Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Wiley is collaborating with JSTOR to digitize, preserve and extend access to Strategic Management Journal.
http://www.jstor.org
PETER H. FRIESEN
McGillUniversity,Montreal,Canada
Summary
Two very difJerent nmodelsoJ pr-oductinnovation ar-epostulated and
tested. The con rvative m-odelassumes that ininov-ationis peijormed
reluictanitly-,mainl/ in r-esponiseto serious challenges. It therefore
pr-edic-ts that innoration will co rrelate positively with environmental,
injformation processing, structlural and decision nmaking variables that
INTRODUCTION
There is much controversy in the literature on organizational innovation. According to Downs and
Mohr (1976:700):
Perhaps the most alarming characteristic of the body of empirical study of innovation is
the extreme variance among its findings, what we call instability. Factors found to be
important for innovation in one study are found to be considerably less important, not
important at all, or even inversely important in another study. This phenomenon occurs
with relentless regularity ... In spite of the large amount of energy expended, the results
have not been cumulative.
Of the 38 propositions bearing directly on the act of innovation cited by Rogers and Shoemaker
(1971:350-376), 34 were supported in some studies and found to receive no support in others. The
four consistently predictive propositions were treated in very few studies.
We believe that in studies of product innovation, many of the conflicts in the literature have been
caused by the failure of researchers to take into account the nature of the innovative strategy of the
firm, something that is often determined by executives on the basis of their goals and
temperaments. Some executives decide that regular and extensive innovation in product lines or
services and product designs should be a vital element of strategy. Their 'entrepreneurial' firms may
0143-2095/82/01000
1-25$02.50
c 1982by JohnWiley& Sons, Ltd.
try to obtain a competitive advantage by routinely making dramatic innovations and taking the
concomitant risks. Other organizations are run by more conservative managers who may view
innovation as costly and disruptive to production efficiency. Such 'conservative' firms will innovate
only when they are very seriously challenged by competitors or by shifting customer wants The
thesis of this paper is that the impact upon product innovation of environmental, information
processing, structural and decision making variables will vary significantly and systematically
among entrepreneurial and conservative firms; that future research on the determinants of
innovation must consider organizational strategy.
Based on previous empirical research, the paper develops distinct arguments concerning the
determinants of innovation in conservative and entrepreneurial firms. It then presents data from a
diverse sample of 52 Canadian business firms which show how different are the correlates of
product innovation for both kinds of firms. The scope of the paper is limited to innovations in
product lines, produict designs, and services offered. It does not extend to technological or
administrative innovations. Only business firms are studied. Our arguments or findings may not
hold for other types of organizations. Finally, although we sometimes talk about the
'determinants' of innovation, the direction of causality is always in doubt, and, strictly speaking,
we should refer only to 'correlates' of innovation.
MOMENTUM
AND INNOVATION
Miller and Friesen (1980) have shown that momentum is a pervasive force in organizations; that
past practices, trends and strategies tend to keep evolving in the same direction, perhaps eventually
reaching dysfunctional extremes. For example, the implementation of bureaucratic controls may
be followed by more of the same until the organization begins to take on a very mechanistic mantle.
Centralization of authority often continues until the organization becomes an autocracy, while
decentralization can lead to the proliferation of uncoordinated departmental fiefdoms. Miller and
Friesen (1980) found that the same might be true of innovation. Firms with a propensity to
innovate become still more innovative, sometimes passing the point of dramatically diminished
returns. Conservative firms on the other hand sometimes drift towards complete stagnation.
It is reasonable to believe that momentum and its resultant excesses will be attenuated by
influences that warn of the dangers of conservative and entrepreneurial extremes. There may be a
number of such mitigating influences. First, they may take the form of information processing
devices such as scanning and control systems that bring information about the environment and
operating efficiency to decision makers. Second, they may include structural integration devices
that inform decision makers of the consequences of innovation. Finally, they may comprise
decision making methods such as the amount of analysis, planning and consideration of overall
strategy, which describe how carefully information is being processed. Dangerous momentuminduced extremes of too much or too little innovation are expected to be reduced by these warning
devices.
If this is true, the determinants of innovation must vary as a function of the type or direction of
innovation invoked by a firm's innovation strategy. In samples of innovative firms where the
danger is reaching too high a level of innovation mitigating factors such as the use of information
processing and decision making devices will correlate negatively with innovation. That is, firms
with good warning systems will be less innovative than firms that lack such systems. In contrast, for
samples of firms pursuing conservative low innovation' strategies, the operative danger is most
likely to be strategic stagnation. Here the mitigating factors will correlate positively with
innovation as effective warning systems serve as an incentive for innovation. In other words, the
existence of momentum implies that we need at least two models to predict innovation, the first
applicable in conservative firms, the second, in bold entrepreneurial firms. We shall present and test
these models in an attempt to resolve some of the conflicts in the literature on innovation.
THE CONSERVATIVE
MODEL OF INNOVATION
The literature on product innovation, although fraught with conflict, seems to point
preponderantly to a conservative model of innovation. Basically, the model implies that
innovation is not a natural state of affairs, that it must be encouraged by challenges and threats,
and that it requires a particular type of structure and an effective information processing system to
make conservative managers aware of the need for change. We contend that the conservative
model will apply to firms that perform very little innovation or risk taking. These are roughly
reminiscent of the reactors of Miles and Snow (1978), the stagnating firms of Miller and Friesen
(1978), and the adapters of Mintzberg (1973). Here, innovation is performed infrequently, and,
perhaps, because of its disruptive nature, reluctantly. The conservative model suggests that
innovation will only take place when there are felt pressures. It postulates four types of
prerequisites, or at least strong facilitators, of innovation. First, there must be environmental
challenges before innovation occurs. For example, because they create a need for innovation,
factors such as environmental dynamism and hostility would be expected to correlate positively
with innovation. Second, there must be information about these challenges brought to key decision
makers by effective scanning and control systems. Third, there must be an ability to innovate, that
is created by adequate resources, skilled technocrats, and structural devices. Arnd finally, there
must be decision making methods appropriate for innovation projects. For example, the extent to
which key decision makers analyse innovation-related information and use it for planning and
strategy development is expected to correlate positively with innovation. As we shall see, many of
the findings in the literature seem to support our conservative model. What follows are some
specific predictions implied by this model.
