WHEN IT COMES TO PRODUCT INNOVATION, WHAT IS SO BAD ABOUT BUREAUCRACY?
DEBORAH
DOUGHERTY McGill Universiy
SARAH
M. CORSE
University of Virginia
In theory, bureaucracy is said to be “bad” for innovation. Yet large bureaucratic organizations often need to be both adept at innovation and capable of ongoing routinized production. Unfortunately, little is actually known about how bureaucracy hinders innovation, since research on the relationship of structural characteristics of organization and innovation is ambiguous. This study attempts to illuminate the specific relationship between bureaucracy and innovation so as to permit progress beyond the innovation-bureaucracy stalemate. Building on an interpretive rather than structure perspective, we analyzed 134 people’s experiences with product innovation in 15 large bureaucracies. We found four patterns of bureaucratic thinking and acting that systematically inhibited effective action in defining, organizing, evaluating, and staffing the innovation effort. These patterns are described and illustrated at some length. Then, drawing on Weber’s original theory of bureaucracy and other classics of innovation, we theorize that it is an interpretive system of instrumental rationality, not only a bureaucratic structure, that creates and maintains these patterns. From this insight we speculate on how large bureaucracies can become more innovative without losing their efficiency.
Heightened global competition and changing customer needs are pressuring large, bureaucratic firms to change their product offerings, or even to develop new ones. Unfortunately, organization theory has little to say about how bureaucratic organizations can become more innovative and still generate established products efficiently. On the one hand, theorists argue that the bureaucracy is here to stay because it is an efficient way to organize complex but routinizable tasks (Meyer 1990; Perrow Direct all correspondence to: Deborah Dougherty, Department of Management, McGill University, Montreal, Canada H3A IG5; Sarah M. Corse, Department of Sociology, University of Virginia, Charlottesville, VA 22903. The Journal of High Technology Management Research, Volume 6, Number 1, pages 55-76. Copyright @ 1995 by JAI Press, Inc. All rights of reproduction in any form reserved. ISSN: 1047-8310.
56
THE JOURNAL OF HIGH TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. I/ 1995
1986). Ideally speaking, according to Weber (1947, p. 339): “Bureaucratic administration means fundamentally the exercise of control on the basis of knowledge. This is the feature of it which makes it specifically rational.” On the other hand, evidence suggests that bureaucracy is bad for innovation because it can squelch the creativity, nonstandard activity, and speed of response innovation requires (Galbraith 1982; Quinn 1985; Johnson 1988; Pearce & Page 1988). Popular writers admonish managers to “bust” their bureaucracies to become innovative, but define the alternate organizing form only vaguely in terms of nonbureaucratic structure, that is, nonhierarchical, informal, decentralized (see Dumaine, 1991). Managers of large bureaucratic organizations are caught in a paradox: they must innovate and maintain bureaucratic efficiency, but in theory the two do not mix. One reason for this apparent paradox is that we do not know how bureaucracy is bad for innovation. Following Blau and his colleagues (e.g., Blau & Meyer, 1987) and the Aston group (Pugh, Hickman, Hinings, & Turner, 1969), many innovation in narrow structural terms such as hierarchy, researchers define “bureaucracy” centralization, size, specialization, and formalization. But as Mohr (1982) and Nord and Tucker (1987) show in two thorough reviews, these structural characteristics do not explain the relationship between bureaucracy and innovation very well. Studies on the relationship generate ambiguous results. For example, centralization removes decision making from those who may know the most, but direct participation of senior management focuses a team’s efforts and removes barriers (Day, 1994). Adherence to formal rules inhibits creativity (Dougherty & Heller, 1994) but formal processes such as phase reviews enhance a product’s design and development speed (Cooper 1983). Centralized, informal structures enhance the adoption of radical process technologies, but complex, decentralized structures enhance new product introductions (Ettlie, 1988). Nord and Tucker (1987) conclude from their in-depth analysis of the literature that innovation requires an organization that is flexible, concentrates power sufficiently, generates effective inputs, and has appropriate levels of technology. They found that a variety of structural configurations can produce these conditions, and conclude that structural views of organization cannot adequately explain how to organize, or not, for innovation. The goal of this study is theory building to get beyond the paradox of innovation and bureaucracy. For all its flaws, the bureaucracy remains society’s most legitimate form of organizing, and is taken for granted by many. People cannot simply abandon it for a new system of organizing, especially if that system is not fully described. Theories of change management make it clear that people need to know both what is wrong with their current practices (i.e., unfreeze), and how they should change (change and refreeze), see Schein, 1979). In this paper we take a first step toward developing a more useful theory of organizing for innovation by identifying what specific aspects of bureaucracy hinder new product development. From that analysis we develop some insight into why these barriers arise and speculate on what managers can do to overcome them.
