the final mailing included only 5,584 questionnaires. The surveys were mailed in January 1996 to the CEOs of these firms as they were considered to be best informed about the extent of innovative activity within the firm. The questionnaires were followed-up by two written reminders. A total of 267 undeliverable questionnaires were returned. Of the 5,050 remaining firms, 696 returned questionnaires. Of these, 12 could not be used for further studies. The final count of usable questionnaires was 684 for a response rate of 13.5 per cent. The second study, the knowledge-sector study, was part of a larger study on innovation and performance in small and new firms within knowledge-based sectors in Norway. The study, which was initiated and sponsored by the Norwegian Research Counsil, focused primarily on relatively new (< ten years old) SME's (< 100 employees). To highlight the research question, a field survey was designed, and survey data were collected from the Norwegian IT-sector. This sector was chosen in part because a priori studies of the Norwegian IT-sector (STEP, 1995) indicated a higher level of innovations in this sector than in other industry sectors. This gave us reason to believe that the IT-sector was, on average, more knowledge intensive than the average of all other sectors which we regarded

related to the domain into which the innovation is adopted. In other words, to assess the nature of an innovation, we need simultaneously to consider the ``relevant unit of adoption''. Both Cooper (1993) and Kotabe and Swan (1995) argued that innovation can be investigated in terms of both newness to the company (the firm-based framework), and newness to the market (the newness to the market framework). Although the firm-based framework is unlikely to reflect a product's impact on either competitors or customers (Kotabe and Swan, 1995), from a broader perspective, the measure does capture the ability of a firm to service and continue to update the innovative technology which are key consumer concerns. Thus, even innovations that are primarily new within a firm may have an impact outside the firm. Booz, Allen & Hamilton (1982) combined the two approaches in a framework that identifies six levels of product innovativeness. However, since innovations can materialize both as new products and new processes (Utterback and Abernathy, 1975; Damanpour, 1996), we argue that newness to a market framework represents a view that is too narrowly focused on product innovations. To encompass both product and process innovations, we suggest that newness to the industry, rather than newness to the market, represents a more broadly-construed and inclusive framework. Thus, by ascribing to Zaltman et al.'s (1973) notion of ``relevant units of adoption'', we envision a continuum of units of adoption that is roughly parallel to the degree of radicalness continuum. That is, as the economic unit that adopts an innovation becomes more broadly-defined or encompassing, the impact of the innovation is more likely to be radical, that is, organizationchanging or paradigm-shifting. In operationalizing this distinction, therefore, it is important to inquire about both within-firm and industry-level innovations to address the question, ``new to whom?''

Methodology Sample and data The findings reported in this paper are based on data from two separate studies. For the first study, labeled the ``general'' study, data were collected from eight different industry groups. In the second study, called the 24

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

activity. Hence, for both studies, we used six variables to assess the innovativeness of the firms we examined. These variables are: (1) new products (NEWPROD) (2) new services (NEWSERVI); (3) new methods of production (NEWMETO); (4) opening new markets (NEWMARK); (5) new sources of supply (NEWMATER); and (6) new ways of organizing (NEWORG).

as important for testing our research questions. In order to encompass most of the value chain-related activities of the IT-sector, we included both hardware and software producers, as well as sales and services connected to IT in our sector definition. A total of 12 four-digit NACE codes were selected for this study: 3000, 3200, 3300, 5164, 6420, 7133, 7260, 7210, 7220, 7230, 7240 and 7250. Firms were selected from the same Norwegian database (Bedriftsdatabasen), containing all incorporated firms in Norway. To meet the focus of the overall research program, the criteria for selection were: . firms had to be ten years old or less; and . firms had to have more than two and fewer than 100 employees.

The variables are adopted and deduced from Schumpeter (1934; 1939; 1942)[1] and Kirzner (1976; 1985)[2], but have also been used separately by a numbers of researchers (e.g. Utterback and Abernathy, 1975; Damanpour and Evan, 1984; McGrath et al., 1996; Damanpour, 1996). However, the composition of innovation variables which are used in the present study has not, to the best of our knowledge, been previously used in this form for studying innovation. In the general study, respondents were asked to indicate on a five-point scale the degree to which their company had made changes within the last three years to any of the six innovation variables (mentioned above) that were of such a nature that they were perceived to be new to the company. For each of the areas, respondents were asked to indicate the extent of change using a five-point Likert-type scale ranging from 1 = ``To no extent'' to 5 = ``To a very great extent''. In the knowledge sector study, we endeavored to distinguish between radical and incremental innovations. Drawing on the line of arguments made in the literature review, we propose that innovation radicalness may be distinguished by classifying it on the grounds of newness to the ``relevant unit of adoption''. Hence, in order to link the issue of ``how new?'' to the question, ``new to whom?'', we collapsed innovation radicalness into the following two definitions: Incremental innovations are any idea, practice or material artifact that is perceived to be new to the firm, but which may have been previously used by other firms. Radical innovations, by contrast, are any idea, practice or material artifact perceived to be new to the industry. Thus, in the knowledgesector study, we used the firm and the industry in which the firm operates as the distinguishing criteria. Respondents were

Initially, 5,631 companies were considered; 4,551 companies were excluded as they did not meet the above criteria, leaving 1,080 companies to survey. Questionnaires were mailed in March 1996 to the CEOs of these firms as they were considered to be best informed about the firm's innovative activity. The questionnaires were followed-up by two written reminders. A total of 63 undeliverable questionnaires were returned. Of the 1,017 remaining firms, 200 returned completed questionnaires for a response rate of 19.6 per cent. This rate would seem comparable to that achieved in similar studies. Measures An extensive literature review was conducted prior to operationalizing the constructs to enhance the construct validity of our measures. To capture the essence of Zaltman et al.'s (1973) definition of innovation as ``ideas, practices or material artifacts perceived to be new'', we used six variables to reflect perceptions of different types of innovation. For the knowledge-sector study, we also used a question that distinguished innovations based on the degree of radicalness. To operationalize innovation effectively, we have argued that we need to ask the question: what is new? We argue at the outset of this paper that the potential for innovations is found in the ``gaps'' that open up as the result of new disequilibrium situations arising from continuous changes in the state of knowledge. Consistent with prior research, we suggest that, within a paradigm, these gaps can be filled with six different types of innovative 25

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

sector study in Tables IIa and IIb. The Tables display the component loadings, communalities, and sum of the squares, as well as descriptive statistics. As only one component was extracted for each innovation category in both studies, the solution converged after one rotation for both incremental and radical innovations. Hence, the component loadings showed that the analysis grouped together items, both for radical and incremental innovations, which on a priori ground, might have been regarded as very similar. Although a number of studies have suggested that we should classify innovations according to whether they are technical or administrative, or process- versus product-oriented (e.g. Utterback and Abernathy, 1975; Dosi, 1988; Teece, 1989; Damanpour, 1996), our findings support Van de Ven (1986) and Nohria and Gulati (1996) who argue that such distinctions resulted in an unnecessary fragmentation of innovation.

asked to indicate (yes or no) whether the company had made changes within any of the above mentioned six innovation variables within the last three years, which were of such a nature that they were perceived as new to the company, but which had been previously used by other firms. They also responded to measures of radical innovation when asked to indicate (yes or no) whether the company had made changes within the same six areas within the last three years, which were of such a nature that they were perceived as new by the industry in which they operated. For an overview of innovation measures used in the two studies see the Appendix. Analysis For the innovation variables, separate principal component factor analyses were used to reduce the data and develop and test the validity of meaningful constructs. An eigenvalue of one was used to determine the number of components to extract for further analysis (Kim and Mueller, 1978). To ensure convergent validation, items were retained on a factor if their loading exceeded 0.5 on the primary factor; to ensure discriminant validity, no variable could load above 0.4 on any secondary factors. To clarify the loadings, the components were rotated. The varimax rotation method was selected for its simplicity and rigor (Nunnally, 1978). Cronbach's alpha were calculated to ensure internal consistency.

Table IIa Varimax-rotated component loadings for incremental innovation variables in the knowledge-based study Variable NEWPROD NEWSERVI NEWMETO NEWMARK NEWMATER NEWORG

NMean

SD

198 195 195 197 194 192

1.18 1.31 1.46 1.30 1.41 1.35

Eigenvalue % Variance Alpha

Findings

Components 0.38 0.46 0.50 0.46 0.49 0.48

0.56390 0.55382 0.74293 0.51997 0.74867 0.51991

Communality 0.31798 0.30672 0.55195 0.27037 0.56051 0.27031

2.77 38.00 0.67

Note: Listwise missing value treatment

The results from the principal component factor analysis for the general study are exhibited in Table I, and for the knowledgeTable I Varimax-rotated component loadings for innovation variables in the general study

Table IIb Varimax-rotated component loadings for radical innovation variables in the knowledge-based study

Variable

Variable

NEWPROD NEWSERVI NEWMETO NEWMARK NEWMATER NEWORG Eigenvalue % Variance Alpha

NMean

SD

684 685 681 682 669 688

3.02 2.86 2.73 2.79 2.08 3.00

Components 0.98 0.92 1.12 1.02 1.02 0.99

0.77077 0.72582 0.72471 0.70185 0.58661 0.64187

Communality

NEWPROD NEWSERVI NEWMETO NEWMARK NEWMATER NEWORG

0.59409 0.52681 0.52521 0.49259 0.34412 0.41199

Eigenvalue % Variance Alpha

2.89 48.20 0.86

NMean

SD

192 190 191 191 190 190

1.40 1.66 1.70 1.54 1.58 1.67

Components 0.49 0.47 0.46 0.50 0.49 0.47

2.86 47.70 0.78

Note: Listwise missing value treatment

Note: Listwise missing value treatment

26

0.73170 0.63946 0.74635 0.65021 0.76191 0.59944

Communality 0.53538 0.40891 0.55703 0.42278 0.58051 0.35933

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

This finding is also consistent with Damanpour (1991) who found that type of innovation did not appear to influence the relationship between organizational determinants and innovation. The principal component factor analysis of the six innovation variables, in both studies and for both incremental and radical innovations, confirmed the unidimensionality of the scale. All six variables loaded strongly on the same factor. Further, reliability analysis performed on the scale also confirmed the contribution of the six items as indicators of an overall construct of innovation. The innovation construct also met the requirement of convergent validity as there were no statistically significant negative relationships, as well as the requirement for discriminant validity, as there were no significant correlations or factor loadings of measures that were assumed to be distinct. Thus, the results satisfied the requirements necessary to validate a undimensional construct (Campbell and Fiske, 1959). This indicates that the composition of variables deducted from frameworks suggested by Schumpeter and Kirzner shows a very promising direction for measuring innovation, as it enables us to treat innovation as a single concept.

Discussion and implications Innovation is a critical activity that is vitally important for most firms to embrace in order to create and sustain a competitive advantage. Drucker describes innovation as ``the specific instrument of entrepreneurship'' (1985, p. 30) and thus, among entrepreneurial firms, it may be the most critical success factor. Given the pivotal role of innovation to entrepreneurship and business success within increasingly knowledge-based and hypercompetitive environments, the need to understand innovation has become paramount. In order to do so, we have endeavored to explore the concept of innovation and how to measure it. We have seen that there exists some disagreement between the definition and the measure of innovation, and in some cases there is little consistency between definitions and measures. The objective of this paper was to search for a meaningful measure of 27

innovation by focusing on the ``common denominator'' of innovation: newness. Our results indicate that innovation at the organizational level can be defined and measured as a single construct, distinguished only by the degree of radicalness. That is, by focusing on newness as the central concept of innovation, we found that innovation met the requirement of convergent and discriminant validity necessary to support a undimensional construct (Campbell and Fiske, 1959). Although newness has several components ± as indicated by our three subquestions: what is new, how new, and new to whom? ± our results suggest that innovation ranges across a single continuum that encompasses all three aspects. This is consistent with prior research which suggests that the innovation construct need not be fragmented into separate categories or types (Nohria and Gulati, 1996; Van de Ven, 1986). Our findings are also supportive of the view that how new an innovation is perceived to be is closely linked to the issue of who perceives it as such. That is, as the economic unit that recognizes the newness of an innovation increases in size or scope, the more radical the innovation is considered to be. These findings, which enable us to treat innovation as a single construct, are very promising for future innovation and entrepreneurship research. First, by focusing on newness as the essence of innovativeness, it provides a useful starting point for applications of the innovation concept. In this way, it confirms Slappendel's (1996) assertion that innovation is something more than mere change. Thus, for example, in arenas such as process reengineering or organizational design, it allows us to distinguish between changes that are simply alternatives or copies, and changes that are novel and original. As such, the focus on newness can be an indicator of the buildingblocks of sustainable competitive advantage by highlighting factors that are, because of their newness, rare and inimitable, consistent with a resource-base approach to developing competitive advantage (Barney, 1991). Even though our findings support a unidimensional innovation construct, the method used here allows us to consider a range of innovative activities, that is, types of newness. Thus, innovation as newness does not focus too narrowly on elements such as how highly technical an innovation is, or the