Environmental variables
Myers and Marquis (1969) found that 53 per cent of the product and technological innovations in
their sample came in response to market, competitive, or other external environmental influences.
The more diynamic and hostile (i.e. competitive) the environment, the greater the need for
innovation and the more likely it is that firms will be innovative. When competitors' products
change rapidly or when customer needs fluctuate, the conservative model hypothesizes that
innovation will be common. In stable environments this is less likely to be true (Burns and Stalker,
1961). Another environmental dimension may also be germane: namely, that of heterogeneity.
Firms operating in many different markets are likely to learn from their broad experience with
competitors and customers. They will tend to borrow ideas from one market and apply them in
another. According'to Wilson (1966), the greater the diversity of the organization, the greater the
probability that innovations will be proposed, and the greater the probability that organization
members will conceive major innovations. Of course diversity in organization personnel, operating
procedures, technologies and administrative practices increases with environmental heterogeneity
(Peters, 1969).
Information processing variables
Burns and Stalker (1961) have argued that mechanistic structures impede innovation while organic
structures facilitate it, in part because the former have much less information processing capacity.
Subsequent literature has shown that there are at least two types of information processing
categories that can influence innovation. Aguilar (1967), Baker, Siegman and Rubenstein (1967),
Utterback (1971). Keller and Holland (1975) and Tushman (1977) have called attention to the role
of scanning the environment, claiming that a primary limitation on a firm's innovativeness is its
ability to recognize the needs and demands of its external environment. Perceived market needs
accounted for 75 per cent of the ideas for innovation (Baker et al., 1967). This is confirmed by the
work of Carter and Williams (1957), Myers and Marquis (1969) and Mueller (1962), and the survey
of Rogers and Shoemaker (1971:372-373). Controls also are said to facilitate innovation. Controls
that monitor task performance and financial results are said to identify areas of weakness and to
prompt remedially oriented innovations (Rosner, 1968; Downs, 1966: 191).
Structural variables
One important structural variable that has been demonstrated to be associated with innovation is
(or concentration) of authority for decision making. According to Thompson
centralilation
(1969: 25):
... dispersal of power is important because concentrated power often prevents
imaginative solutions of problems. When power meets power, problem solving is
necessarily called into play ... Dispersed power, paradoxically, can make resources more
readily available to support innovative projects, because it makes possible a larger
number and variety of subcoalitions. It expands the number and kinds of possible
supporters and sponsors.
Hage and Aiken (1970) seem to concur, but Normann (1971) disagrees, claiming that the major
innovations, which he calls 'reorientations', were made in companies which were either family
owned or otherwise had a strong concentration of power. Only a powerful leader is able to
overcome resistance to change and to make bold innovations. Rogers and Shoemaker's (1971: 384)
review indicates that Normann's (1971) position is the one most strongly supported by previous
studies.
Technocrats and professionals such as scientists and engineers possess the knowledge and
training that often make them most capable and motivated to discover new products and
processes. Professional employees may best be able to recognize the need for change (Hage and
Aiken, 1970: 33) and therefore firms that have a high percentage of influential technocrats will tend
to be the most innovative.
Mohr (1969) has emphasized the need for organizational resources in prompting innovation.
Most major innovations are too costly to be undertaken by organizations that are short of financial
capital. Abundant material, capital equipment, and human resources are also necessary. For
example, some kinds of innovation require laboratories, scientists, and financial slack resources
that are not needed for day-to-day operations. New product introductions often require much
expenditure for R & D, test-marketing and changes in production facilities.
The final structural dimensions that we shall consider have been introduced by Lawrence and
Lorsch (1967). They are dijjerentiation and integration, respectively. For our purposes, the first will
refer to the extent to which an organization's products require different marketing and production
methods and procedures (our indicants of this scale will differ from Lawrence and Lorsch (1967)).
Hage and Aiken (1970) have argued that the existence of very different groups in the firm will make
available more varied sources of information for developing new programmes. Complex
innovations require a diversity and richness of inputs which are most likely to be available only in
differentiated organizations (Wilson, 1966; Thompson, 1969). But differentiation causes conflicts
among sub-units and departments unless there are integrative devices used to ensure effective
collaboration. It is necessary to keep departmental parochialism to a minimum. In carrying out
complex new product innovations it may be necessary for members of the R & D, marketing,
finance, and production departments to work together intensively. Unless there are integrative
devices such as task forces, interdepartmental committees, integrative personnel, or matrix
structures, collaboration is difficult and conflicts and mistakes result (Lawrence and Lorsch, 1967;
Galbraith, 1973).
Decision making variables
The final set of variables that can stimulate innovation in conservative firms describes the way
executives use and process information in decision making. Given that the organization gathers the
appropriate information about the environment and about organizational performance through
its scanning and control systems, and given that this information is communicated to appropriate
decision makers, it is still necessary for this information to be used and evaluated by executives
charged with making key decisions. For example, if conservative executives ignore relevant
information that signals the need to innovate, then innovation will not take place. Some important
decision making style variables are degree of analysis of information, amount of planning, and the
amount of explicit conceptualization of strategies. The more anal sis is performed by key decision
makers, that is, the greater the tendency to search deeper for the roots of problems and to generate
the best possible solution alternatives, the more likely it is for innovation opportunities to be
discovered and actualized. Managers who make seat-of-the-pants decisions are unlikely to spend
the time and effort required to recognize the need for innovations. This may be true, for example,
of Cyert and March's (1963) satisficing and uncertainty avoiding firms, and of Lindblom's (1968)
remedially focused organizations.