ANINTERPRETIVEPERSPECTIVEOFBUREAUCRACY The ambiguous research record indicates of bureaucracy are not strongly associated
that conventional structural characteristics with innovation. An interpretive perspective
Bureaucracy and Product Innovation
57
on bureaucracy offers another vantage point from which to explore how it might hinder innovation. An interpretive perspective considers the ways in which people define a situation, develop and maintain shared understandings, and make sense of their and others’ beliefs and experiences. The conventional structural view sees “organization” as a fairly static network of roles and relationships which channel information, decisions, and authority. In contrast, an interpretive perspective sees “organization” as a process of social action which creates and channels meaning. People understand their world as they act, socially constructing meaning in context with reference to others (Daft & Weick, 1984; Smircich & Stubbart, 1985; Weick 1979). In that process, people use shared understandings or cognitive maps which may range from simple “theories of action” (cf. Argyris & Schon, 1978) to institutionalized systems of knowledge and beliefs (Berger & Luckmann, 1967). Bureaucracies are key institutions of modern society and, we believe, engender their own set of shared understandings, some of which, we will show, impede product innovation. Both bureaucracy and innovation have been studied from interpretive perspectives. According to Weber (1947), the bureaucracy embodies a particular type of rationality, which he defined as a form of social action or a worldview. Weber distinguished two types of rationality. Substantive rationality is oriented to values, such as religion or duty, which are pursued for their own sake. A substantively rational organization would emphasize complex qualitative goals such as justice or quality of life and commitment to those goals. Choices are guided by values rather than by rules, but are rational because values are consistent. In contrast, instrumental rationality is oriented to “the methodological attainment of a definitely given and practical end by means of an increasingly precise calculation of means” (Weber 1946, p. 293). An instrumentally rational organization would emphasize what Parsons (in Weber, 1947, p. 36) called “enormously simplified” goals and standards, universally agreed upon rules, quantification, and formalistic impersonality. Weber based his theory of bureaucracy primarily on instrumental rationality, not substantive rationality, even though many theorists gloss over his distinctions and use the general term “rationality” in discussing bureaucracy. March & Simon (1958) att~buted a narrow form of rationality to limits in human ability rather than to a worldview. Nonetheless their theory of organization is both similar to Weber’s, as Perrow (1986) argues, and based on an interpretive view. To them, the organization is comprised of programs, decision premises, and repertoires of action that operate unobtrusively to limit information content and flow, highlight some aspects of a situation over others, and limit search. A number of interpretive views on organizations have emerged since, such as “garbage can” decision-making from a loose formation of decision rules (Cohen, March, & Olsen, 1972), organizational learning (Hedberg 1981; Starbuck & Milliken, 1988), organizational cognition (Sims & Gioia, 1986), and organization culture and symbols (Gagliardi, 1990; Schein, 1985). Together these and related works provide a good base for an interpretive view of bureaucracy, but they have not been applied systematically to product innovation. Innovation researchers have also developed interpretive approaches, however they are in turn unconnected from organizational ones. Perhaps the most famous, but unacknowledged, is Burns & Stalker (1966). They defined their mechanistic and organic organizational forms as “interpretive systems,” not structures, even though most have redefined their forms in structural terms. Burns & Stalker’s non-innovative mechanistic
THE JOURNAL OF HIGH TECHNOLOGY
58
MANAGEMENT
RESEARCH
Vol. ~/NO. I/ 1995
interpretive system was characterized by a differentiation of tasks that have been abstracted from the whole, a precise definition of rights and obligations, hierarchical communication and control, and an emphasis on internal (local) knowledge. Their innovative organic system was characterized by an orientation to the interconnectedness of tasks, the spread of commitment and responsibility beyond functionary definition, network control and communication, and an emphasis on external (cosmopolitan) knowledge. Burns & Stalker argued that “codes of conduct” for practice sustained the two forms, but they never defined these codes beyond a vague sense of culture. Unfortunately, innovation research since has not developed Burns and Stalker’s ideas on “codes of conduct” either. Others have, however, reiterated their idea that an overall perspective like a mechanistic interpretive system (which can be seen as the worldview of instrumental rationality) thwarts innovation. For example, Schon (1967) argues that what he called “technical rationality” hinders innovation by reinforcing sequential development, functional separation, and precise measurement; Nelson & Winter (1977) suggest that “economic rationality” ignores the complex, uncertain realities of innovation; and Kanter (1983) describes the negative effects of a “segmentalist culture.” Research has also considered the more day-to-day level of social action, where things like decision premises or codes of conduct thwart innovation. For example, Burgelman (1984) describes the anti-innovation effects of “administrative neglect” and an orientation to the current strategic and structural context. Van de Ven & Polley (1992) show how tacit patterns of funding-based planning, fluid participation, and a proliferation of activities inhibited effective learning during a new product effort. Similarly, Dougherty (1992a) shows how entrenched routines for defining markets and evaluating products kept innovators from understanding actual market needs. Research Questions. These various threads suggest that an interpretive perspective can help us understand how bureaucracy hinders product innovation. The interpretive perspective does not assume that people’s perceptions are necessarily accurate from an “objective” viewpoint. It does assume that people act on the basis of their perceptions of the situation, so studying those perceptions is important. To develop the potential of this interpretive approach, we focus on the particular activities of new product development in order to ground the analysis in both the specifics of new product development and the everyday world of people at work. A large literature has established that product innovators ought to do four things: 1) 2) 3) 4)