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

outcome of large investments in R&D. Instead, it considers a broad array of innovations ± characterized by the six areas addressed in this study ± and acknowledges that innovation might occur in any of these areas. In this way, innovation as newness cuts across the four approaches to innovation outlined in the literature review. That is, an innovation is just as likely to come from an individual inventor as from an organizational initiative or a government-sponsored program. The success of an innovation, therefore, is determined more by the extent of its adoption than by who originates it or how technologically advanced it is. What makes it innovative is its newness. Numerous studies have focused on the external and internal influences that are associated with innovativeness in organizations. The insights from this study regarding the importance of newness to innovative processes may help organizations further innovation by suggesting ways to manage and predict factors that are antecedent to and corollaries of innovation. External factors influencing innovation that have been highlighted in the literature include customer-supplier relations (Von Hippel, 1989), network studies (HaÊkanson, 1989; Midley et al., 1992), market conditions (Ames and Hlavacek, 1988; Ancona and Caldwell, 1992), and external knowledge infrastructures (Lundvall, 1992; Nelson, 1993). Each of these is a potential source of ``ideas, practices, and material artifacts'' for innovation. As an example, it can be argued that the richest sources for stimulating innovation are those factors that are most new. For network relationships to contribute positively to innovativeness, therefore, it may be necessary to emphasis ``weak ties'' (as opposed to ``strong ties'') because weak ties provide more diverse and rich links to the kind of novel and unique information that may be needed to generate innovative activity (Granovetter, 1973). Other external factors that emphasize newness may also be fruitful for understanding and promoting innovation. Internal factors associated with innovation might also be profitably managed by focusing on newness. Internal factors that may be critical for utilizing the innovation potential in companies include cultural factors (Hage and Dewar, 1973), structural links, i.e. information, communication and learning processes (Teece, 1986; 1988; Tushman and

Nadler, 1986), internal competencies (Drucker, 1985; Quinn, 1992), and the role of management information and communication technology (Freeman, 1991; Antonelli, 1993). Maintaining an internal awareness of the importance of newness to innovation may aid a firm's innovation efforts. For example, factors such as cultural and structural forces often tend to be impediments to innovation because they lockin traditional methods and thus shut out new ideas and practices (Miller, 1990). Even core competencies such as values, technical skills, and management systems that served a company well in the past can evolve into ``core rigidities'' that hinder effective innovation as Leonard-Barton (1992) found when studying new product development. To keep such tendencies from inhibiting innovativeness, firms need continually to infuse their knowledge sets with newness: new ideas, new skills, new personnel and new forms of organizing. Thus, by emphasizing the importance of newness to the process of innovation, managers may identify more targeted interventions that can be used to generate innovativeness more effectively. Future research should endeavor to overcome some of the limitations of this study. We relied on a single respondent to address an organization-level question; future research might seek the views of multiple organization members to reduce the potential for common method variance. Nevertheless, a data reduction technique such as factor analysis as was used here should not be troubled by the types of problems often associated with hypothesis guessing or common method variance (Cook and Campbell, 1976). Our measure of the range of radicalness was essentially dichotomous. Future research might benefit by taking a finer-grained approach to measuring the ``relevant units or adoption'', consistent with Zaltman et al.'s (1973) approach to innovation. Additionally, more specific measures of how new an innovation is perceived to be would strengthen this approach. Future research might also consider how innovation as newness is related to performance and other internal and external factors. For example, Nelson and Winter's (1982) study suggests that innovators may not be as profitable in the long run as imitators. Is there a price to pay for being the most 28

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

innovative? Or can a firm focus on the newness aspect of innovation without jeopardizing profitability? The innovationperformance relationship might also be tested with regard to potential contingencies. How, for example, do industry effects and environment relate to issues such as the newness or radicalness of an innovation? For new entrants, how important is innovativeness to launching a successful start-up? For established firms, what are the internal organizational factors that contribute to or detract from effective innovation? Such studies, by maintaining a focus on the role of newness in the process of innovation, can help researchers unravel the underlying assumptions of innovation, and help managers apply appropriate interventions to pursue innovation. To meet this challenge, we suggest that future researchers endeavor to understand more precisely these basic questions about innovation ± what is new, how new, and new to whom?

Booz, Allen & Hamilton (1982), New Product Management for the 1980s, Booz, Allen & Hamilton, New York, NY. Campbell, D.T. and Fiske, D.W. (1959), ``Convergent and discriminant validation by the multitraitmultimethod matrix'', Psychological Bulletin, Vol. 54 No. 2, pp. 81-105. Child, J. and Smith, C. (1987), ``The context and process of organizational transformation'', Journal of Management Studies, Vol. 24, pp. 565-93. Cook, T.D. and Campbell, D.T. (1976), ``The design and conduct of quasi-experiments and true experiments in field settings'', in Dunnette, M. (Ed.), Handbook of Industrial and Organizational Psychology, Rand McNally, Chicago, IL. Cooper, R.G. (1993), Winning at New Products: Accelerating the Process from Idea to Launch, 2nd ed., Addison-Wesley, Reading, MA. Covin, J.G. and Miles, M.P. (in press), ``Corporate entrepreneurship and the pursuit of competitive advantage'', Entrepreneurship: Theory & Practice. Cyert, R.M. and March, J.G. (1963), A Behavioral Theory of the Firm, Prentice-Hall, Englewood Cliffs, NJ. Daft, R.L. and Becker, S.W. (1978), The Innovative Organization, Elsevier, New York, NY. Damanpour, F. (1991), ``Organizational innovation: a meta analysis of effects of determinants and moderators'', Academy of Management Journal, Vol. 34, pp. 555-90. Damanpour, F. (1996), ``Organizational complexity and innovation: developing and testing multiple contingency models'', Management Science, Vol. 42 No. 5, pp. 693-716. Damanpour, F. and Evan, V.M. (1984), ``Organizational innovation and performance: the problem of organizational lag'', Administrative Science Quarterly, Vol. 29, pp. 392-409. D'Aveni, R. (1994), Hypercompetition: The Dynamics of Strategic Maneuvering, Basic Books, New York, NY. Dewar, R.D. and Dutton, J.E. (1986), ``The adoption of radical and incremental innovations: an empirical analysis'', Management Science, Vol. 32, pp. 1422-33. Dosi, G. (1982), ``Technological paradigms and technological trajectories: a suggested interpretation of the determinants and directions of technical change'', Research Policy, Vol. 11, pp. 147-62. Dosi, G. (1988), ``Sources, procedures, and microeconomic effects of innovation'', Journal of Economic Literature, Vol. 36, pp. 1126-71. Drazin, R. and Schoonhoven, C.B. (1996), ``Community, population, and organization effects on innovation: a multilevel perspective'', Academy of Management Journal, Vol. 39 No. 5, pp. 1065-83. Drucker, P.F. (1985), Innovation and Entrepreneurship: Practice and Principles, HarperBusiness, New York, NY. Drucker, P.F. (1993), Post-Capitalist Society, Butterworth, Heinemann, NY. Edquist, C. (1997), ``Systems of innovation approaches ± their emergence and characteristics'', in Edquist, C. (Ed.), Systems of Innovation, Pinter, London. European Commission (1991), Four Motors for Europe: An Analysis of Cross-regional Cooperation, Fast Occasional Paper no. 241, CEC, DG XII, Vol. 17.

Notes 1 Schumpeter (1934) also used ``change of market form'' as one type of innovation. We exclude this variable from our analysis as it represents a change that we consider revolutionary, hence, outside the focus of this study. 2 New services was not used originally by either Schumpeter or Kirzner, but is included here to better capture a contemporary phenomenon. This is consistent with Quinn et al. (1996) who note that 75 per cent of all economic activity now involves delivery of services and intangibles.

References Ames, B.C. and Hlavacek, J.D. (1988), Market Driven Management: Prescription for Survival in a Turbulent World, Irwin, Homewood, IL. Ancona, D.G. and Caldwell, D.F. (1992), ``Bridging the boundary: external process and performance in organization teams'', Administrative Science Quarterly, Vol. 37, pp. 634-65. Antonelli, C. (1993), ``The dynamics of technological inter-relatedness: the case of information and communication technologies'', in Foray, D. and Freeman, C. (Eds), Technology and the Wealth of Nations, Pinter, London. Barney, J.B. (1991), ``Firm resources and sustained competitive advantage'', Journal of Management, Vol. 17, pp. 99-120. Blau, P.M. and McKinley, W. (1979), ``Ideas, complexity and innovation'', Administrative Science Quarterly, Vol. 24, pp. 200-19.

29

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

European Commission (1995), Green Paper on Innovation, Brussels. Freeman, C.C. (1991), ``Networks of innovators: a synthesis of research issues'', Research Policy, Vol. 20 No. 5, pp. 499-514. Gartner, W.B. (1988), ```Who is an entrepreneur?' is the wrong question'', American Journal of Small Business, Vol. 12 No. 4, pp. 11-32. Gersick, C.J.G. (1991), ``Revolutionary change theories: a multi-level exploration of the punctuated equilibrium paradigm'', Academy of Management Review, Vol. 16, pp. 10-36. Granovetter, M. (1973), ``The strength of weak ties'', American Journal of Sociology, Vol. 78, pp. 136080. Hage, J. (1980), Theories of Organizations, Wiley, New York, NY. Hage, J. and Dewar, R. (1973), ``Elite values versus organizational structure in predicting innovation'', Administrative Science Quarterly, Vol. 18, pp. 279-90. HaÊkanson, H. (1989), Corporate Technological Behavior: Corporation and Networks, Pinter, London. Henderson, R. and Clark, K. (1990), ``Architectural innovation: the reconfiguration of existing product technologies and the failure of established firms'', Administrative Science Quarterly, Vol. 37, pp. 9-30. Jacobson, R. (1992), ``The Austrian school of strategy'', Academy of Management Review, Vol. 17 No. 4, pp. 782-807. Johannessen, J-A. and Hauan, A. (1994), ``Communication: a system theoretical point of view'', Systems Practice, Vol. 7 No. 1, pp. 63-73. Kanter, R.M. (1985), ``Supporting innovation and venture development in established companies'', Journal of Business Venturing, Vol. 1, pp. 47-60. Kim, J. and Mueller, C.W. (1978), Factor Analysis: Statistical Methods and Practical Issues, Sage, Beverly Hills, CA. Kirzner, I.M. (1976), ``On the method of Austrian economics'', in Dolan, E.G. (Ed.), The Foundation of Modern Austrian Economics, Sheed & Ward, Kansas City, MO, pp. 40-51. Kirzner, I.M. (1985), Discovery and the Capitalist Process, University of Chicago Press, Chicago, IL. Kotabe, M. and Swan, K.S. (1995), ``The role of strategic alliances in high technology new product development'', Strategic Management Journal, Vol. 16 No. 8, pp. 621-36. Lawless, M.W. and Anderson, P.C. (1996), ``Generational technological change: effects of innovation and local rivalry on performance'', Academy of Management Journal, Vol. 39 No. 5, pp. 1185-217. Leonard-Barton, D. (1992), ``Core capabilities and core rigidities: a paradox in managing new product development'', Strategic Management Journal, Vol. 13, pp. 111-25. Lumpkin, G.T. and Dess, G.G. (1996), ``Clarifying the entrepreneurial orientation construct and linking it to performance'', Academy of Management Review, Vol. 21 No. 1, pp. 135-72. Lundvall, B.AÊ. (Ed.) (1992), National Systems of Innovation, Pinter, London. Lundvall, B.AÊ. and Johnson, B. (1994), ``The learning economy'', Journal of Industry Studies, Vol. 1 No. 2, pp. 23-42.