Planning horizons (or Juturity, as we shall call the variable) are also very likely to influence
organizational innovation. Executives who are concerned with putting out fires will be too
preoccupied with such matters of the moment to be able to assess the long-term adequacy of their
product lines and product designs. They will fail to perceive the need for innovation. The more
future-oriented the firm, the greater the concern with change and therefore with innovation
(Ansoff, 1965; Andrews, 1980).
Our final variable is the consciousness of strategy and concerns the degree to which strategies
have been explicitly considered and deliberately conceptualized. Executives whose attention is
devoted exclusively to non-strategic matters tend to muddle-through and are much less likely to
engage in product innovation (Mintzberg, 1973; Miles and Snow, 1978; Miller and Friesen, 1978),
but where there is a concerted attempt to decide upon the product-market orientation of the firm,
there is a greater likelihood that target markets will be defined more broadly. Consideration is
given to goals and opportunities, and therefore to programmes of innovation.
THE ENTREPRENEURIAL
MODEL OF INNOVATION
In sharp contrast to the conservative model, we now outline the entrepreneurial model which
applies to firms that innovate boldly and regularly while taking considerable risks in their productmarket strategies. The entrepreneurial strategy might be followed, for example, by Collins and
Moore's (1970) and Mintzburg's (1973) entrepreneurial firms, Miles and Snow's (1978)
prospectors, and Miller and Friesen's (1978) adaptive, innovative, and impulsive firms. According
to the entrepreneurial model, innovation is seen as good in itself, as a vital and central part of
strategy. The entrepreneurial model postulates that firms will engage in much innovation unless
there are certain key obstacles acting to stop it. First, innovation will be very high unless good
scanning or control systems reveal it to be too expensive or wasteful, that is, there will be negative
correlations of scanning and controls with innovation. Second, effective analysis of decisions,
futurity, and explicit and conscious considerations of strategy will also guard against the natural
tendency towards innovative excess. Here again we expect to find negative correlations. Third,
because strategy is expected to be the main driving force behind innovation, the role of
environment as an innovation incentive will be reduced. However, because innovation can itself
induce environmental dynamism and heterogeneity, there still should be positive correlations
between innovation and environment. Finally, the frequently observed positive covariance
between innovation and structural factors such as technocratization, and differentiation should
prevail, but at a lower level of significance than for conservative firms. This is again because goals
and strategy, not structure or environment, are claimed to be the prime causes of innovation in
entrepreneurial firms. Now we can derive some predictions that follow from the entrepreneurial
model.
Environmental variables
For entrepreneurial as for conservative firms, environmental variables are expected to relate
positively to innovation. Entrepreneurial firms are often found in dy)namic and hostile
environments because their venturesome managers prefer rapidly growing and opportuneful
settings; settings which may have high risks as well as high rewards. Such firms may even be partly
responsible for making the environment dynamic by contributing challenging product innovations
(Peterson and Berger, 1971). Because innovation prompts imitation, the more innovative the firms,
the more dynamic and competitive (hostile) their environments can become. Innovation is also
likely to be positively correlated with heterogeneity because firms that innovate are more likely to
come up with products and services that can be exploited in different markets (Chandler, 1962).
Notice that in entrepreneurial firms, unlike conservative firms, innovation may cause dynamism,
hostility, or heterogeneity, rather than the other way around. If so, the greater latitude for strategic
choice (e.g. to innovate in stable environments) will cause correlations between innovation and
environment to be lower in entrepreneurial samples than in conservative samples.
Information processing variables
Traditionally, the impact of information processing variables upon innovation has not been clear.
Although there have been numerous empirical studies, these often conflict. For example, Rogers
and Shoemaker (1971:373-374) found 46 studies that showed early adopters of innovations to have
greater exposure to information channels than later adopters, while 14 studies did not support this
finding. They also found that 12 studies showed that earlier adopters scan or seek information
about innovations more than later adopters, while two studies contradicted this. We believe that
for studies of business firms, this conflict can be resolved by examining the role of strategy. While
we postulated that samples of conservative firms would demonstrate correlations between
innovation and information processing, an opposite relationship might be obtained in
entrepreneurial samples. Some entrepreneurial firms may have a tendency to innovate too much. A
proclivity towards taking risks, and an innovation-embracing ideology can cause firms to squander
resources in the pursuit of superfluous novelty. An effective control framework can flag the need to
reduce the scope and expense of projects and to slow down an overly rapid pace of innovation.
Scanning the environment to monitor the more parsimonious strategies of competitors is also
expected to have a dampening effect upon innovation as opportunities for resource savings are
discovered.
Structural variables
For much the same reasons as those presented in the corresponding section of the discussion of the
conservative firms, most structural variables are predicted to have a positive correlation with
innovation in entrepreneurial firms. The only qualifier we must add is that, in general, the positive
relationships between structure and innovation should be weaker in the entrepreneurial sample.
This is because the innovativeness of entrepreneurial firms is believed to be determined more by the
strategy of the firm and the aims of its venturesome top managers than by structure. It would not be
surprising to find some entrepreneurial firms that have a tendency to innovate a great deal even
though their structures are less than ideal for this, according to the literature supporting the
conservative model.
The integration variable should be negatively correlated with innovation in entrepreneurial
samples. Integrative devices such as committees. task forces, and integrative personnel bring
important facts to bear upon decisions. The innovation proposals of enthusiastic but reckless
executives are likely to be pared down by departments whose aim it is to ensure effective resource
management and efficiency.