work closely with potential customers; work creatively to integrate technology with market needs; collaborate in multifunctional teams to pool diverse perspectives and coordinate efforts; and build on the firm’s resources and strategy (Cooper 1983; Souder 1987; Zirger & Maidique, 1990).
Research also shows that people have difficulty in carrying out these necessary activities. We investigate why these difficulties occur in bureaucracies first by looking for underlying patterns of thought and action that may get in the way of effective product innovation. We do not assume that simply generating innovation is good, but rather ask why people have such difficulty even when they decide that innovation is necessary
Bureaucracy
and Product Innovation
59
for a particular situation. We do assume that people in large firms wish to innovate more effectively (as indicated by surveys, e.g., Arthur D. Little, 1991), but not necessarily that they wish to innovate more often. Second, we develop a conceptual understanding to make sense of what we find. We draw on the ideas summarized above to suggest why these patterns exist and what can be done about them.
METHODS
Smircich & Stubbart (1985) suggest that to develop an interpretive understanding, one must consider people’s reasons for their actions and the meanings they assign to events. To generate data which reflect interpretive insights, we first approached bureaucratic firms and asked to interview people who were actively working on a new product. The 15 firms that gave us access represent a variety of industries, and average 96 years of age, 54,000 employees, and $9.4 billion annual revenue. Next we selected products in consultation with a firm’s representative that embodied unfamiliar technology, were intended for unfamiliar markets, or both. As shown in Table 1, the products vary on their current success status, which was determined through follow-up phone calls two years later, and on degree of innovativeness (see Table 1 for how each was coded). Then we interviewed 134 people from a variety of departments who were working on one of the products identified. Our focus in the interviews was on the person’s understandings of the specific activities of innovation in the particular case; why events happened as they did and how the person interpreted them. To that end, the innovators were asked to tell the story of the product, and to describe how much they knew about the market and technology, how they worked with other departments, two problems in the innovation effort that came up and how they were solved, and if the firm had “got in the way.” The interviews averaged one hour each. The people were all white collar managers, engineers, or scientists from different departments, and averaged nine years of tenure with the firm. Eleven described products that had already been cancelled, while 123 described products that were still in development. Our primary unit of analysis is the innovators’ experiences, not the merits of a particular product. For each product, we tried to talk to at least several people who played an important role to get multiple views about events in the product’s history. For 30 of the 40 cases, we talked to all those identified by the contact person or product manager as playing an important role in the product. We acquired multiple interviews for five of the remaining cases. This is a nested design (Yin, 1989) in which the experiences of innovation are nested within a variety of products which in turn are nested in a variety of firms. The extensive variation in the data across industries, firms, and product types meets the criterion of richness recommended by Bailyn (1977) and Eisenhardt (1989), allowing us to develop a more broadly grounded understanding that is not bound to only one firm or innovation type. In addition, multiple researchers brought different perspectives to bear on the data. Data were collected over a two year period for many of the cases as well, so we have some (albeit limited) insight into processes over time. We also combined first order qualitative data with second order quantitative data. As Eisenhardt (1989) suggests, combining these two data types helps researchers avoid focusing in on vivid but perhaps limited qualitative impressions, while at the same time understanding processes behind quantitative relationships. One important limit is that the data come
Notes:
5 5
Food Item (3) Chemical (4)
Bond (3)” Food Item (7) Document System (2)
Success
12 8
Medical Product (3) Computing System (2) Robot (5) Office Machine (1)
Specialty Chemical (1) Lubricant (3) Building Product (1) Machine (4) Solvent (4) Office Machine (1)
Bond (1) Information Package (1)
Cancelled in Developmenl
4 4
Communication System (2) Site Mgt System (2) Waste-fix (4) Food Item (10)
(shows that results not due to data coming from only 1 or 2 firms)
16 IO
Medical System (2) Expert System (4) Expert System (7) Warehouse System (1)
Rock Crusher (2) Document System (5) Office Product (4) Chemical (2) Chemical (3) Information Package (I) Food Item (5) Chemical (3) Communication System (4) Plastic (6)
Bond (8) Information Package (2)
In Development
Current Status
’ Number of individuals interviewed. ’ Five different ways to be innovative: new users, new uses, new product technology, new manufacturing, new distribution.