McGrath, R.G., Tsai, M.H., Venkatraman, S. and MacMillan, I.C. (1996), ``Innovation, competitive advantage and rent: a model and a test'', Management Science, Vol. 42 No. 3, pp. 389-403. Manu, F.A. and Sriram, V. (1996), ``Innovation, marketing strategy, environment, and performance'', Journal of Business Research, Vol. 35, pp. 79-81. Midley, D.F., Morrison, P.D. and Roberts, J.H. (1992), ``The effect of network structure in industrial diffusion processes'', Research Policy, Vol. 21 No. 6, pp. 533-52. Miller, D. (1987), ``Strategy making and structure: analysis and implications for performance'', Academy of Management Journal, Vol. 30 No. 1, pp. 7-32. Miller, D. (1990), The Icarus Paradox, HarperCollins, New York, NY. Miller, D. and Friesen, P.H. (1978), ``Archetypes of strategy formulation'', Management Science, Vol. 24 No. 9, pp. 921-33. Nelson, R.R. (Ed.) (1993), National Innovation Systems, Oxford University Press, Oxford. Nelson, R.R. and Winter, S.G. (1982), An Evolutionary Theory of Economic Change, Harvard University Press, Cambridge, MA. Nohria, N. and Gulati, R. (1996), ``Is slack good or bad for innovation?'', Academy of Management Journal, Vol. 39 No. 5, pp. 1245-64. Nonaka, I. and Takeuchi, H. (1995), The Knowledge Creating Company, Oxford University Press, Oxford. Nunnally, J.C. (1978), Psychometric Theory, 2nd ed., McGraw-Hill, New York, NY. Pettigrew, A.M. (1985), The Awakening Giant: Continuity and Change in ICI, Basil Blackwell, Oxford. Quinn, J.B. (1992), Intelligent Enterprise, Free Press, New York, NY. Quinn, J.B., Anderson, P. and Finkelstein, S. (1996), ``Leveraging intellect'', Academy of Management Executive, Vol. 10 No. 3, pp. 7-27. Schumpeter, J.A. (1934), The Theory of Economic Development, Harvard University Press, Cambridge, MA. Schumpeter, J.A. (1939), Business Cycles: A Theoretical, Historical and Statistical Analysis of the Capitalist Process, McGraw-Hill, New York, NY. Schumpeter, J.A. (1942), Capitalism, Socialism, and Democracy, Harper & Brothers, New York, NY. Scott, S.G. and Bruce, R.A. (1994), ``Determinants of innovative behavior: a path model of individual innovation in the workplace'', Academy of Management Journal, Vol. 37, pp. 580-607. Slappendel, C. (1996), ``Perspectives on innovation in organizations'', Organization Studies, Vol. 17 No. 1, pp. 107-29. STEP (1995), Innovation Performance at Industry Level in Norway: Information Technology Industry, working paper, W13, p. 95. Stevenson, H.H. and Jarillo, J.C. (1990), ``A paradigm of entrepreneurship: entrepreneurial management'', Strategic Management Journal, Vol. 11, pp. 17-27. Teece, D.J. (1986), ``Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy'', in Teece, D.J. (Ed.), The Competitive Challenge: Strategies for Industrial Innovation and Renewal, Ballinger, Cambridge, MA.

30

Innovation as newness

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 20±31

Jon-Arild Johannessen, Bjùrn Olsen and G.T. Lumpkin

Teece, D.J. (1988), ``The nature and the structure of firms'', in Dosi, G., Freeman, C., Nelson, R., Silverberg, G. and Soete, L. (Eds), Technical Change and Economic Theory, Pinter, London. Teece, D.J. (1989), ``Inter-organizational requirements of the innovation process'', Managerial and Decision Economics, Special Issue, pp. 35-42. Tushman, M.L. and Nadler, D. (1986), ``Organizing for innovation'', California Management Review, Vol. 28 No. 3, pp. 74-92. Tushman, M.L. and Romanelli, E. (1985), ``Organizational evolution: a metamorphosis model of convergence and re-orientation'', in Staw, B. and Cummings, L. (Eds), Research in Organizational Behavior, Vol. 7, JAI Press, Greenwich, CT, pp. 171-22. Utterback, J.M. and Abernathy, W.J. (1975), ``A dynamic model of process and product innovation'', Omega, Vol. 3 No. 6, pp. 639-56. Van de Ven, A.A. (1986), ``Central problems in the management of innovation'', Management Science, Vol. 32 No. 5, pp. 590-607. Van de Ven, A.A., Angle, H.L. and Poole, M.S. (Eds) (1989), Research on the Management of Innovation: The Minnesota Studies, Harper & Row, New York, NY. Van de Ven, A.A. and Rogers, E.M. (1988), ``Innovations and organizations: critical perspectives'', Communication Research, Vol. 15, pp. 623-51. Van de Ven, A.A. and Poole, M.S. (1988), ``Paradoxical requirements for a theory of organizational change'', in Quinn, R.E. and Cameron, K.S. (Eds), Paradoxes and Transformation: Toward a Theory of Change in Organization and Management, Ballinger, Cambridge, MA, pp. 19-63. Vesper, K.H. (1988), ``Entrepreneurial academics ± how can we tell when the field is getting somewhere?'', Journal of Business Venturing, Vol. 3, pp. 1-10. Von Hippel, E. (1989), Sources of Innovation, Oxford, London. Walton, R.E. (1987), Innovation to Compete, Jossey-Bass, San Francisco, CA. Young, J. (1994), ``Innovate or die'', Forbes, February, Vol. 28, p. 106. Zaltman, G., Duncan, R. and Holbeck, J. (1973), Innovations and Organizations, Wiley, New York, NY.

To no To a little To some

To a great

To a very

extent

extent

extent

extent

great extent

New products

1

2

3

4

5

New services

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

1

2

3

4

5

New methods of production Opening new markets New sources of supply New ways of organizing

Additional innovation measures used in the knowledge-sector study 1. Incremental innovation. Has your company made changes during the last three years that were perceived to be new for the company, but which have previously been used by other firms, within the following areas? (Please circle one response in each row). New products New services New methods of production Opening new markets New sources of supply New ways of organizing

Yes Yes Yes Yes Yes Yes

No No No No No No

2. Radical innovation. Has your company made changes during the last three years that were perceived to be new to the industry in which the company operates, within the following areas? (Please circle one response in each row).

Appendix

New products New services New methods of production Opening new markets New sources of supply New ways of organizing

Innovation measured used in the ``general'' and ``knowledge sector'' studies Has your company made changes during the last three years that were perceived to be new for the company, within the following areas? (Please circle one number in each row).

31

Yes Yes Yes Yes Yes Yes

No No No No No No

Introduction

Factors influencing product development team satisfaction

Cross-functional teams (CFTs) are being used 70-75 per cent of the time for new product development (NPD) (Cooper and Kleinschmidt, 1994; Griffin, 1997). Despite the attractiveness of such teams, many companies and managers are finding it difficult to create teams that are effective. Effectiveness is partly a function of team members being satisfied with their team and their project. Not surprisingly, most of the elements driving team member satisfaction and thereby, team effectiveness, are controllable by senior management and team leaders. The challenge is identifying those drivers so that appropriate actions can be taken to establish teams with a greater probability of success. This article reports the results of 71 interviews with members of NPD teams from a variety of industries. Our emphasis is on the factors that influence team member satisfaction and the actions managers need to take to ensure that these drivers are fully implemented.

Gloria Barczak and David Wilemon

The authors Gloria Barczak is Associate Professor of Marketing, Northeastern University, Boston, Massachusetts, USA. David Wilemon is Director, Earl V. Snyder Innovation Management Center, Syracuse University, Syracuse, New York, USA. Keywords Teams, New product development, Effectiveness, Methodology Abstract

Methodology

The increasing use of cross-functional teams for new product development (NPD) belies the difficulty managers face in creating teams that are truly effective. Effectiveness depends, in part, on having members who are satisfied with their team and their project. This article reports the results of 71 interviews with members of NPD teams, with a particular focus on the drivers of team member satisfaction. These drivers include: team characteristics, clear project goals, clarity about evaluation and rewards, effective leadership, management support, and manageable levels of conflict and stress. To create satisfied team members, we discuss actions that can be taken at the senior management, project leader and team member level.

In-depth interviews using a structured protocol were conducted with 71 members of NPD teams in 18 diverse technology-based companies. These companies were drawn from a variety of industries including medical technology, software, electronics, industrial automation, telecommunications, biotechnology, and computer hardware. Particular firms were chosen due to their long-time use of cross-functional NPD teams, their accessibility to the researchers, and their willingness to participate in the study. Individual team members were selected for similar reasons. All interviews were taperecorded and transcribed. The average length of each interview was about one and a half hours. All individuals and companies were based in the USA. About 68 percent of respondents were from engineering while the remaining 32 percent of team members were from other functional areas such as marketing, quality assurance, and manufacturing. About 54 percent of team members had bachelor's degrees while 34 percent had Master's degrees. Team members had been with their company an

Electronic access The research register for this journal is available at http://www.mcbup.com/research_registers The current issue and full text archive of this journal is available at http://www.emerald-library.com/ft

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . pp. 32±36 # MCB University Press . ISSN 1460-1060

32

Factors influencing product development team satisfaction

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 32±36

Gloria Barczak and David Wilemon

average of ten years. The average length of an NPD project was 1.8 years with a range from one to five years.

others on the team. Thus, team members need to possess several different interpersonal skills as well as functional expertise to be effective. (2) Setting clear project goals that are communicated to and understood by team members pays off in greater focus and thus, greater satisfaction. Lack of clear project goals and continual shifting of goals makes it difficult for the team to know how to proceed. It leads to frustration and dissatisfaction and ultimately, less successful project outcomes. Team members desire to know the project's purpose, the goals of the project as well as their role in accomplishing these. Moreover, they need stable goals, which they can continually use as a guide to keep them on-track and focused. (3) Clearly articulating how team members are evaluated and rewarded is crucial to team member satisfaction. Most NPD team members are uncertain as to how they are evaluated and/or rewarded. Firms rarely specify NPD or any other performance evaluation criteria or the reward for good performance. Although the types of reviews and rewards vary, we found that team members are most frustrated by the lack of knowledge and understanding they have about the evaluation and reward systems for new product development work. To do their jobs well, NPD team members require clear communication of evaluation and reward criteria, mechanisms and procedures. Members need to know who is doing the evaluation (i.e. their functional manager or the NPD project leader), what the criteria are (i.e. contribution to functional group or contribution to the NPD project/team), and the specifics of the rewards for good performance (i.e. keeping one's job, bonus/raise, recognition, attaboys). (4) Effective team leaders possess a variety of skills important to enhancing team member satisfaction. One important characteristic of an effective leader is the ability to manage the team. This includes taking actions that serve to motivate, coordinate and facilitate team members in their efforts.

Drivers of team member satisfaction Six key factors from our study were identified as drivers of team member satisfaction. They are: (1) team characteristics; (2) clear project goals; (3) clarity about evaluation and rewards; (4) effective leadership; (5) management support; (6) manageable levels of conflict and stress. Let us now consider each driver with evidence of both good and poor practice within the companies studied and the implications for management: (1) Team characteristics are critical to team member satisfaction. Positive characteristics such as sharing a common purpose, having a team orientation and members with the ``right'' skill mix lead to greater satisfaction. Team members who have a shared understanding of a common goal are aligned in their efforts to achieve that goal. Clear roles and responsibilities help keep each member focused on their specific task and provide a sense of how their task fits into the ``big picture''. Teams require members who are cooperative and get along with others. They thrive with members who put the needs of the team and the project before their own needs and agendas. Effective team members possess a set of skills that facilitate their functioning in a team environment. Although expertise in one's functional area is necessary, it is not sufficient for being a member of a team. We found that team members also need to have effective work-styles which means they have a strong work ethic, are disciplined, determined and motivated. In other words, effective team members are committed to a common objective and will do what it takes to achieve that objective. Strong communication skills are also important to effective team members. Our interviews noted that members must be willing and able to share ideas, listen, and be open-minded about the views of 33

Factors influencing product development team satisfaction

Gloria Barczak and David Wilemon

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 32±36

Leaders also need skill in managing the process/project itself. It is important that they set schedules, define roles and responsibilities, organize and manage meetings, establish budgets and review performance. Personal qualities of the leader are also important to team member satisfaction. Traits such as being approachable, respected, motivated, visionary and easy to work with are qualities highly regarded by team members. Leaders who possess such valued personal qualities may command more respect and cooperation, thereby enhancing the team's effectiveness. In addition, such leaders may attract individuals with the proper mix of task and interpersonal skills to improve the team's performance Although leaders need strong technical skills, these alone are not sufficient for effectiveness. Thus, leaders need to possess three sets of skills: interpersonal, project management, and technical. Through effective utilization of these skills, the team leader is able to motivate team members, generate involvement, manage conflict constructively, manage the environment of the team, and overall, achieve continuous levels of satisfactory behavior. (5) Support from senior management in the form of clear goals and priorities, adequate resources, appropriate schedule, and a sense of commitment to the project engender satisfaction. Lack of senior management support is a common source of team member dissatisfaction. Support deficiencies include having unclear goals and priorities, inadequate resources, unattainable schedules, and a lack of commitment by senior management. We found that one of the greatest frustrations of team members is thinking that their project is not of value to the company or of interest to senior management. Team members need to know that management supports their project. This must be done not simply through words but through actions. For example, management can show commitment by knowing the status of the project, talking with team members about problems and issues, and by providing requisite resources when a strong, logical rationale is offered.