Decision making variables
The variables called anal}ysis, Jufturit.1, and consciousness oj strategy are expected to correlate
negatively with the degree of product innovation. Essentially the same rationale applies to
substantiate these predictions as was presented in the section on information processing variables.
Analysis, planning, and the deliberate attempt explicitly to formulate strategies will provide the
firm with a better knowledge of its opportunities and excesses. Any tendency to overindulge in
product innovation may be curbed by these activities.
METHODOLOGY
The variables and questionnaires
In order to test the predictions derived in the last section we employed a lengthy questionnaire to
gather information on variables of environment, information processing, organization structure,
decision making style, product innovation, and risk taking. Appendix 1 presents the questionnaire.
All scale items were averaged for each variable to obtain the variable scores. Table 1 presents the
construct reliability measure of each of our variables. In every instance, the Cronbach alpha
measure (which averaged 0.74 for all variables) well exceeded the guidelines set up by Van de Ven
and Ferry (1980:78-82) for measuring organizational attributes. Construct reliability therefore
appeared to be very acceptable.
Data sample
Our data sample consists of 52 business firms that range in size from sales of less than $2,000,000 to
those of over $1 billion. Mean sales are $237 million and the standard deviation is $649 million. The
average number of employees is 2270. Firms are in industries as varied as retailing, furniture
manufacturing, broadcasting, pulp and paper, food, plastics, electronics, chemicals, meatpacking,
publishing, construction, and transportation. No industry represents more than 10 per cent of the
sample. Still, we cannot pretend here to have a random sample since its geographic area is restricted
to the Montreal region, and because firms were chosen by teams of second year MBA students
according to their personal interests. However, because of the broad representation of types and
sizes of businesses, and because no one type of firm dominates the sample, these exploratory
findings should have a high degree of generality.
Variables
Environment
Dynamism
Heterogeneity
Hostility
Information processing
Scanning
Controls
Structure
Centralization
Technocratization
Resources
Differentiation
Integration
Decision making
Analysis
Futurity
Consciousness of strategy
Product innovation
Risk taking
*
Entrepreneurial
sample
N= 18
Mean
S.D.
Total
sample
N=52
Mean
S.D.
Cronbach
alpha
3.7
3.5*
3.9t
1.3
1.6
1.1
4.4
4.7*
4.7t
1.6
1.4
1.0
3.9
4.1
4.2
1.4
1.6
1.1
0.74
0.84
0.55
4.6
4.3
1.3
1.5
4.8
4.6
1.4
1.9
4.7
4.4
1.4
1.7
0.74
0.69
5.2
3.4t
4.4
2.8t
4.8
1.0
1.3
1.4
1.7
1.2
4.9
4.9t
4.2
4.6t
4.9
1.5
1.7
1.2
1.4
1.0
5.1
4.0
4.3
3.5
4.8
1.2
1.7
1.3
1.7
1.2
0.79
0.69
0.68
0.88
0.71
3.9
3.8
3.4*
2.3t
3.Ot
1.0
1.4
1.6
0.8
1.0
4.3
4.5
4.5*
5.3t
4.7t
1.5
1.4
1.4
0.8
0.8
4.0
4.1
3.9
3.5
3.6
1.3
1.5
1.6
1.6
1.2
0.62
0.83
N/At
0.77
0.91
All responses to the questionnaire were obtained by interviews. This ensured that executives
could have any vague items explained to them and it removed any problem of missing data. While it
is difficult to estimate a response rate, most interviewing teams were able to obtain cooperation
from the first company they contacted. About 30 per cent of the teams approached two or three
firms before they were able to gain admission to the firm to carry out their field study.
All respondents used in the analysis had the rank of divisional vice president or higher. In 67 per
cent of the cases, more than one respondent per firm completed the questionnaire. In such a case,
the ratings of the highest ranking respondents were used. Where responses differed by more than
two points on the scales among the respondents, responses were averaged. This happened for 8 per
cent of all the scores. In 73 per cent of the cases, the data was supplied by the chief executive. Inter-rater
reliability was adequate across all of the variables. The scores of the raters were significantly
correlated at beyond the 0.01 level of significance for all of the variables. In cases of diversified and
divisionalized companies, each division was treated as a separate entity to ensure that questions
could be answered unambiguously. Thus, five of the 'firms' in the sample were really 'divisions
rather than autonomous organizations. In every case the divisions represented profit centres and
were controlled on the basis of their financial performance.
To carry out our analysis, we had to split the sample into two groups which were unambiguously
conservative and entrepreneurial. Two dimensions were used to achieve this: innovation and risk
taking. Firms whose scores on innovation and risk taking averaged less than or equal to 3.5 on the 7
point scales were classified as conservative (innovation and risk taking were positively correlated
with a product moment correlation of 0.51 and an N of 52). Such firms tended to be risk averse and
engaged in relatively little product innovation. Firms whose score on innovation and risk taking
averaged greater than or equal to 4.5 on the 7 point scales were classified as entrepreneurial. Firms
with average scores of greater than 3.5 and less than 4.5 tended to be in a grey area. They manifested
high risk taking and low innovation, or vice versa, and therefore were deleted from the sub-sample
analysis because they could not be unambiguously classified.
10
having nothing to do with the strategies implied by our models, sources such as environmental
dynamism, information processing and decision making characteristics, resource availability,
industry, and firm size. Perhaps all firms receiving and analysing the same information about their
environments and having the same resources would be equally innovative. Or perhaps, instead of
being related to strategy, innovation is solely a function of industry differences, firm size, or
environmental dynamism. Fortunately, these alternatives could be ruled out. Table 1 reveals no
significant differences in the levels of dynamism, scanning, controls, analysis or resources between
the samples. Also, firms do not differ significantly in size (number of employees averaged 1830 for
entrepreneurs and 2480 for conservatives these figures were not significantly different at the 0. 10
level). Finally, we could detect no systematic differences in the industry composition of the
samples. For instance, both conservative and entrepreneurial firms were found in industries such as
retailing, food, banking, telecommunications, transportation, meatpacking, electronics, furniture
manufacturing, aircraft parts, and chemicals. Thus overlap was considerable. Some industries that
were unique to conservatives were drugs, distilling, apparel manufacturing, automotive parts and
computer services, while industries unique to entrepreneurial firms included boat-building,
plastics, and pollution control. It is reasonable to believe that both entrepreneurial and
conservative firms could be found in all of these industries.