No. Products: No. Firms in Cat:
High (Unfamiliar in four or five ways)
Medium (Unfamiliar in three ways)
(U~amiliar in one or two ways)
LOW
Znnovativenessb
In Market Success Not Clear
TABLE 1 New Product Cases by Current Status and ln~~vatjvaness
3 3
Operating System (4)
Specialty Chemical (3)
Bond (4)
Failed
Bureaucracy and Product Innovation
61
only from large bureaucracies. We cannot compare our underlying patterns with those in nonbureaucratic firms to assure that they are different, so important follow-up research remains to be done. To search for possible patterns of bureaucratic thinking and acting in the interview data, we used the procedures for qualitative analysis described by Strauss (1987). The first step in the Strauss approach is “open coding.” Working together and with research assistants in multiple coding sessions, we closely scrutinized people’s descriptions in a subset of interviews of how established practices failed them, how procedures that did not fit were imposed anyway, and problems with understanding customers, working with others and so forth. We were looking at how their ideas were organized, and trying to understand the expectations, rationales, and attributions they used in order to identify patterns of thinking and acting that lay behind the problems people described. For example, when people described problems with other departments, we asked: What conditions contributed to this inability to collaborate? When they said that senior managers did not know much about innovation, we asked: Why were senior managers seen as inaccessible or incompetent? The outcome of this step was a set of initial patterns of thinking and acting that hindered product innovation. The second step is “axial coding.” Here, each initial theme was explored across the data in different types of situations, product types, and firms in order to “test” the robustness of the idea and/ or clarify it. The analysis was guided further by descriptions of how problems were sometimes solved, because (in this case) people often explained why usual approaches did not work and how they developed alternate approaches. Continuous comparative analyses such as these helped to sharpen the underlying patterns. The third step is “selective coding.” In this step, the boundaries of each pattern are clarified to ensure conceptual distinction by content-analyzing the interviews. First, we specified problem indicators for each pattern, and had research assistants code the descriptions of organizational problems in all interviews into one of the four patterns based on the indicators. The problems had been previously identified by Dougherty & Heller (1994). The inter-rater consistency among the assistants was high, which suggests that the indicators for each pattern could be distinguished.’ The codes also allowed us to count how often each pattern occurred, and to compare and contrast the relative frequencies across different groups of innovators to check our inferences further. Our analysis identified four patterns of bureaucratic thinking and acting that inhibited the activities of effective product innovation. While we cannot assure that the patterns are exhaustive and mutually exclusive, the Strauss method emphasizes the development of categories with these characteristics. 1.
Definition of the Product through Inward Orientation: Following this bureaucratic pattern, innovators defined their new product narrowly from the inside out, by emphasizing their firm’s internal resources and technology. This penchant to concentrate on the technological side suppressed people’s ability to appreciate the external world of customers and markets, to see how customers would use the product, to understand how they perceive value, and to translate customer issues into product attributes. Instead, innovators tended to ignore or misinterpret customer needs.
62
THE JOURNAL
2.
3.
4.