(6) Manageable levels of conflict reduce the amount of stress experienced by team members and, thus, increase satisfaction. The primary conflict experienced by NPD teams is conflict between them and senior management or between the team and other organizational units (functional departments). Our interviews indicated that friction with senior management centers around policies and procedures, support, and resources over which the team has little control. For example, the project schedule is a primary source of conflict between the team and senior management. This conflict leads to a high degree of stress on individual team members and the team to get the project completed quickly in spite of resource constraints and technical difficulties. Discord between the teams and other organizational units generally takes the form of politics or turf battles. Team members also experience conflict around teamwork. Strife around teamwork stems from interpersonal conflicts and ineffective communication between members as well as having members with ``big egos'' who put themselves before the team and/or the project. Such conflicts can lead to finger pointing; members not pulling their weight on the team; and withdrawal by some team members. These interpersonal conflicts lead to stress. Thus teamwork ± learning how to work together as a team ± can be a major source of stress. Another source of conflict emanates from the task itself. Shifting goals, priorities and schedules cause conflict. So do roadblocks such as not being able to solve technical issues quickly or completely. We found that conflict between the team and management, between the team and other units, and between team members has a greater impact on team members than conflict about the task. However, too much conflict can result in negative feelings about the project, frustration, stress and apathy. These feelings, in turn, often negatively affect morale and commitment to the project. Negative feelings about the project can also affect the quality of work. Thus, conflict can lead to poor direction, poor coordination and communication 34

Factors influencing product development team satisfaction

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 32±36

Gloria Barczak and David Wilemon

between members, lower quality work, and more mistakes and errors. Clearly, these negative results can create even more conflict resulting in even lower quality project work. It is important, therefore, for team members and their leaders to deal with the conflict so that this iterative cycle does not spiral into deeper and deeper negativity.

.

.

What can be done? To create satisfied team members, several actions must be taken at the senior management, project leader and team member level. Specific recommendations at each level are discussed below. Implications for senior management Build an organizational culture that supports teamwork. This can be achieved by: creating models of effective teamwork and sharing exemplary stories throughout the organization, emphasizing crossfunctional integration, recognizing and rewarding cooperation, and providing training in teamwork skills. . Support NPD projects by setting clear goals, providing resources, and protecting teams from political in-fighting. Failure to provide these elements can lead to conflict and feelings of frustration and dissatisfaction on the part of team members. This, in turn, can lead to negative attitudes and behaviors when individuals are assigned to a team. . Ensure that team leaders possess three sets of critical skills: interpersonal, project management and technical. As a first step, management needs to consider carefully whom they place into team leadership positions. In addition, however, as teams become more prevalent, training needs to be provided to various levels of managers for the purposes of building and enhancing their interpersonal as well as project management skills. . Reconsider hiring criteria to include interpersonal skills along with functional expertise. Both qualities can be used for evaluating NPD team leaders as well as team members. By including these criteria in their hiring and performance evaluation processes, management sends

a clear signal that both skills are important. Improve how team members are evaluated and rewarded for their NPD efforts and clearly communicate this to NPD personnel. Although the intrinsic challenge of an NPD project can be highly rewarding, team members' comments suggest that knowledge of how they are evaluated and rewarded is important to their behavior and motivation towards the project. Be aware of the causes of stress and develop activities to help team members deal with stress. The fact that the very process of team membership is stressful to team members implies that activities that focus on team building, communication, stress management, and healthy living can help members manage stress.

Implications for team leaders . Develop, manage, and sustain the team's relationship with senior management. Since disagreements about organizational issues create conflict and affect team members' satisfaction, engage in activities aimed at promoting and generating support for the project. To do this, one must possess strong interpersonal skills which allows him/her to forge positive relationships with senior management, with other functional groups, and with customers. . Understand the internal dynamics of crossfunctional teams and take appropriate actions to manage these relationships. Since CFTs require strong interpersonal skills as well as functional expertise, special care should be taken when selecting/ assigning individuals to particular teams. Although it is not always possible or practical to consider carefully the mix of people on a team, evidence indicates that selecting the right team members can have significant benefits. More importantly, however, since individuals often need help being part of a team, team leaders need to take responsibility for securing and providing training in teamwork to NPD team members. Particular skills which team members need to develop include: conflict resolution, stress management, negotiation and communication. Training in these interpersonal skills can lead to better internal team functioning

.

35

Factors influencing product development team satisfaction

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 32±36

Gloria Barczak and David Wilemon

.

which, in turn, can result in the team reaching its goals. Reconsider and communicate the criteria used for performance evaluation. Since many team members are uncertain about how they are evaluated, leaders need to articulate clear performance criteria. These criteria should include assessments of interpersonal skills (ability to work with others) as well as functional expertise. .

Implications for team members . Possess the capabilities needed to be an effective team member. If management is unwilling or unable to take action on this issue, find ways to develop your own teamwork capabilities. For example, team members can: seek out and participate in training seminars in areas such as team building and interpersonal communication and identify and develop relationships with others who have strong interpersonal and team skills and solicit them as a mentor. . Recognize that strong communication skills are not sufficient to be an effective team

member. This does not suggest that communication is not important to member or team effectiveness. Rather, it indicates that effective communication needs to be integrated with a willingness to support others as well as to work towards common goals. By displaying these characteristics, team members show their willingness and ability to truly be a ``team''. Proactively acquire what is necessary for project success. If goals or performance criteria or rewards are not clear, seek clarity from team leaders and managers.

References Cooper, R. and Kleinschmidt, E.J. (1994), ``Determinants of timeliness in product development'', Journal of Product Innovation Management, Vol. 11, pp. 381-96. Griffin, A. (1997), ``The effect of process and teams on product development cycle time'', Journal of Marketing Research, Vol. 34, pp. 24-35.

36

1. Introduction

Implementation of success factors in new product development ± the missing links?

Both academia and practitioners acknowledge that new product development (NPD) is a crucial activity to most companies in order for them to secure longterm survival and growth (Brown and Eisenhardt, 1995; Clark and Fujimoto, 1991; Craig and Hart, 1992). Product development, however, is both expensive and risky with a large portion of product development budgets going towards new product failures (Page, 1993). For years researchers have therefore been interested in explaining what constitutes new product success. They have investigated factors leading to success and failure of NPD in an effort to uncover factors that can enhance product development success in companies. Since the mid-1950s research has been published in a large and continuous number of studies probing into the factors of success and/or failure for NPD. More than 200 studies have been carried out in various industries, geographical settings, and also with various methodological approaches, all sharing the objective of understanding new product success and failure and deriving normative implications for companies. Despite the fact that the studies point to a fairly consistent list of success factors, it seems that only a few companies have implemented these identified success factors ± indicated by the fact that recent studies show companies to make the same mistakes they did 30 years ago. By focusing on aspects relating to companies, a few researchers have tried to look into why the identified success factors do not appear in most new product development projects. For instance, Cooper (1999) tries to outline company specific reasons why business managers fail to heed researchers' normative advice and why businesses fail to implement success factors. Another approach to the puzzle is to take a look at the research conducted and the success factors identified in order to determine whether researchers can do anything to improve the identified success factors, or make them more accessible to business. The purpose of this research project is precisely that, which means that the purpose of the study is to find relevant areas which are not adequately covered by the literature on success factors. This is done

Bjarne Jensen and Hanne Harmsen

The authors Bjarne Jensen is a PhD student and Hanne Harmsen is an Associate Professor, both at the MAPP Centre (Market based Process and product Innovation), Aarhus School of Business, Denmark. Keywords Implementation, New product development, Success, Competences Abstract This paper addresses companies' lack of implementation of success factors in new product development. Drawing on theory in the competence perspective and an exploratory empirical study, the paper points to two major areas that have not been covered by previous studies on new product development success factors. The two areas are knowledge and skills of individual employees, values and norms and it is suggested that increased understanding of these two areas holds potential in making identified success factors more accessible to companies. Electronic access The research register for this journal is available at http://www.mcbup.com/research_registers The current issue and full text archive of this journal is available at http://www.emerald-library.com/ft

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . pp. 37±52 # MCB University Press . ISSN 1460-1060

37

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

by scrutinizing the product development literature on success factors, the literature on company competences and through empirical investigations in four successful Danish food companies.

2. The literature review This section discusses first the literature on NPD success, second the literature on competences and finally the NPD success factors are looked at from a competence perspective. 2.1 No success from new product development success factors According to Montoya-Weiss and Calantone (1994), empirical studies that deal with determinants of new product performance can be divided into three overall categories: (1) studies that identify factors leading to success; (2) studies that identify factors leading to failure; and (3) studies that distinguish between success and failure, with increased focus in recent years on the third category. In short, the methods used in NPD research are primarily inductive, i.e. researchers ask managers to identify a number of factors underlying success or failure of product development. The studies of these three categories have led to the identification of a number of determinants of new product performance, and most often researchers suggest normative implications of how to enhance the chances of success or how to avoid failures in NPD. Generally, the normative implications of the empirical studies are a list of key success factors in order of priority and despite the diversity of designs, methods, and operationalisations applied, several literature reviews point out that the findings are consistent and similar across the various studies (Barclay, 1992; Brown and Eisenhardt, 1995; Cooper and Kleinschmidt, 1987; Craig and Hart, 1992; Harmsen, 1992; Lilien and Yoon, 1989; Montoya-Weiss and Calantone, 1994; Rothwell, 1977; Schewe, 1991). According to Craig and Hart (1992), one explanation of the consistency and similarity is that the various studies have a tendency to 38

build on factors identified in previous studies, the consequence being that the same variables are researched over and over again which leaves little chance of discovering new explanatory factors (Harmsen, 1992). Therefore, the conclusion of the literature review is that the NPD literature on success factors seems to agree on which factors will enhance the chances of success. However, other studies show that company practice has not changed much since the first results on success factors were published and that to a large extent companies still face the same problems and make the same mistakes that they have made for many years (Cooper, 1998; 1997; Cooper and Kleinschmidt, 1995; 1986; Craig and Hart, 1992; Page, 1993). This suggests that companies have not been able to implement the normative advice. Other studies have actually shown that companies have great difficulties in implementing the success factors in their new product development practice (Barclay, 1992; Biemans and Harmsen, 1994; 1995). This has been explained by lack of operational, normative implications and general barriers to change in companies. A related issue to the lack of operational implications is whether the success factors were initially identified in a manner that makes it possible to present operational normative advice. Critical voices point out that the NPD research focuses solely on identifying the factors leading to success and failure (Foss and Harmsen, 1996; Montoya-Weiss and Calantone, 1994), that often the success factors can be characterised as descriptors (an example is the typical success factor: ``the market should perceive the product as better as competitors' products''), and that the normative implications are very brief. The question ± how do companies go about implementation? ± has not really been addressed by the literature (O'Connor, 1994) ± an exception is Cooper (1990). Researchers have not been concerned with how to operationalise the success factors or how companies can improve their product development process by implementing the identified success factors (Foss and Harmsen, 1996; Harmsen, 1996). Foss and Harmsen (1996) further state that, due to the lack of detail, the identified success factors can only serve as a check-list

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

encompassing the necessary activities and resources for NPD success as the results say nothing about how to acquire superiority. Considering the interest ``successful product development'' has attracted, it is somewhat surprising that such limited attention has been paid to applying the research results and that companies have been unable to implement the success factors. This situation is less than flattering for NPD researchers. Therefore, we set out to look into possible areas which have not been well covered by the NPD literature on success factors to see whether NPD researchers can do anything to improve or make the identified success factors more accessible to business. To that end, we tried to link concepts of competence theory with the NPD literature on success factors. The competence perspective is particularly appealing for this purpose, since the perspective like the product development literature on success factors stresses that company internal factors are the main drivers of company success. Furthermore, compared to the new product literature on success factors, the competence perspective in general has a larger focus on the mechanisms and elements that create company action, competitive advantage and thus company success. 2.2 The competence perspective: where would they seek success factors? The competence perspective can be traced back to the resource-based theory of the firm which arose in the wake of Penrose's (1959) work. In short, the resource-based view of the firm is, like the NPD literature on success factors, guided by the research question of why some firms are persistently more successful than others. In contrast to the NPD literature the results of resource-based view of the firm are deduced from theory. The resource-based view of the firm has been further developed by researchers like Wernerfelt (1984) and Barney (1991). However, it never really ``caught on'' to practitioners and it was not until Prahalad and Hamel (1990) published their article on core competence that the ``thoughts'' gained in popularity (Wernerfelt, 1995). Since 1990 the competence perspective has gained increasing interest and popularity among practitioners and researchers in the strategic management literature (Prahalad and Hamel, 39