To conclude, we seem, roughly speaking, to have effected a categorization that manifests both
discriminant and convergent validity. However, a word a caution is in order. We have shown that
so far our sub-sample statistics are consistent with our models of conservative and entrepreneurial
behaviour. We are therefore tentatively justified in applying the interpretations and predictions
derived from these models to the respective sub-samples. This does not mean, however, that there
are no interpretations inconsistent with our models that can also be used to explain the figures of
Table 1 ol the correlational and regression results that we shall be discussing. This is always the
case, but particularly here. Our models were couched in dynamic terms and relate to managerial
motives and cognitions, but the data are cross-sectional and do not measure motives or cognitions
directly. Thus, while our interpretations are and will be consistent with our data, they will be by no
means uniquely determined by them. Subsequent longitudinal research will be necessary further to
explore the findings and inferences of this exploratory study.
CORRELATIONAL
ANALYSIS
We can now begin to test the predictions that were implied by the conservative and entrepreneurial
models of innovation. Some of the relevant data are to be found in Table 2, which presents the
product-moment correlations between product innovation and all other variables for conservative
and entrepreneurial sub-samples and for the total sample. It shows rather strong confirmation of
many of the predictions of the respective models. We shall discuss the findings for each class of
variables in turn.
Environmental variables
Our discussion predicted that there would be significant positive correlations between
environmental variables and innovation for both types of firms, but that strategic choice rather
than environmental pressures would play a greater role in promoting innovation in entrepreneurial
firms. The opposite was postulated for conservative firms. The implication was that environmentinnovation correlations would therefore be higher for conservatives. While our results were
consistent with this conjecture they cannot be unambiguously interpreted. Although all
correlations are significant at or beyond the 0.05 level in the conservative sample, only one
11
Variables
Environment
Dynamism
Heterogeneity
Hostility
Information processing
Scanning
Controls
Structure
Centralization
Technocratization
Resources
Differentiation
Integration
Decision making
Analysis
Futurity (planning)
Consciousness of strategy
Conservative
sample
N = 29
Entrepreneurial
sample
N= 18
Total
sample
N= 52
Significance levels
of differences in
rs between samples
(Fisher transform)?
0.32t
0.36t
0.33t
0.36*
0.28
0.25
0.361
0.491
0.43$
NS
NS
NS
0.27*
0.38t
-0.36*
-0.41t
0.08
0.07
0.05
0.005
0.15
0.18
0.13
0.26*
0.11
0.25
0.15
-0.30
-0.09
-0.33*
-0.03
0.44$
-- 0.05
0.48t
-0.03
NS
NS
0.09
NS
0.09
0.13
0.41$
0.47t
-0.39t
-0.67t
-0.47t
0.09
0.20*
0.33t
0.05
0.001
0.001
t, t, respectively, indicate that the correlation coefficient is significant at the 0.10, 0.05, and 0.01Ilevel of significance.
? We wished to determine if the correlation coefficients of both sub-samples represented populations having the same true
correlation p. We tested the hypothesis using the ratio
Z1 -Z2
UT(ZI
Z2)
where Z. represents the Fisher transformed value of the correlation coefficients r" such that,
Z
log,(I
+ rx)
rxy
and
U(Z,-Z2)=
(N1-3
N2-3).
12
between the conservative and entrepreneurial samples of firms. Scanning serves to bolster
innovation in firms that are classified as conservative. Attempts to gather information from the
environment may make managers aware of the disadvantages of their own product lines and the
superiority of the product lines of competitors. Scanning also can point out changing customer
desires and buying patterns. In other words, scanning demonstrates the need for innovation to
conservative managers. It tells them that it is time to change products and product lines. The
conservative nature of these firms ensures that unless this occurs there will be very little
innovation.
The entrepreneurial firms show an opposite relationship between scanning and innovation. A
significantly negative correlation is manifested. Perhaps this finding is caused by information that
induces highly innovative and risk embracing executives to slow down. Scanning may reveal that
competitors succeed without introducing many new products, that they cut costs by taking
advantage of long production runs, product standardization, and the 'economies of stability'.
Market research may show that customers favour established brands, or pay attention to factors of
price and quality more than variety and novelty. The effect of such information might be to reduce
superfluous and expensive product-innovation.
The same duality in the findings occurs when we examine the relationship between the use of
organizational controls and innovation. Controls indicate the need for innovations in conservative
firms while pointing to the need to curb innovative excesses in entrepreneurial firms. It is interesting
that the relationship between controls and innovation is greater than that between scanning and
innovation. Controls provide concrete financial information that is harder to explain away or to
rationalize. Controls may indicate to entrepreneurial executives that a great deal of money has been
spent on innovation and that very little return has been forthcoming. They may show that reserves
of capital have been badly depleted, that productivity and efficiency has fallen, or that scrap rates
have escalated because of too much product line diversity or change. In conservative firms controls
may reveal significant declines in market share, a dramatic reduction in the sales of older, more
obsolete products, and declining profitability.