OF HIGH
TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. l/ 1995
Organization of Innovation through Linear Progression: Following this pattern, the innovation process was organized in a fixed, sequential, and linear fashion. Not only was work segmented, but innovation was presumed to progress methodically along already defined paths, where activities were only a small, incremental step from routine. This pattern suppressed people’s ability to react quickly to unexpected developments, to take up discontinuous lines of thought or action, to incorporate surprising feedback on changing conditions, or to learn from problems. Instead, people tended to conform to routinized processes even though they did not work. Evaluation of Progress through Detached Judgement: Following this pattern, people evaluated the new product with standard rules or measures that were detached from the complex reality of the market and technology. This pattern suppressed the ability to evaluate the product’s new or unfamiliar aspects, to develop workable heuristics, or to devise useful milestones to monitor progress. Instead, the pattern produced false impressions of measurability and control and kept people from understanding real progress. Staffing Projects Through Dislocated Responsibility: This pattern chopped up accountability and responsibility into fixed and narrow roles, which made it hard to develop the collective accountability innovation requires. Ironically, this pattern located the responsibility for product innovation in individuals, rather than in organizational positions as is the case with normal bureaucratic work. Dislocated responsibility discouraged people from becoming involved because innovation was not seen as legitimate work, forced innovators to adopt a marginal or maverick role, and prevented the development of a sense of collective commitment to innovation.
Table 2 shows that problems with each pattern are widely distributed across the products and firms. This is consistent with our idea that these patterns do not derive solely from specific organizational cultures or innovation experiences, but rather are properties of large bureaucracies in general. If these patterns thwart product innovation,
TABLE 2 Proportion of Firms and Products with Bureaucratic Patterns Inward Orientation (Problems with understanding customers, markets)
Linear Progression (Problems organizing project, connecting to firm)
Dislocated Responsibility Detached (Problems with Judgemen t (Problems of risk averse climate, getting imposition of commitment) standards, rules)
Percent of Firms with one or more people with indicator
87%
100%
93%
100%
Percent of products with one or more people with indicator
80%
93%
88%
93%
63
Bureaucracy and Product Innovation
TABLE 3 Average Proportion of Problems that were Solved by Products’ Success Status Inward Orientation (Problems with understanding customers, market) N= 74 Total Average** Success Status 1. Successful 2. Out, uncertain 3. In development 4. Cancelled in development 5. Failed
Notes:
Linear Progression (Problems organizing project, connecting to firm) N=98
Dislocated Responsibility Detached (Problems with Judgement (Problems of risk averse climate, imposition for getting standards, rules) commitment) N= 74 N=86
40%
38%
26%
42%
75% 45 25
47% 35 40
57% 30 18
62% 38 51
42 9 *p = .07 without failures
60 6
08 0 p = .02 without failures
32 0
NS without
Products innovativeness, stage of development, * based on one-way ANOVA ** only those who had problems are included
failures
people’s department
NS
do not affect conclusions.
we would also expect that successful product innovators would solve problems more frequently than others. Table 3 shows that, on the whole, successful innovators overcame these anti-innovation patterns more often. Looking at specific patterns, successful innovators solved more problems associated with designing and development the product (i.e., Patterns 1 and 3). They did not solve more problems associated with organizing their work (Patterns 2 and 4), although the data for Pattern 4 are in the expected direction. We infer that connecting products to the firm and getting commitment were not as necessary to a particular product’s commercial success as its design. However, difficulty with these organizational activities even for successes suggests that the organizations overall were not conducive to innovation. Below we describe each pattern in detail. We also describe how people who overcame the pattern tended to create alternate approaches to thinking and acting which enabled them to carry out the innovation activity more effectively. Definition of the Product By Inward Orientation Research shows that successful product innovators link technological possibilities with market needs to create a comprehensive package of design attributes (for reviews see Dougherty, 1992b; Rothwell, 1977; Zirger & Maidique, 1990). Successful product innovators have more knowledge about how the product fits into customers’work, what the customers’ problems are, and how customers evaluate solutions to those problems (Cooper, 1983; von Hippel, 1986). However, when we examined the difficulties with product conception and incorporating market insights, we found a tendency to define new products primarily in terms of technologies. Many innovators focused on the product’s physical features such as how fast, hard, or resistant it was. Some did not
64
THE JOURNAL OF HIGH TECHNOLOGY MANAGEMENT RESEARCH Vol. ~/NO. I/ 1995
consider knowledge of customer needs to be necessary. Others knew knowledge was important, but could not figure out how to make sense example, this planner was aware that she and her colleagues did not customers as well as they should, so they hired a consultant to teach the But here she explained how gathering customer data was so foreign to their ways that they would rather, in effect, “retreat” to the technology:
that such of it. For appreciate team how. established
We generally approach things too piecemeal; we muddle along. So we called in a consultant to help us [learn to talk to customers]. The consultant had an iterative process to get ideas and insights from customers in one and a half weeks. He gets into the interactive process quickly. We were uncomfortable with decision making that goes that fast. Instead, we like to go away for three months and work on the product, and then come back and say to customers: “Here it is.”