1990; Hamel and Heene, 1994; Teece et al., 1990). The competence perspective is still in its infancy and therefore the perspective on concepts and terms is not all that clear. This is evident in that researchers in the competence perspective have many suggestions on how to classify and define a competence. In that connection, concepts and terms like assets, capabilities, skills, resources, competences, knowledge and management processes are often mentioned and defined (LeonardBarton, 1992; Sanchez et al., 1996). Furthermore, the competence perspective faces some problems related to aspects such as to grasp and describe more precisely the concept of competence, to operationalise the concept, and to show how competences can be identified. These problems have had implications for our empirical study as explained in section 3. A few researchers in the competence perspective have provided frameworks for what actually constitutes a company's core competences. Leonard-Barton (1992) adopts a knowledge-based view of the firm and views a core capability as the knowledge set that distinguishes and provides a firm with a competitive advantage. Leonard-Barton's (1992) knowledge set consists of four different knowledge dimensions: (1) Knowledge and skills embodied in employees. This encompasses both firm-specific techniques and scientific understanding. (2) Employee knowledge and skills embedded in technical systems. These are the results of accumulated, codified and structured knowledge in people's heads over the years. Often knowledge and skills are laid down in systems, routines, procedures and tools. (3) The managerial system is the creation and control of knowledge. It represents both formal and informal ways of creating and controlling knowledge. Examples are reward systems and various kinds of report systems. (4) The fourth dimension, values and norms, is infused through the other three dimensions. These include ``values assigned within the company to the content and structure of knowledge (e.g. chemical engineering versus marketing expertise), means of collecting knowledge, and controlling knowledge (e.g. individual empowerment

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

versus management hierarchies)'' (Leonard-Barton, 1992, p. 114).

related to strategy, NPD management and company characteristics predominantly cover aspects of the knowledge dimension related to managerial systems. Success factors related to the NPD process, people and information primarily cover technical systems and managerial systems and a few are related to knowledge and employee skills. This suggests that studies on successful product development are comprehensive with regard to understanding systems supporting NPD success. The concrete activities, organisational structures and descriptors of successful product development have been identified in some detail. In Table I, these factors correspond to technical systems and managerial systems. One reason for the predominance of these two knowledge dimensions might be that they are easier to identify across companies due to their general nature. However, according to the competence perspective and Leonard-Barton (1992), one must take in all four knowledge dimensions to understand variance in company performance and success. Furthermore, researchers in the competence perspective would argue that these two knowledge dimensions would be relatively easy for other companies to copy, and, as a consequence, they cannot alone explain differences in performance and success (Barney, 1991; Prahalad and Hamel, 1990). Knowledge and skills embedded in employees as well as values and norms are the two dimensions poorly represented in the literature on NPD success as shown in Table I. The understanding of NPD literature on knowledge and skills embedded in employees is limited to the importance of cross-functional co-ordination of information and the various roles of staff involved in NPD. Few researchers have been interested in understanding the role and importance of the individual employee (as opposed to manager) knowledge and skill with regard to cross-functional co-operation, and as a consequence the understanding of this knowledge dimension is limited, rather general and cursorily appreciated. The same goes for understanding the importance of individual employee roles. In those instances where roles and skills are researched, it is typically related to the project leader.

Leonard-Barton (1992) thus argues that core capabilities are in fact different forms of knowledge accumulated over time and sees this knowledge as interrelated and interdependent. How much each of the four knowledge dimensions contributes to a given competence will depend on the competence and the company. From this also follows that it makes little sense to try to capture a capability, like for instance a product development capability, without covering these four dimensions. This also means that, compared to new product development, Leonard-Barton (1992) has a slightly different view on what constitutes successful NPD competence. How different or similar is discussed in the next section. 2.3 A rejoinder to the new product success factors. What is missing? In this section, we apply concepts from the competence perspective to the NPD literature in order to find possible areas which are not sufficiently covered by the new product literature on success factors. More specifically, Leonard-Barton's (1992) four knowledge dimensions are compared to the new product success factors. As most of the literature reviews on new product success factors have reached consistent conclusions regarding the most important success factors, we use Craig and Hart's (1992) literature review as the basis for comparing identified NPD success factors with the competence perspective. As in most other literature reviews Craig and Hart (1992) have identified groups of similar success factors. They identified six groups of success factors: (1) management; (2) process; (3) company; (4) people; (5) strategy; and (6) information. In Table I the new product success factors, represented by Craig and Hart's (1992) dimensions, are compared to LeonardBarton's (1992) four dimensions. Table I compares Craig and Hart's (1992) six groups of success factors with LeonardBarton's (1992) four knowledge dimensions. As Table I illustrates, the success factors 40

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Table I Success factors in new product development and four knowledge dimensions of competence Critical success factors in NPD

Knowledge and skills

Technical systems

Strategy

NPD management

Company characteristics NPD process

Initial screening Market assessment Technical assessment Market study Business analysis Product development In-house product test Test market Trial production Pre-business analysis Production Launch Functional co-ordination R&D/marketing co-ordination Cross-functional communication

People

Information

Roles: Idea generator Entrepreneurs and champions Project leader Gatekeeper Sponsor Functional co-ordination of information

Functional co-ordination of information

With regard to values and norms, the understanding is limited to strategic orientation, risk taking and climate. Examples of research on values and norms are Gupta and Wilemon (1988) (risk taking and top management support), McDonough (1986)

Managerial systems NPD as a part of corporate strategy Strategic orientation Technology and marketing Proactive Product differentiation Synergy Risk taking Top management support Balanced managerial orientation Involving top management Top management roles (climate) Organisation ± structure and design Initial screening Market assessment Technical assessment

Values and norms

Strategic orientation Proactive

Risk taking Top management support

Top management roles (climate)

Business analysis Product development Test market Pre-business analysis

Functional co-ordination R&D/marketing co-ordination Cross-functional communication Organisational structure Project management

Functional co-ordination of information

(top management support), Voss (1985) (risk taking). In sum, few studies have touched upon the importance of values and norms with the consequence that the understanding is limited, rather general and not well understood. 41

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Summing up, looking at the groups of success factors identified by Craig and Hart (1992) in Table I, it seems that in the NPD literature the understanding of the role and importance of knowledge and skills embedded in employees, and values and norms in connection with NPD success is limited. Since all four dimensions are necessary for success, a thorough understanding of all four knowledge dimensions is required in order to understand a core competence. Therefore, improving the understanding of the two knowledge dimensions (knowledge and skills embedded in employees, and values and norms) shows the way to understanding what constitutes successful product development.

structure their product development competence. The theoretical rationale for taking point of departure in managers' subjective understanding is grounded in managerial cognition theory, which has its origin in cognitive psychology. Briefly explained, cognitive psychology says that actions should be understood as directed by goals. That is, in a given situation a person will choose the action that leads to fulfilment of desired goals (Anderson, 1983; Huff, 1990). In order to uncover managers' understanding of NPD competences we used the laddering method, which is a tested and theory-based method (cognitive psychology) in consumer behaviour. The method explicitly connects means and ends in networks. The laddering method is therefore intuitively a suitable method for uncovering competence structures as several researchers have argued that competences are made up of bundles or networks of various organisational factors like knowledge, skills and processes (see for example, Day, 1994; Durand and CreÂmadez, 1999; Sanchez et al., 1996; Leonard-Barton, 1992). Since our point of departure is managers' subjective understanding and knowledge of competences, it is important to limit our role and influence as interviewers as much as possible. This makes the laddering method particularly attractive as stated by Grunert and Sùrensen (1996): ``Laddering's most important virtue is that the collection of raw materials is driven by the respondent's cognitive structure and processes rather than by those of the researcher''. In other words, it is the manager's own understanding of causes and effects that guides the actual interview. More precisely, the interview procedure has been the following: we started with the product development competence and asked the respondent ``What is it that makes the company good at product development?'' Typically, the respondent would mention three to five factors or competence elements (e.g. market orientation). We then followed up on each of these competence elements by asking ``what is it that makes the company good at this competence element?'' (e.g. what is it that makes the company good at market orientation?) and so on. We followed up on each competence element until the respondent was unable to elaborate on the

3. The empirical investigation of companies' product development competence The purpose of the empirical part is to investigate how successful food companies structure their product development competence. By focusing on competences, we aim at embracing as much as possible of the companies' ability to develop new products. The primary aims are partly to determine whether Leonard-Barton's (1992) four knowledge dimensions do indeed apply in a product development context and partly to determine possible areas in companies' product development practice which have not been covered adequately by the NPD literature on success factors. 3.1 Methodology As our interest lies in companies' view on product development competences we applied an inductive approach implying that we do not have an ex ante definition of product development competence. Our starting point is that a product development competence is something that a company has or is able to do. Furthermore, analogously to Day (1994), Durand and CreÂmadez (1999), Eriksen and Foss (1997) and Sanchez et al. (1996) we regard product development competences as entities consisting of networks. Our inductive approach also means that our point of departure is managers' subjective understanding of how companies 42

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

competence element. This enabled us to draw up chains containing various competence elements and showing the relationship between them, i.e. we were able to map the structure of the companies' product development competence. The maps show how various competence elements through interrelations and interdependence together form the product development competence in the four companies. We investigated the product development competence in four successful, medium-sized food companies. They were primarily selected because of their excellent product development competence. In each food company, we interviewed the employee with the greatest and deepest understanding of the company's product development competence. The managing directors of the companies selected their particular ``expert'' employee, as follows: (1) Company A ± sales and marketing manager. (2) Company B ± product manager. (3) Company C ± marketing manager. (4) Company D ± product development manager.

As can be seen from Figure 1, there are three competence elements that, according to the sales and marketing director, directly support company A's product development competence. These three elements are briefly touched on in the following. The company's consumer focus is an important competence element in connection with the company's product development. The sales and marketing director points out that consumers play no role in the actual idea generation phase but that consumers are important in all the remaining phases where the company conducts a number of market analyses in order to secure that the new product meets consumer wants and wishes. Therefore, the ability to conduct market analyses is decisive for the company's consumer focus. This ability, together with employees' ability to put themselves in the consumers' shoes, secures the consumer focus in the NPD. The company's ability to generate new ideas is also a decisive competence element in its NPD. The sales and marketing manager states that of the three competence elements that support its ability to generate new ideas the company's ability to take point of departure in new product possibilities and trends on other markets (typically South Europe and the USA) is especially important. The third and last competence element that directly supports the company's product development competence is the company's ability to balance technical and commercial development. As Figure 1 shows the ability to balance these two aspects is supported by a product development committee (consisting of the sales and marketing director and the technical director) and cross-functional project teams. This competence element is supported by two other and more specific competence elements. The first is the company's ability to ensure that employees from various functions are represented in each project from day one. The second is that the project leader always is the product manager. And as can be seen from the map in Figure 1 the product manager's personal skills and knowledge are specified as necessary elements that indirectly support the company's product development competence.

The next section presents and discusses the results of the empirical investigation. 3.2 Results The results of the empirical investigation are shown in four competence maps of how the product development competence is structured. The maps show the direct and indirect elements supporting the product development competences. This section presents two of the four product development competence maps and discusses these in relation to previously identified NPD success factors. Since all four maps give rise to the same comments in relation to previously identified NPD success factors the remaining two maps are illustrated in the Appendix and serve mainly as a support of the discussion put forward in connection with the two competence maps presented in the following. The first map is shown in Figure 1. The company's product development function refers formally to both the marketing and technical functions as the company believe the two functions that play an equally important role in the development of new products. The expert respondent is the sales and marketing manager. 43

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Figure 1 The product development competence of company A

Relating the map in Figure 1 to Craig and Hart's (1992) six groups of success factors, it can be seen that many of their groups of success factors appear in the company's product development competence. However, it is also evident that not all of the identified competence elements can be categorised into the six groups of success factors. The square and circles indicate competence elements that do not fit easily into any of Craig and Hart's (1992) six groups of success factors. The squares refer to competence elements that are related to employees' understanding and qualifications (according to Craig and Hart (1992) there are few studies focusing on aspects of ``people'' in NPD and the understanding of ``people'' in new product development literature is indeed limited). The circles refer to competence elements that are related to values and norms in Leonard-Barton's (1992) terminology (these competence elements do not really fit into any of Craig and Hart's (1992) groups). Let us now look at the product development competence of company B as shown in Figures 2 and 3. Company B's 44

product development function refers formally to the marketing function and the expert respondent is the product manager. In brief, Figure 2 shows the four competence elements that directly support the company's product development competence. These are: (1) project formulations and action plans; (2) market orientation; (3) know-how and market knowledge; (4) clear goals for product development. Figures 2 and 3 show that other and more specific competence elements support each of the four competence elements. Looking at Figure 2, the company's ability to formulate projects and action plans is very important as this is how the company ensures that the stage for the project's further progression is defined. It is important that the project formulation and action plans are in writing as it helps all involved parties work toward the same goal. Another competence element is the company's market orientation. A decisive success criterion for the company is that its products provide value to consumers. Therefore, it is very important that the

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Figure 2 Company B's product development competence (continues in Figure 3)

Figure 3 Company B's product development competence (continued from Figure 2)

company involves consumers from the very beginning of new projects ± preferably basing idea generation on consumer demands and needs. The third competence element that directly supports the product development competence is the company's ability to follow the clear objectives set for its product

development ± both short- and long-term objectives. To a large extent this competence element rests on personal characteristics of the product manager who typically also is the project leader. Finally, Figure 3 illustrates the competence structure related to the last of the four competence elements that directly supports 45