Structural variables
Modest support has been found for the predictions of Normann (1971) and Rogers and Shoemaker
is positively correlated with innovation. The reasons
(1971: 384) who postulated that centralization
for this prediction were given earlier. However, we hesitate to place much reliance on the finding
since none of our coefficients is significant at the 0.10 level, and the negative sample-wide
correlation coefficient appears to confuse things. Perhaps in some types of organizations,
centralization boosts innovation, while in others it serves to obstruct it. Miller (1979) found that this
relationship varied in magnitude and direction according to the developmental or evolutionary
path being followed by the firm. The confusion in the literature surrounding this relationship may
be due to a failure to distinguish carefully among different organizational types. Unfortunately,
our findings are not very enlightening on this point.
Technocratization is positively correlated with innovation in both of the sub-samples as well as in
the total sample. This is in line with most of the predictions in the literature (c.f. Hage and Aiken,
1970; Zaltman, Duncan and Holbek, 1973). It is interesting that the sample-wide correlation is
much more significant than the correlations of the conservative and entrepreneurial sub-samples.
This may be because the conservative firms are less technocratized and less innovative than the
entrepreneurial firms. When the sub-samples are combined, a very positive correlation results.
Unfortunately, this result too contributes little new information since it merely supports the
consensus in the literature.
13
14
= 3.74-0.17
SCANNING
0.14CONTROLS
+ 0.27 TECHNOCRATIZATION
-
0.31 FUTURITY
+ 0.20 CENTRALIZATION
+ 0.30 DIFFERENTIATION
0.15 RESOURCES
Partial Fratios for the variables were 2.1, 3.1, 4.9, 9.8, 7.2, 7.7, and 1.6, respectively, with 1, 16
degrees of freedom. All but the first and last were significant at beyond the 0.10 level. The R2 was
0.785, the multiple R was 0.886, and the adjusted R2 was 0.634. The overall F statistic was 5.21,
which is significant at beyond the 0.01 level with 7, 10 degrees of freedom.
The entrepreneurial model is strongly supported. It is notable that while environmental
dynamism and hostility were prominent in the conservative regression model, no environmental
variables approached significance in the entrepreneurial model. Perhaps then, as predicted,
strategic goals and top executive motivations are more important than environment in promoting
innovation in entrepreneurial firms. This is consistent with the significant relationship between
centralization and innovation. Powerful entrepreneurial top executives can more easily implement
bold innovations than can those who must share their power with more conservative counterparts.
15
The observed levels of significance for variables of technocratization and differentiation probably
result from the tendency for innovation to require experts and more complex structures, or,
perhaps, vice versa (Burns and Stalker, 1961; Lawrence and Lorsch, 1967). Finally, the negative
regression coefficients that were expected for information processing and decision making
variables also were found.
CURVILINEAR REGRESSION ANALYSIS OF THE COMBINED SAMPLE
The correlational and regression results suggest that there might be a curvilinear relationship
between information processing and decision making variables and the level of product
innovation. At low levels of innovation there should be a positive correlation between innovation
and information processing or decision making variables, while at high levels of innovation a
negative relationship would be manifested. An inverted U-shaped curve such as the one in Figure 1
is expected to result.
Information
Processing
Decision
Making
Variab les
HIGH
orr
LOW
\_X
LOW
H IGH
Product
Innovation
16
1. Scanning-2.61
F*-
2.
3.
4.
5.
6.
7.
8.
Total F=2.2
R2 =0.08
Total F=2.4t
R2=0.09
Total F= 1.3
R2=0.05
Total F=5.2?
R2=0.17
Total F=8.5?
R2=0.26
Total F=3.61
R2=0.13
Total F=7.0?
R2=0.22
Total F=6.6?
R2=0.21
* In all cases the partial F statistic is given for the 'innovation' and 'innovation squared' independent
variables, with 1, 49 degrees of freedom. The total Fapplies to the whole regression analysis and has 2,
49 degrees of freedom.
t,t ? indicate the statistical significance of an F-test at the 0.10, 0.05 and 0.01 levels, respectively.
Though the R2 statistics are small, this is not a serious problem since we are not using the
regressions to predict the scores along the dependent variables. We merely wished to show that the
relationship between innovation and intelligence variables is of the inverse-U type.
(formed by taking the mean score of analysis, futurity, and consciousness of strategy) shown in
Equation 7 is highly significant. Again, it is possible that there exists some substitutability among
the variables.
Finally, we decided to take a total composite of all five variables relating to information
processing and decision making. Again the curvilinear relationship between the composite and
innovation variables is highly significant, as we can see from Equation 8. In all except Equation 3,
our predictions were well supported.
CONCLUSI ON
Two very different models of innovation were proposed and tested. Each seemed to be substantially
borne out in different sub-samples of firms. The 'conservative' model views product innovation as
something done in response to challenges, occurring only when very necessary. The model predicts
that innovation will not take place unless: (a) there are serious challenges, threats, or instabilities in
the environment; (b) these are brought to the attention of managers and consciously analysed by
them, and (c) structural, technocratic, and financial resources are adequate for innovation. In
short, positive and significant correlations are expected of innovation with environmental,
information processing, decision making, and structural variables. The predictions of the
conservative model were borne out for our sub-sample of conservative firms.
A very different 'entrepreneurial' model was also proposed. This model predicts that innovation
is a natural state of affairs; that it will be boldly engaged in unless there is clear evidence that
resources are being squandered in the pursuit of superfluous novelty. The model postulates that
17
innovation will tend to be excessive and extremely high unless: (a) information processing
(scanning and control) systems warn executives of the dangers of too much innovation, and
(b) analytical and strategic planning processes and structural integration devices do the same. In
other words, negative correlations of innovation with information processing, decision making,
and structural integration devices are expected. The entrepreneurial model also predicts low order
positive correlations of innovation with environment and more significant positive correlations
with structural devices. Goals and strategies rather than environment or structure are seen to be the
key impetuses to iinnovation. Most of the predictions derived from the entrepreneurial model were
borne out by our sub-sample of entrepreneurial firms.