This inward orientation could also be seen in descriptions of usual approaches analysis, where “the market” was seen in piecemeal, idiosyncratic terms:
to market
Doing market research is new at ZCO. With the old process, product design evolves as customer needs present themselves or as design problems come up. There was no marketing per se.. . We never did analyses in the traditional marketing sense. It was always sales support. We would develop a product, and then find a customer to try it on. A lot of [these products] failed because if they didn’t work for the first customer, that would be it.. . The market focus in this division is zero. Say we have a product that we think Proctor and Gamble might like. We go to them and say: “Do you like it?” If they say: “Well, maybe not,” that would be the end of that business. In fact, all the soapers like Lever Brothers may be interested, but we never ask.
In another technology in:
example, a sales director for a cancelled product recalls his attention to the and his concomitant glossing of what customers were actually interested
When we first talked to possible customers, we said: “Look at all the good information the device can capture abut the operation of your machinery.” They said: “Great, but what do we do with all that information?” So Jack worked like a tiger for three months writing new software that would summarize the information. We went back and said: “Remember when you asked what could you do with all this information? Well, here are bar charts and histograms.” They said: “Great, but what do we do with the bar charts and histograms?” Boy, I tell you, they sure had us there.
Defining the product as technology was also self-consciously overcame this pattern of thinking and acting. Here, a planner first approached potential customers in the “usual fashion”:
addressed by some who described how his group
Our division typically sells amine, which is a nitrogen with a few hydrogens. Another manufacturer can combine these with other material-for example, amine and acid makes nylon. We found out that we could make a few different kinds, so we went to the marketplace and said: “Hey, we’ve got great amine if you want to buy them.”
Bureaucracy and Product Innovation
65
The group quickly found out that customers were much more interested in another intermediary that combined amine with other chemicals. The innovators explained that this time they caught themselves and deliberately stepped out of their accustomed behavior. Instead of simply pushing their new amine as usual, they listened to customers, and then developed a more comprehensive understanding of their product that incorporated how customers would use the product and the actual performance requirements they had. This more complete understanding helped the innovators see the product attributes as a complete package, and thus to make informed trade-offs among product quality, cost, delivery schedules, and performance as the development proceeded. Organization of Innovation Through Linear Progression Research shows that developing a viable new product requires creativity, experimentation, and occasional discontinuous leaps in imagination (see Van de Ven, 1986). To craft the full product under such conditions, product innovators need to pursue multiple paths, test a variety of attribute mixes, and respond rapidly to customer feedback. This requires mutually adaptive interdepartmental collaboration, in which people jointly create the product over time (Mintzberg & McHugh, 1985). When we examined how people actually organized their innovation activities or organized the product’s connection to the firm, we found an underlying understanding of appropriate work as a fixed, sequential, and linear process. One aspect of linear progression was to separate development activities, as in this planner’s description of how they normally first create and refine a plastic, and only then think about possible markets: [We normally] would do resins in ZCO very methodically. Something would break in the lab or a research guy working along a particular avenue would create a unique material, and then they would hunt around for a market. For example, we developed a fiber product that skipped the knitting and weaving process. Our initial thrust was to try and sell it for fabric, but now 20 years later we use it to line drainage ditches.