Implementation of success factors in new product development

Bjarne Jensen and Hanne Harmsen

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

the company's product development competence. This competence element, know-how and market knowledge, is dominated by two other blocks of competence elements. The first block is related to the company's ability to analyse the market. Looking at that block of competence elements one sees that the competence element is supported by, among other elements, its ability to buy external market analyses. As the company does not have specialists that can conduct all the necessary market analyses, the company's ability to buy market analyses becomes crucial in connection with product development. In this connection, it is decisive that the company is able to find agencies with complementary competences. The respondent emphasises this aspect, as the company's product development is very much based in emotions and intangibles. The company as well as consumers find it difficult to express these aspects. Therefore, the quality of the analyses rest on the company, together with the agencies, spending much time and many resources discussing how best to ask the consumers about the intangible aspects and emotions. The other block of competence elements in Figure 3 is related to know-how and market knowledge embedded in product managers. According to the respondent, it is decisive that the product manager as project leader is well-informed about the product and the process ± without such knowledge the product manager has no background for conducting excellent product development. Relating the structure of company B's product development competence to Craig and Hart's (1992) six groups of success factors reveal a number of the NPD success factors in the map. However, again a number of competence elements do not fit easily into any of Craig and Hart's six groups. The squares and circles indicate these competence elements. The squares refer to competence elements that are related to employees' understanding and qualifications (NPD literature has limited understanding of the role of this competence element). The circles refer to competence elements that are related to values and norms in Leonard-Barton's (1992) terminology (NPD literature has no or very

limited understanding of the role of this competence element). Especially the squares are quiet salient and suggest that employee knowledge and skills are decisive in ensuring the distinguishing character of the company's product development competence. Comparing the two companies' product development, it is evident that the product development competence of both companies include some of the identified NPD success factors. However, it is also evident that the companies structure these success factors in different ways. Two of the prevalent differences are their idea generation and the way they co-ordinate and manage the individual projects. Company A primarily generates ideas based on product possibilities (product focus) and trends on other markets as the company believes that consumers are unable to articulate latent needs. Company B, on the other hand, emphasises the importance of generating ideas based on consumer needs and wants, and the company believes that consumers are indeed able to articulate needs and wants, the consequence of the two quite different approaches being that company A to a much larger extent develops new-fromscratch products (in a market related sense) whereas company B primarily develops improvements and variations of its existing products. With regard to co-ordination and management of projects, in company A the product development committee (consisting of the marketing and the technical directors) decides, co-ordinates and manages all projects from idea to launch. In company B, co-ordination and management is left to the project leaders. Within the frames of the overall product development strategy, the project leaders co-ordinate and manage their own projects. The two remaining maps in the Appendix also reveal that the product development competence of companies C and D include some of the identified NPD success factors and it is also clear that they each structure the product development competence differently. Comparing all four maps, we see that the four companies structure product development quite differently, which, for one thing, indicates that there is ``room'' for building up product development differently and still ± or maybe rather therefore ± be successful in developing new products. If one looks at the structure of the four competences, it is also

46

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

evident that they are rather complex entities made up of a number of different competence elements. As stated, to various degrees the competence structures of the four maps are made up of elements that correspond to Leonard-Barton's (1992) four knowledge dimensions. Therefore, the empirical investigation leads to the same conclusions as the theoretical investigation when comparing the empirical results with the new product development literature on success factors. Furthermore, the maps show that these four types of elements are interrelated and interdependent which points to the fact that it might be necessary to have a thorough understanding of all four competence elements in order to understand successful product development. Furthermore, the competence perspective and the empirical study suggest that in order actually to change the product development competence, one needs to have a thorough understanding of all four knowledge dimensions and not least of the importance and role of the knowledge dimensions related to employee knowledge and skills and values and norms. However, it is also clear that these two dimensions are the hardest to investigate. Knowledge and skills are hard to investigate as this kind of knowledge is embedded in employees and therefore most often is implicit. Values and norms are maybe even harder to investigate as the knowledge entailed in this knowledge dimension is implicit and hard to specify. The complex and implicit nature of these two knowledge dimensions seriously questions whether NPD researchers can improve their understanding of these two knowledge dimensions and whether such an improved understanding actually can be generalised. However, this should not be a convenient excuse for the NPD researcher as it seems that understanding of the two knowledge dimensions is important in understanding what constitutes successful product development. Also, we believe as some researchers already have shown that it is possible at a higher level of abstraction to generalise such understanding. We acknowledge that it is not an easy task and the real challenge for NPD researchers will be to find ways of operationalising the understanding of the two knowledge dimensions. 47

In sum, by relating these four dimensions to the literature on NPD, it becomes obvious that there is a thorough understanding of the importance and the role of the knowledge dimensions related to technical and managerial systems, whereas both the theoretical and the empirical studies reveal that understanding the importance and the role of the two other knowledge dimensions ± employee knowledge and skills, and values and norms ± is rather limited or even absent. This observation might also provide insightful understanding of why companies experience difficulties in implementing the identified success factors as it is clear that all four knowledge dimensions are important. Furthermore, in order actually to change the product development competence, one must have a thorough understanding of all four knowledge dimensions, especially of the importance and role of the knowledge dimensions related to employee knowledge and skills and values and norms. The NPD literature demonstrates little understanding of these two knowledge dimensions, even though improved understanding might hold the key to making the identified success factors more relevant for companies.

4. Conclusion and implications The purpose of this research project was to find areas not adequately covered by the new product development literature on success factors. We did this by looking into the product development literature on success factors, the literature on company competences and through empirical investigations of the product development competence of four successful Danish food companies. The NPD literature seems to agree on which factors lead to NPD success. Despite the diversity of designs, methods, and operationalisations applied, the findings showing factors leading to success are consistent and similar across various studies. Generally, the normative implications of the empirical studies are a list of key success factors in order of priority. The literature study also reveals that success factors have only been sparringly implemented in companies' NPD practice. Part of the explanation is to be found in the

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

literature not having devoted much attention to how companies can implement these factors, and another part in the fact that the identified success factors are not operational in nature. Rather, researchers have concentrated on identifying success factors with the consequence that the normative implications are very brief and incomplete. Drawing on theory in the competence perspective, we argue that by looking for factors leading to success across companies, maybe the identified factors are not the ``real'' explanations of success. We pointed out theories in the competence perspective question whether the identified success factors are real factors distinguishing success from less successful product development practice. Two aspects underline this doubt: that the identified success factors are common to a large number of companies and that they are fairly easy to copy. With the goal of revealing possible areas that have not been adequately covered by the NPD literature on success factors, we looked further into theories in the competence perspective and investigated empirically the structure of four successful Danish food companies' product development competence. Both the theoretical and the empirical part point towards the potential benefits of improved understanding of the role and importance of knowledge and skills embedded in individuals, and values and norms in relation to successful new product development and in relation to making identified NPD success factors more operational and thus more implementable. The literature on NPD success factors displays limited knowledge of the role and importance of these two aspects. We believe that discussions of NPD can gain much from increased understanding of or at least some reference to those involved in the development of new products. In this respect answers to questions like: What kind of dynamics are at play? How do people involved act? What knowledge is needed and how is it applied? How do organisational aspects like routines, procedures and culture influence actions and knowledge in product development? hold great potential for advancing our understanding of what constitutes successful new product development.

References Anderson, J.R. (1983), The Architecture of Cognition, Harvard University Press, Cambridge, MA. Barclay, I. (1992), ``The new product development process: past evidence and future practical application, part 1'', R&D Management, Vol. 22 No. 3, pp. 255-63. Barney, J. (1991), ``Firm resources and sustained competitive advantage'', Journal of Management, Vol. 17, pp. 99-120. Biemans, W.G. and Harmsen, H. (1995), ``Overcoming the barriers to market-oriented product development'', Journal of Marketing Practice, Vol. 1 No. 2, pp. 7-26. Brown, S. and Eisenhardt, K.M. (1995), ``Product development: past research, present findings, and future directions'', Academy of Management Review, Vol. 20 No. 2, pp. 343-78. Cooper, R.G. (1990), ``New products: what distinguishes the winners?'', Research and Technology Management, November-December, pp. 27-31. Cooper, R.G. (1998), ``Benchmarking new product performance: results of the best practices study'', European Management Journal, Vol. 16, pp. 1-17. Cooper, R.G. (1999), ``From experience. The invisible success factors in product innovation'', Journal of Product Innovation Management, Vol. 16, pp. 115-33. Cooper, R.G. and Kleinschmidt, E.J. (1986), ``An investigation into the new product process'', Journal of Product Innovation Management, Vol. 3, pp. 71-85. Cooper, R.G. and Kleinschmidt, E.J. (1987), ``New products: what separates the winners from the losers'', Journal of Product Innovation Management, Vol. 4 No. 3, pp. 169-84. Cooper, R.G. and Kleinschmidt, E.J. (1991), ``New product processes at leading industrial firms'', Industrial Marketing Management, Vol. 20, pp. 137-47. Cooper, R.G. and Kleinschmidt, E.J. (1995), ``Benchmarking the firm's critical success factors in new product development'', Journal of Product Innovation Management, Vol. 12, pp. 374-91. Craig, A. and Hart, S. (1992), ``Where to now in new product development research?'', European Journal of Marketing, Vol. 26, pp. 1-46. Day, G.S. (1994), ``The capabilities of market-driven organizations'', Journal of Marketing, Vol. 58, October, pp. 37-52. Durand, T. and CreÂmadez, M. (1999), ``When efficient operations means weak product development'', Proceedings of the 6th International Product Development Management Conference, Cambridge, 5-6 July, pp. 349-68. Eriksen, B. and Foss, N.J. (1997), Dynamisk kompetenceudvikling. En ny ledelsesstrategi, Handelshùjskolens Forlag, Kùbenhavn. Foss, K. and Harmsen, H. (1996), ``Studies of key factors of product-development success: a resource-based critique and reinterpretation'', in Foss, N.J. and Knudsen, C. (Eds), Towards a Competence Theory of the Firm, Routledge, London, pp. 133-50. Grunert, K.G. and Sùrensen, E. (1996), Perceived and Actual Key Success Factors: A Study of the Yogurt

48

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Market in Denmark, Germany and the United Kingdom, MAPP Working Paper No. 40, The Aarhus School of Business. Gupta, A.K. and Wilemon, D. (1988), ``The credibilitycooperation connection at the R&D-marketing interface'', Journal of Product Innovation Management, Vol. 5, pp. 20-31. Hamel, G. and Heene, A. (1994), Competence-based Competition, John Wiley & Sons, Chichester. Harmsen, H. (1992), Determinanter for produktinnovationssucces, MAPP Working Paper No. 5. The Aarhus School of Business. Harmsen, H. (1994), Tendencies in Product Development in Danish Food Companies. Report of a Qualitative Analysis, MAPP Working Paper No. 17, The Aarhus School of Business. Harmsen, H. (1996), ``Succesfaktorer i produktudvikling og deres implementering i mellemstore fùdevarevirksomheder'', PhD dissertation, MAPP Centre/Department of Marketing, The Aarhus School of Business (in Danish). Huff, A.S. (1990), ``Mapping strategic thought'', in Huff, A.S. (Ed.), Mapping Strategic Thought, John Wiley & Sons, Chichester, pp. 11-49. Leonard-Barton, D. (1992), ``Core capabilities and core rigidities: a paradox in managing new product development'', Strategic Management Journal, Vol. 13, pp. 111-25. Lilien, G. and Yoon, E. (1989), ``Determinants of new product performance: a strategic reexamination of the empirical literatures'', IEEE Transactions on Engineering Management, Vol. 36 No. 1, pp. 3-11. McDonough, E.F. III (1986), ``Matching management control systems to product strategies'', R&D Management, Vol. 16, pp. 141-9. Montaya-Weiss, M. and Calantone, R. (1994), ``Determinants of new product performance'', Journal of Product Innovation Management, Vol. 11, pp. 397-417. O'Connor, P. (1994), ``From experience. Implementing a stage-gate process: a multi-company perspective'', Journal of Product Innovation Management, Vol. 11, pp. 183-200. Page, A. (1993), ``Assessing new product development practices and performance: establishing crucial norms'', Journal of Product Innovation Management, Vol. 10, pp. 273-90. Penrose, E. (1959), The Theory of the Growth of the Firm, Basil Blackwell, London. Plichta, K. and Harmsen, H. (1994), Studies of Key Success Factors of Product Development Success: A Resource-based Critique and Reinterpretation, working paper no. 94-14, Copenhagen Business School, Institute of Industrial Research and Social Development. Prahalad, C.K. and Hamel, C. (1990), ``The core competence of the corporation'', Harvard Business Review, May-June, pp. 79-91. Rothwell, R. (1977), ``The characteristics of successful innovators and technically progressive firms'', R&D Management, Vol. 7 No. 3, pp. 191-206. Sanchez, R., Heene, A. and Thomas, H. (1996), ``Introduction: towards the theory and practice of competence-based competition'', in Sanchez, R., Heene, A. and Thomas, H. (Eds), Dynamics of

Competence-Based Competition, Pergamon, Oxford. Schewe, G. (1991), Key Success Factors of Successful Innovation Management, working paper No. 274, Institute for Business Administration, University of Kiel. Teece, D.J., Pisano, G. and Shuen, A. (1990), Firm Capabilities, Resources and the Concept of Strategy, working paper No. 90-9, University of California at Berkeley, Berkeley, CA. Voss, C.A. (1985), ``Determinants of success in the development of application software'', Journal of Product Innovation Management, Vol. 2, pp. 122-9. Wernerfelt, B. (1984), ``A resource-based view of the firm'', Strategic Management Journal, Vol. 5 No. 2, pp. 171-80. Wernerfelt, B. (1995), ``The resource-based view of the firm: ten years after'', Strategic Management Journal, Vol. 16, pp. 171-4.