A central message that emanates from this research is that the determinants of product
innovation in firms are to a very great extent a function of the strategy that is being pursued. The
impact of structural, information processing, decision making, and even, to a lesser degree, some
environmental and structural devices appears to be a function of whether firms have adopted a
conservative or an entrepreneurial strategy. Many of the conflicts in the innovation literature that
have been highlighted by Rogers and Shoemaker (197 1), Downs and Mohr (1976), and Zaltman et
a!. (1973) show promise of being resolved when we begin to look at strategy as a mediating influence
in the relationships between innovation and its context. Indeed, one must take very seriously John
Child's (1972) suggestion that we view organizations in a less deterministic light and pay more
attention to the role of strategic choice. To understand the relationships among innovation,
structure and environment, it may be necessary to study managerial motives, ideologies, and goals.
On a somewhat more practical plane, the research suggests that in addition to making remedial
efforts to stimulate innovation in stagnating firms, it might also be useful to take care that
innovation does not become an end in itself. It is necessary to ensure that the rate of innovation
does not outstrip its utility or the organization's ability to pay. If Miller and Friesen (1980) are
correct, there may be a tendency for any organizational trend to have momentum, that is, to feed
upon itself, perhaps being protracted past the point of usefulness. This might be true of the drift
towards excessive conservatism or excessive entrepreneurialism. Practitioners should begin to
focus upon the second danger as well as the first.
We shall close with a note of caution. Our findings suggested the applicability of two distinct
models of innovation in different contexts, but these models are probably not the only ones
consistent with our data and indeed go beyond the data in the explanations that are offered.
Further longitudinal research into the impact of managerial motives and goals upon innovation,
and their relationships to environment and structure would at this time be most useful to help
provide a more solid basis for our conclusions.
APPENDIX:
QUESTIONNAIRE
Please answer the following questions for the industry that accounts for the largest ',>of your sales
(in other words, your principal industry). Always answer by circling the correct digit unless
otherwise noted. How rapidor intense is each of the following in your main industry? Please circle
the number in each scale that best approximates the actual conditions in it.
Environmental dynamism (V. 1)
I. Our firm must rarely change
its marketing practices to
keep up with the market and
competitors.
X
-
1 2
18
1 2
1 2
1 2
1 2
Are there great differences amongst the products/services you offer, with regard to:
About the same for all our
products
II
1 2
1 2
1 2
4 5
4 5
I 1 2
I
3
I
4
4 5
1 2
4 5
Il
1 2
4 5
1 2
1 2
15. governmentinterference
19
I
4
Scanning (V. 4)
Rate the extent to which the following scanning devices are used by your firm to gather
information about its environment:
Not ever used
1 2
lI
3 4
Il
1 2
1 2
1 2
Controls (V. 5)
Rate the extent to which the following control devices are used to gather information about the
performance of your firm:
Used rarely or for small
part of operations
Used frequently or
throughout the firm
i
1 2
I
l
1 2 3
--
1 2
1 2
1 2
I I
I
3
1 2 3 4
I
5
20
Centralizatlon(V. 6)
Which levelsof managementare usually responsiblefor makingdecisions of the following
types?
Divisional top
Functional top executives or
executives if functional ones
Topmost
if no divisional
Middle
divisional
levels of
managers
structure
structure
management
Capital budgeting
Acquisitions of firms
3
5
(Scales 26, 32, and 38)
3
5
(Scales 27, 33, and 39)
3
5
(Scales 28, 34, and 40)
3
5
(Scales 29, 35, and 41)
5
3
7
7
7
7
7
7
personnel
Technocratization(V. 7)
44. In decision making, there is
great reliance on personnel
with experienceand common
sense.
I I
1 2 3 4 5 6 7
Some of
both
professionals such as
engineers, scientists, and
accountants (less than I per
cent of people other than first
line production workers).
1 2 3 4 5 6 7
Varies
considerably
by functional
area
I I
1 2 3 4 5 6 7
About
5 per cent
of work
force
A minimum of a bachelor's
degree with specialization(e.g.
consulting, engineeringfirms,
etc.).
21
47. Capital
1 2
48. Skilled labour
4---
4-4+ 1
4 5
1 2
1
1 2
!
3
DIifferentiation(V. 9)
How many distinctly different (i.e. unrelated) product lines or services does your firm market?
51. Only one.
1 2
How similar are these product lines or services in terms of (i) the technology used to produce
them and (ii) their markets'?
52. Technology: very similar
technologies (e.g. all produced
with similar equipment).
53. Markets: very similar in terms
of required marketing
strategy, types of customers,
pricing, etc. (e.g. one product,
one market).
I I
1 2
I
4
1 2
22
l
1 2
To what extent is decision making at top levels in your firm characterized by participative,
cross-functional discussions in which different departments, functions, or divisions get
together to decide the following classes of decisions?
Rare use of committees or
infrequent informal
collaboration
Frequent use of
committees and/or
informal interdepartmental
collaboration
l l
1 2
I
1 2
1 2
II
1 2
1 2
3 4 5 6
Decisions of
the difficult
departments
neither help
or hinder
each other.
- l
1 2
l
3
II
senior managementgroups
for novel solutions to
problems.
1 2 3.4 5 6 7
l
1 2 3 4 5 6 7
1 2 3 4 5 6 7
23
on major decisions.
1 2 3 4 5 6 7
Futurity(V. 12)
67. Decisions aimed at the
1 2 3 4 5 6 7
1 2 3 4 5 6 7
Medium
term
orientation.
To what extent are the following activities carried out?
opportunities in the
environmentare most common.
Long term (over 5 years)
goals and strategies are
emphasized.
Very frequentlyand
intensively
Very
rarely/or
haphazardly
69. Long term forecasting of
sales, profits and the nature
of markets.