Appendix The Appendix shows the product development competence of companies C and D. Company C The map of company C's product development competence is shown in Figures A1 and A2. The point of departure is not product development competence as such, the reason being that the main raw material that this company uses to manufacture its product makes up a substantial part of the physical product and thus limits potential development possibilities. This has forced the company to focus on development of concepts in its ``product development''. This means that the company does not change the physical product very much but concentrates its development efforts on ``everything that is around the physical product''. It also means that ``product development'' refers formally to the marketing function and thus the respondent is the marketing director. Company D The map of company D's product development competence is shown in Figures A3 - A6. The product development function refers formally to the production department. The respondent is in this case the product development manager (with a technical background). 49

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Figure A1 Company C's product development competence (continues in Figure A2)

Figure A2 Company C's product development competence (continued from Figure A1)

50

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Figure A3 Company C's product development competence (continues in Figure A4)

Figure A4 Company C's product development competence (continues in Figure A5)

51

Implementation of success factors in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 37±52

Bjarne Jensen and Hanne Harmsen

Figure A5 Company C's product development competence (continues in Figure A6)

Figure A6 Company C's product development competence

52

Introduction

Resource adequacy in new product development: a discriminant analysis

New product development (NPD) is widely regarded as an important business activity in creating and maintaining an organisation's competitiveness. Its importance can be seen in the number of studies undertaken over the past three decades to identify the factors that are associated with successful new product performance. These studies have investigated new product performance from different perspectives, including new product success (Globe et al., 1973; Langrish et al., 1972), failure (Cooper, 1975; Crawford, 1979), or a comparison between success and failure (Cooper, 1979a; Link, 1987; Maidique and Zirger, 1984; SPRU, 1972). As a result, a large number of factors have been identified, including the NPD process itself, management systems and organisational characteristics, strategy and people (Craig and Hart, 1992). In particular, an organisation's marketing and technical resources and skills seem to impact on the effectiveness of its NPD processes (Cooper and Kleinschmidt, 1988; Song and Parry, 1997) and on its new product performance (Cooper, 1979b; Cooper and Kleinschmidt, 1987; Song and Parry, 1994; Song and Parry, 1997). The NPD process is a series of activities that transforms new product ideas into marketable products. It is an uncertain process that requires marketing and technical skills, as well as financial resources. In a recent study, Cooper and Kleinschmidt (1995) found that resource commitment is one of three cornerstones of successful product development, which is an important finding from a small to medium enterprise (SME) viewpoint. While SMEs have long been recognised as innovative organisations, their lack of skills and resources has been suggested as a potential weaknesses when innovating (Nooteboom, 1994). Their lack of resources, together with an increasingly competitive environment, increased industrial specialisation (Teece, 1992), and the nature of NPD, has made the role of inter-firm cooperation in NPD crucial for many SMEs. An SME can cooperate with others to make use of their complementary contributions (Day, 1990), particularly in terms of providing additional resources and capabilities (Bleeke and Earst, 1991). One of

Xueli Huang Geoffrey N. Soutar and Alan Brown The authors Xueli Huang is Research Fellow, School of Management, Edith Cowan University, Churchlands, Australia. Geoffrey N. Soutar is Director, Graduate School of Management, The University of Western Australia, Nedlands, Australia. Alan Brown is Acting Executive Dean, Faculty of Business, Edith Cowan University, Churchlands, Australia. Keywords New product development, Resources, Skills, Small to medium-sized enterprises Abstract New product development (NPD) is crucial to the survival and thriving of a business entity and a firm's sources of advantages are important to the NPD success. This paper explores the marketing and technical resources adequacy of Australian small and medium enterprises (SMEs) in NPD. A survey of 276 Australian SMEs in the chemical and machinery industries was conducted. Analytical procedures include factor analysis, cluster analysis, and discriminant analysis. Findings from these analyses suggest that three distinct groups in terms of their NPD resources exist in Australian SMEs: one group with rich marketing and technical resources and skills, one with rich technical resource only, and one with rich technical skill only. The organisational and managerial characteristics of each group of these firms are described. The findings imply that different resource groups need to adopt different strategies in NPD. Electronic access The research register for this journal is available at http://www.mcbup.com/research_registers The current issue and full text archive of this journal is available at http://www.emerald-library.com/ft European Journal of Innovation Management Volume 4 . Number 1 . 2001 . pp. 53±59 # MCB University Press . ISSN 1460-1060

53

Resource adequacy in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 53±59

Xueli Huang, Geoffrey N. Soutar and Alan Brown

the barriers to inter-firm cooperation in innovation, however, is a lack of information about the characteristics of innovative small firms. There has been little attempt to profile the resource and skills of innovative SMEs and their key decision-makers. The present paper attempts to fill this gap.

Table I Descriptive statistics for the level of resources and skills Adequacy of resources and skills in NPD R&D skills Engineering skills Manufacturing skills R&D resources Engineering resources Manufacturing resources Marketing research skills Salesforce skills Distribution skills Advertising/promotion skills Marketing research resources Salesforce resources Distribution resources Advertising and promotion resources Financial resources

The present study Information about the characteristics of innovative firms and their key decisionmakers in product innovation, as well as the level of their resources and skills was collected as part of a larger study into innovation in Australian SMEs involved in the chemical and machinery industries. These two industries were selected as they are the most active in product innovation (Australian Bureau of Statistics, 1998) and account for a majority of Australia's manufacturing R&D (61 per cent in 1994-95) (Industry Commission and Department of Industry Science and Technology, 1997).

Mean

Standard deviation

3.88 3.96 3.93 3.31 3.51 3.67 3.23 3.27 3.33 3.10 2.97 3.13 3.26 3.00 3.16

0.92 0.88 0.80 1.05 0.92 0.87 1.05 1.06 1.04 1.03 1.03 1.06 1.05 1.06 1.15

population of interest. A search of the database found a little over 10,000 such firms and a sample of 3,440 was selected by choosing every third such firm in the database. The questionnaires were mailed to the selected firms, resulting in a return of 440 useful questionnaires, with 276 from innovative firms and 164 from non-innovative firms. A firm was considered as innovative if it has developed a new product in the last five years. In addition, 375 questionnaires were returned because the addressed person had left the business or because of an incorrect address. The effective response rate from the innovative firms was 27 per cent based on the percentage of innovative firms in both industries (Australian Bureau of Statistics, 1998), which is a normal response rate for most surveys of SMEs (e.g. Sullivan and Kang, 1999). The analyses undertaken were based on the 276 innovative respondents.

The questionnaire used A questionnaire was developed based on previous research that had examined NPD. Respondents were initially asked if they had developed a new product in the last five years and, if they had, they were asked to reply to a number of questions in relation to their most recent new product. Such respondents were asked a series of questions about their organization and its NPD processes. Some personal background data were also collected. Within the survey, 14 questions, developed by Song and Parry (1997), asked about their organisation's marketing and technical resources and skills. An additional question about financial resources was included and these 15 items were of central interest to the present study. Respondents were asked about the adequacy of the 15 skills or resources shown in Table I for the last NPD process they had undertaken using a five-point scale ranging from strongly agree (1) to strongly disagree (5).

Data analysis Initially means and standard deviations were examined to gain a feel for the data obtained. Subsequently a factor analysis was undertaken on the skills and resources section of the questionnaire to find if there were any underlying dimensions that should be used in the subsequent analysis. As will be noted subsequently, several such dimensions were found, which suggested that skill and resource availability is multidimensional. Consequently, any understanding of the phenomenon requires an examination of patterns of adequacy. A procedure that has

The sample Respondents were drawn from a database purchased from Dun & Bradstreet. Firms that employed fewer than 200 people and, as already noted, worked in the chemical or machinery industries were chosen as the 54

Resource adequacy in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 53±59

Xueli Huang, Geoffrey N. Soutar and Alan Brown

been found useful in this situation is cluster analysis, which places objects (here, respondents) into groups based on the similarity over some criteria (here, the adequacy of skills and resources) (Savery and Soutar, 1990; Soutar et al., 1995; Soutar and Williams, 1985). Once such groups have been found, discriminant analysis can be used to examine intergroup differences (Klecka, 1980). In particular, analyses were undertaken to examine the association of the adequacy of resources and skills with the organisational and managerial characteristics in SMEs. The results of these analyses are outlined in the next section.

seems to provide reasonable results with data of this sort. Using the commonly accepted ``eigen values greater than one'' rule to determine the number of factors to retain, three factors emerged that, together, explained 68 per cent of the variance in the data. The rotated factor loadings, communalities and percent of variance explained by each factor are shown in Table II. Table II shows that the communalities ranged from 0.488 to 0.766, suggesting that the three factors explained an acceptable amount of the variance in each of the items, while the percentage of variance explained (68 per cent) suggests that there is little more information to be explained. The alpha coefficients from the summated scales suggested by the three factors ranged from 0.94 to 0.72, suggesting that all factors were sufficiently reliable for further analysis. Consequently, the three-factor solution was accepted. As can be seen from the rotated factor loadings, the first factor was related to marketing skills and resources, the second to internal (technical and financial) resources and the third to technical skills. In order to understand respondents' views about these three factors, summated scales were computed. The descriptive statistics are shown in Table III. The mean scores support earlier comments, respondents were more positive about their technical skills than their technical resources and more positive about their technical resources than their marketing skills and resources, which was confirmed by computed paired sample t statistics for the three pairs of factors. All were significant well beyond the 1 per cent level, suggesting that the differences are real and that SMEs have quite different views about the three underlying dimensions. As can be seen from the Table, three skill and resource adequacy dimensions emerged from the factor analysis and, as this suggested a multidimensional view of resource adequacy, the three scores computed for each factor were used as a basis for clustering respondents into groups with similar satisfaction patterns using Howard and Harriss's (1966) clustering procedure.

The results obtained Level of the resources and skills in SMEs As already noted, the data of central interest to the present study were those related to resource availability and the descriptive statistics for these items are shown in Table I. As can be seen from the Table, standard deviations for all items were around the one value, suggesting a reasonable spread of opinions about skills and resources adequacy across the sample. Mean scores were all less than four on the five-point scale used, suggesting that respondents were not overly optimistic about the adequacy of their NPD skills and resources. Interestingly, respondents were more positive about internal factors (e.g. engineering and R&D skills) than about external factors (e.g. marketing and advertising and promotion resources), which may reflect the technical background of many SME managers in the industries surveyed. While these results provide useful information, patterns of responses are of more interest when trying to understand the nature of skills and resources issues in SMEs NPD processes and the next stage of data analysis attempted to provide such understanding. Factors of resources and skills in SMEs It seemed likely that some skills and/or resources are related to each other. Therefore, before examining the relationship between skills and resources adequacy and new product success, the 15 shown in Table I were factored analysed. A principal components analysis with a varimax rotation of the retained factors was used as this procedure

Groups of SMEs with different levels of resources and skills The resultant factor scores were used as inputs for cluster analysis. A three-cluster 55

Resource adequacy in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 53±59

Xueli Huang, Geoffrey N. Soutar and Alan Brown

Table II Factor analysis results ± 15 skills and resources Skills (resources) more than adequate

Factor 1

Salesforce resources Advertising/promotion resources Distribution resources Advertising/promotion skills Salesforce skills Marketing research resources Marketing research skills Distribution skills Manufacturing resources Engineering resources R&D resources Financial resources Manufacturing skills R&D skills Engineering skills Note:

a

0.863 0.851 0.839 0.828 0.802 0.800 0.800 0.795 0.854 0.754 0.634 0.604 0.530

Marketing skills and resources Internal resources Technical skills

0.446 0.851 0.794

Communality 0.770 0.705 0.488 0.526 0.682 0.747 0.659 0.662 0.649 0.699 0.668 0.772 0.751 0.766 0.613

Factor loadings less than 0.40 were excluded to improve readability

Table III Skills and resource adequacy factors Factor

Factor loadingsa Factor 2 Factor 3

Mean score

Standard deviation

Alpha coefficient

3.16 3.40 3.92

0.87 0.75 0.70

0.94 0.74 0.72

background. A stepwise approach was used because of the exploratory nature of this study. Following the procedure recommended by Hair et al. (1995), the statistical significance for two discriminant functions was checked first, both functions were significant at 0.05 level. The hit ratio was 50.2 per cent, which was bigger than 1.25 times chance. The Press's Q was also calculated, valued at 35.3, also significant at 0.05 level. All these indicated that the multiple discriminant analysis was statistically significant for the data. The rotated structural loadings and group centroids of the multiple discriminant analysis are presented in Table V, following the suggestions of Soutar and Clark (1981). The results are also graphically displayed in Figure 1. The structural loadings of variables were depicted with vectors and the group centroids were also plotted on Figure 1. This graphical display of structural loadings and group centroids together highlights the characteristics of the three groups. It can be seen from Figure 1 that the first discriminant function is the primary source of difference between groups 2 and 3 versus group 1, implying that it mainly represents skills. The second discriminant function separates groups 1 and 3 from group 2, reflecting the resource difference between groups. The group profile can also be delineated from Figure 1. Group 3 (43 per cent of the

solution was considered as most appropriate for the data as a sudden large jump was observed in going from three to two clusters (Hair et al., 1995). The details of these three groups are reproduced in Table IV (for details, see Huang et al., 1999). These three clusters represent technical-resource rich (group 1), technical-skill-rich (group 2) and all-resource-and-skill-rich (group 3).

Profile of SMEs with different levels of resources and skills A multiple discriminant analysis was conducted to profile the members of each of the three groups. The three-cluster membership was used as the dependent variable and the characteristics of the firms and their key decision-makers as independent variables. The firm's characteristics included turnover, number of employees, major business activities, age of firm, industry and geographic location (State). The key decisionmaker's characteristics covered educational level, years of experience, age and 56

Resource adequacy in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 53±59

Xueli Huang, Geoffrey N. Soutar and Alan Brown

Table IV The adequacy of resources and skills of three clusters of firms Resources and skills R&D skills Engineering skills Manufacturing skills R&D resources Engineering resources Manufacturing resources Marketing research skills Salesforce skills Distribution skills Advertising/promotion skills Marketing research resources Salesforce resources Distribution resources Advertising/promotion resources Financial resources

Overall

Cluster 1 (N = 50)

Cluster 2 (N = 100)

Cluster 3 (N = 115)

ANOVA (Scheffe)

3.88 3.94 3.93 3.31 3.51 3.67 3.23 3.27 3.33 3.10 2.97 3.13 3.26 3.00 3.16

3.06 3.42 3.82 3.06 3.38 3.78 2.26 2.24 2.48 2.14 2.12 2.10 2.42 2.02 3.18

4.03 3.94 3.56 2.82 2.91 2.94 3.25 3.21 3.16 3.05 2.84 2.94 3.00 2.82 2.44

4.10 4.21 4.30 3.82 4.06 4.23 3.63 3.77 3.84 3.57 3.46 3.74 3.86 3.61 3.77

3, 2 > 1 3, 2 > 1 3 > 1, 2 3 > 1, 2 3>1>2 3>1>2 3>2>1 3>2>1 3>2>1 3>2>1 3>2>1 3>2>1 3>2>1 3>2>1 3>1>2

Note: Mean score on a five-point Likert scale with 5 strongly agree and 1 strongly disagree Table V The results of multiple discriminant analysis Structural correlation with discriminant function Function 1 Function 2

Variable Engineering background Turnover for 1997-1998 Number of employees Manufacturing-dominated firms Non-manufacturing-dominated firms Years of experience in this industry Western Australia Tradeperson background Group 1 2 3

Figure 1 The plot of structural loadings and group centroids

±0.427 ±0.425 0.380 0.127 ±0.147 0.387 0.058 ±0.177

0.076 ±0.024 0.203 ±0.602 0.600 0.554 ±0.521 0.366

Group centroid Function 1 Function 2 ±0.514 0.138 0.173 ±0.459 0.234 0.227

manufacturing, such as retailing, consulting and wholesaling. Group 1 (19 per cent of the sample) has the smallest number of members among these three groups. They usually have a large sales volume and possess a moderately high level of resources, especially in manufacturing, engineering, and finance. However, they also indicated that they severely lacked resources and skills in marketing. The key decisionmaker in these firms usually has an engineering or tradesperson background and the major business activity of these firms is manufacturing. It is interesting to find that the type of industry does not contribute significantly to

Note: Only those variables with at least one loading greater than 0.3 were listed

sample) members that have rich marketing and technical resources and skills consist of firms with a large number of employees and an experienced key decision-maker in product innovation. These firms undertake manufacturing as their major business activities. Members of group 2 (38 per cent of the sample) have rich technical skills, particularly in R&D and engineering. They also have reasonably adequate marketing and skills and resources. This group includes firms with the major business activity being non57

Resource adequacy in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 53±59

Xueli Huang, Geoffrey N. Soutar and Alan Brown

distinguishing the groups, suggesting indifference in resources and skills of these firms from both industries. A firm's age is also not a good indicator to predict its resources and skills in NPD. Does a firm's size matter in NPD in terms of resources and skill level? The previous analyses show that this depends on how we define it. The number of employees can be used as a strong indicator of resource and skill adequacy in NPD; while a firm's sales volume only partially reflects this adequacy.

engineering and production. The manufactured product could then be sent back the firm in group 1, which have sufficient resources and skills in marketing the new product. For venture capitalists, firms in groups 2 and 3 may be more the target for investing. While firms in group 3 may occasionally need capital for their product innovations, they usually perform well in the NPD process. Investing in these firms could be more secure than firms in other groups. Group 2 lacks resources, both in production and finance, having the most potential focus for external funding in NPD. Managers of firms in group 1 need to rethink the role of product innovation in their survival and growth and the structural impediments of resources and skills. These firms could be jobbers or parts suppliers in a manufacturing chain. In these cases, these firms mainly take orders from others and have few incentives to develop new products (Jones et al., 1989). However, their survival and growth may be under threat because they are usually heavily dependent on other organizations. If the markets for group 2 are competitive and product innovation become crucial for a firm's survival, organizations in group 2 could recruit experienced managers in overcoming its shortage in skills, and then gradually develop its marketing resources and skills.

Conclusions and implications The characteristics of a firm and its key decision-maker can be used to predict a firm's resource and skill level in NPD, and so be able to predict a firm's proficiency in the NPD process and performance. Diversity is the most important and striking characteristics of SMEs (Nooteboom, 1994). Our cluster and discriminant analyses have provided new credence to this. SMEs behave differently in the NPD process and have a wide spectrum of resources and skills in NPD. It can be concluded from our discriminant analysis that several variables representing the organizational and managerial characteristics are useful in distinguishing firms' resources and skills in the NPD Process. The organizational variables include number of employees, turnover, and major business activity. The managerial characteristics consist of years of experience and background of the key decision-maker in the firm. Overall, the manager's experience and number of employees are the strongest indicators for predicting a firm's resource and skill sufficiency in NPD. The findings here also offer several managerial implications. The varied level of resources and skills a firm has in NPD suggests that different strategic focuses need to be adopted by different groups of SMEs. For firms in group 3, they have sufficient resources and skills for NPD; their focuses in successful NPD could be placed on the formulation of appropriate new product strategies and the improvement of the NPD process. Groups 1 and 2 have complementary resources and skills in NPD. Therefore, cooperative NPD may be a viable strategy. For example, a new product idea from group 2 could be passed to firms of group 1, which have more resources in

References Australian Bureau of Statistics (1998), Innovation in Manufacturing: 1996-97, Australian Government Publishing Service, Canberra. Bleeke, J. and Earst, D. (1991), ``The way to win cross-border alliances'', Harvard Business Review, Vol. 69, November-December, pp. 127-35. Cooper, R.G. (1975), ``Why new industrial products fail'', Industrial Marketing Management, Vol. 4, pp. 315-26. Cooper, R.G. (1979a), ``Identifying industrial new product success: project newprod'', Industrial Marketing Management, Vol. 8 No. 2, pp. 124-35. Cooper, R.G. (1979b), ``The dimensions of industrial new product success and failure'', Journal of Marketing, Vol. 43 No. 3, pp. 93-103. Cooper, R.G. and Kleinschmidt, E.J. (1987), ``New products: what separates winners from losers'', Journal of Product Innovation Management, Vol. 4 No. 3, pp. 169-84. Cooper, R.G. and Kleinschmidt, E.J. (1988), ``Resource allocation in the new product process'', Industrial Marketing Management, Vol. 17, pp. 249-62.

58

Xueli Huang, Geoffrey N. Soutar and Alan Brown

Resource adequacy in new product development

European Journal of Innovation Management Volume 4 . Number 1 . 2001 . 53±59

Cooper, R.G. and Kleinschmidt, E.J. (1995), ``Benchmarking firms' new product performance and practice'', Engineering Management Review, Vol. 23, Fall, pp. 112-20. Craig, A. and Hart, S. (1992), ``Where to now in new product development research?'', European Journal of Marketing, Vol. 26 No. 11, pp. 1-49. Crawford, C.M. (1979), ``New product failure rates ± fact and fallacies'', Research Management, September, pp. 9-13. Day, G.S. (1990), Market Driven Strategy, The Free Press, New York, NY. Globe, S., Levy, W. and Schwartz, C.M. (1973), ``Key factors and events in the innovation process'', Research Management, Vol. 16, July, pp. 8-15. Hair, J.F. Jr, Anderson, R.E., Tatham, R.L. and Black, W.C. (1995), Multivariate Data Analysis with Readings, 4th ed., Prentice-Hall, Englewood Cliffs, NJ. Howard, H. and Harris, R.D. (1966), ``A Hierarchical Grouping Routine; IBM360/65 FORTRAN IV program'', University of Pennsylvania, Philadelphia, PA. Huang, X., Soutar, G. and Brown, A. (1999), New Product Development in Australian Small and Medium Enterprises, Working Paper Series, Small and Medium Enterprise Research Centre, Edith Cowan University. Industry Commission and Department of Industry Science and Technology (1997), A Portrait of Australian Business: Results of the 1995 Business Longitudinal Survey, Small Business Research Program, Canberra. Jones, B.L., Dunlop, W.C. and Sheehan, W.J. (1989), Small Business in Australia: Problem and Prospects, Allen & Unwin, Syndey. Klecka, W.R. (1980), Discriminant Analysis, Sage Publications, Newbury Park, CA. Langrish, J., Gibbons, M., Evans, W.G. and Jevons, F.R. (1972), Wealth from Knowledge, Macmillan, London. Link, P.L. (1987), ``Keys to new product success and failure'', Industrial Marketing Management, Vol. 16, pp. 109-18.

Maidique, M.A. and Zirger, B.J. (1984), ``A study of success and failure in product innovation: the case of the US electronics industry'', IEEE Transactions on Engineering Management, No. 4, pp. 192-203. Nooteboom, B. (1994), ``Innovation and diffusion in small firms: theory and evidence'', Small Business Economics, Vol. 6, pp. 327-34. Savery, L.K. and Soutar, G.N. (1990), ``Unions and productivity: some community views'', Labour Economics and Productivity, pp. 120-31. Song, X.M. and Parry, M.E. (1994), ``The dimension of industrial new product success and failure in state enterprises in the People's Republic of China'', Journal of Product Innovation Management, Vol. 11, pp. 105-18. Song, X.M. and Parry, M.E. (1997), ``A cross-national comparative study of new product development processes: Japan and the United States'', Journal of Marketing, Vol. 61, April, pp. 1-18. Soutar, G.N. and Clark, Y. (1981), ``Life style and television viewing behaviour in Perth, Western Australia'', Australian Journal of Management, Vol. 6 No. 1, pp. 109-23. Soutar, G.N. and Williams, T.A. (1985), ``Prison officers' work attitudes: the influence of background and work experience'', Australian New Zealand Journal of Criminology, March, pp. 17-24. Soutar, G.N., McNeil, M.M. and Maisey, G. (1995), ``Community satisfaction with policing: the Western Australian experience'', The Police Journal, Vol. 68 No. 4, pp. 351-7. SPRU (1972), Success and Failure in Industrial Innovation, Centre for the Study of Industrial Innovation, London. Sullivan, P. and Kang, J. (1999), ``Quick response adoption in the apparel manufacturing industry: competitive advantage of innovation'', Journal of Small Business Management, Vol. 37 No. 1, pp. 1-13. Teece, D.J. (1992), ``Competition, cooperation, and innovation: organizational arrangements for regimes of rapid technological progress'', Journal of Economic Behavior and Organization, Vol. 18, pp. 1-25.

59