70. Long term forecastingof the
technology relevant to
products and services offered
by firms.
71. Planning of long-term
investments.
Consciousnessof strategies (V. 13)
72. Administrativeand product/
market strategieshave not
been explicitly conceptualized.
I I
I I I
1 2 3 4 5 6 7
I
1 2 3 4 5 6 7
4-X
1 2 3 4 5 6 7
I
1 2 3 4 5 6 7
24
1 2
How many new lines of products or services has your firm marketed in the past 5 years? Please
exclude mere minor variations.
74. No new lines of product or
services in past 5 years.
75. Changes in product lines have
been mostly of a minor
nature (e.g. putting in towel
with the soap).
Risk taking (V. 15)
76. There is a strong proclivity to
low risk projects (with normal
and certain rates of return).
77. Owing to the nature of the
environment it is best to
explore it gradually via timid,
incremental behaviour.
1 2
i
5
1 2
I
1 2
1 2
II
REFERENCES
Aguilar, Francis. Scanning the Business Environment, Macmillan, New York, 1967.
Andrews, Kenneth. The Concept of Corporate Strategy, Irwin, Homewood, Ill., 1980.
Ansoff, H. Igor. Corporate Strategy, McGraw-Hill, New York, 1965.
Baker, Norman, James Siegman, and Albert Rubenstein. 'Effects of perceived needs on the generation of
ideas in R&D labs', IEEE Transactions on Engineering Management, EM-14, 1967, pp. 156-163.
Burns, Tom and G. M. Stalker. The Management of Innovation, Tavistock, London, 1961.
Carter, Charles and Bruce Williams. Industry and Technical Progress, Oxford, London, 1957.
Chandler, Alfred D. Strategy and Structure, MIT Press, Cambridge, Mass., 1962.
Child, John. 'Organizational structure, environment, and performance: The role of strategic choice',
Sociology, 6, 1972, pp.2-22.
Collins, Orvis and David G. Moore. The Organization Makers, Appleton, Century, Crofts, New York, 1970.
Cyert, Richard M. and James G. March. A Behavioral Theory of the Firm, Prentice-Hall, Englewood Cliffs,
N.J., 1963.
Downs, Anthony. Inside Bureaucracy, Little, Brown, Boston, 1966.
Downs, George W. and Lawrence B. Mohr. 'Conceptual issues in the study of innovation', Administrative
Science Quarterly, 21, 1976, pp. 700-714.
Galbraith, Jay. Designing Conmplex Organizations, Addison-Wesley, Reading, Mass., 1973.
Hage, Jerald and Michael Aiken. Social Change in Complex Organizations, Random House, New York,
1970.
Keller, Robert and Winford Holland. 'Boundary spanning roles in an R&D organization', Academy of
Management Journal, 18, 1975, pp. 388-393.
Lawrence, Paul R., and Jay Lorsch. Organization and Environment, Harvard, Boston, 1967.
25
Lindblom, Charles. The Policy-making Process, Prentice-Hall, Englewood Cliffs, N.J., 1968.
Miles, Raymond E. and Charles C. Snow. Organizational Strategy, Structure and Process, McGraw-Hill,
New York, 1978.
Miller, Danny. 'Strategy, structure and environment: context influences upon some bivariate associations',
Journal of Management Studies, 16, 1979, pp. 294-316.
Miller, Danny and Peter H. Friesen. 'Archetypes of strategy formulation', Management Science, 24, 1978,
pp. 921-933.
Miller, Danny and Peter H. Friesen. 'Momentum and revolution in organizational adaptation', Academy of
Management Journal, 23, 1980, pp. 591-614.
Mintzberg, Henry. 'Strategy making in three modes', California Management Reriewv, 16, 1973, pp. 44-53.
Mohr, Lawrence B. 'Determinants of innovation in organizations', American Political Science Retview, 63,
1969, pp. 11 1-126.
Mueller, W. F. 'The origins of the basic inventions underlying DuPont's major product and process
innovations, 1920-1950', in R. R. Nelson (ed.), The Rate and Direction of Intentite Actitvity, Princeton
University, Princeton, 1962, pp. 323-360.
Myers, Summer and Donald G. Marquis. Successful Industrial Innotation, National Science Foundation,
Washington D.C., 1969.
Normann, Richard. 'Organizational innovativeness: product variation and reorientation', Administratile
Science Quarterly, 16, 1971, pp. 203-215.
Peters, Donald H. 'Commercial innovation from university faculty: A study of the invention and exploitation
of ideas', Sloan School of Management Working Paper, No. 406-69, M.I.T., Cambridge, Mass., 1969.
Peterson, R. and D. Berger. 'Entrepreneurship in organizations', Administrative Science Quarterly, 16, 1971,
pp.97-106.
Rogers, Everett M. and F. Floyd Shoemaker. Communication of Innovations. A Cultural Approach, Free
Press, New York, 1971.
Rosner, Martin M. 'Administrative controls and innovation', Behatvioral Science, 13, 1968, pp. 36-43.
Thompson, Victor A. Bureaucracy and Innovation, University of Alabama Press, Alabama, 1969.
Tushman, Michael L. 'Special boundary roles in the innovation process', Administrative Science Quarterly,
22, 1977, pp. 587-605.
Utterback, James M. 'The process of technological innovation within the firm', Academy, of Management
Journal, 14, 1971, pp. 75-88.
Van de Ven, Andrew and Diane Ferry. Measuring and Assessing Organizations, Wiley, New York, 1980.
Wilson, James Q. 'Innovation in organization: notes toward a theory', in James D. Thompson (ed.),
Approaches to Organizational Design, University of Pittsburgh, Pittsburgh, 1966, pp. 193-218.
Zaltman, Gerald, Robert Duncan, and Jonny Holbek. Innotvations and Organizations, Wiley, New York,
1973.