Other ZCO people told similar stories of how people would invent a technology and then spend years hunting around for applications. One material was developed for automobile tires, but ended up 10 years later being used to make combat gear. Unfortunately, ZCO’s penchant to break up innovation into discrete activities, took far more development time than they could afford as competition increased. Even though people recognized the inadequacies of linear progression, this pattern was easy to fall into. Consider this person’s recollection of the launch of a new product that was designed to prevent the counterfeit of important documents. The description appears ludicrous to an outsider. However, he was the sales manager at the tail end of a progressively linear process, and he was simply doing his job, developing sales materials by following procedures which everyone understood to be appropriate: We launched the product, and we did a brochure, and we did an in-house sales video that won us a very prestigious award for sales education.. . And then we also launched a formal market research program. We wanted to confirm the key vertical
66
THE JOURNAL
OF HIGH TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. I/ 1995
markets, and get specific customer viewpoints on what documents they are concerned about. What we found was that there was zero customer awareness of security as an issue. So we stopped the market research since we couldn’t find anyone to talk to about it.. . These people had no way to circle back and rethink one activity in light of another. Unfortunately, the missing market assessment continued to plague the product, since its sales were still quite modest after two years. The penchant to proceed in a linear fashion is also clearly expressed in descriptions of how this innovation-inhibiting pattern of thinking and acting was overcome. A packaging expert compares his successful team’s development style, in which they created in effect a multi-pronged, interactive network, with his normal work: The reason the product was successful was that it was not in the usual organization. I feel very strongly about that. You need to take all the barriers down and say: “Hey, you guys are good at what you do, so go and do it”. . . . No one from the top bothered us, so we could do what had to be done. What you lose in the matrix [DCO’s normal structure] is contact. But if you are dedicated to a team, you can get involved and spend time in other areas and understand their problems.. . Now I am back in the matrix and I have 30 or 40 products to work on, and I can’t contribute much to any of them.
The manager of a successful product comprehensive organizing mode:
at another
firm also described
a more inclusive,
At EC0 there is a lack of continuity on developing businesses. People move every nine months.. . From the beginning we created a radically new organization. It has two concepts. The first is that we did not want turnover, the theory being that if you are going into a business that you know nothing about, you don’t want to keep retraining people. Second, we had a long term incentive program installed [to compensate for the loss of a career ladder].
These two examples highlight the use of what Yang & Dougherty (1993) label anchoring. Anchoring is a learning process in product innovation in which the team “drops anchor” on one problem area, working it out while working out other issues in terms of it. The team then iterates to another multi-problem task, again anchoring in one function but tying that to all the others. In this study as well, innovators worked out organizational problems caused by linear progression by anchoring: creating an alternate approach in which they could work collaboratively, revising decisions, learning from experiments, and thinking more comprehensively and creatively. As these examples also illustrate, the innovators had to violate the prevailing bureaucratic pattern of linear progression to do so. Project Evaluation
By Detached Judgement
New products can consume considerable resources or easily fail, so it is important to monitor them. Because innovation is inherently ambiguous, monitoring and evaluating new products relies on qualitative judgements, heuristics rather than rules, and knowledge which may be situation-specific and based on trial and error learning
67
(Nelson & Winter, 1977; Quinn, 1978). Instead of monitoring and evaluating their innovations in this creative way, many we interviewed said that they or their managers refied on general standards that were detached from the rich, compiex reality of innovation. The detached judgement suppressed the ability to evaluate the unexpected expenses, ambiguous possibilities, or uncertain knowledge of product innovation. Perhaps the most obvious negative effect of detached judgement on innovation can be seen in the application of short-term accounting measures to the emergent innovation process. Decision makers attempted to abstract out from a complex reality formulaic notions of cost/ benefit analysis, even when the necessary data were not available. This drew people’s attention away from actual progress so that they could concoct “numbers.” Consider the frustrations of these innovators: The hamper to innovation is the analysis process we get into. Rather than use judgement, they would much rather gather lots of data.. . The business analyzers sit here day after day and get reams and reams ofprintouts, all to see if the product is good or bad. That is nonsense. Nobody can look into the future. Traditionally we had to prove gO% assurance that the revenues are there. I. They [management] want all the numbers that tell us it is profitable, with no risk taking an their part. That takes two years. _.
The pattern of detached judgement was also embedded in the general “rules of thumb” used by a firm to decide what is a “proper” new product. For example, one firm’s rule of thumb was that it was a “pioneer” who entered a new market niche first and made money from premium prices. One case in this firm was a follower product, so the innovators had considerable trouble convincing others in the firm that the idea had merit. Another firm had the opposite rule of thumb-it always entered markets as a follower and standardized by emphasizing quality. Here, innovators who developed a pioneering product had the same difficulty as those in the first firm--others were unwilling or unable to consider the potential of their idea. This is not to say that concerns over fit with a firm’s standard approaches were not valid, but in these cases the rules were invoked arbitrarily. Third, this pattern hindered the post-market phase for some products. New products may need extensive revision after launch, but instead they were often assumed to be ready for routine m~a~ement and evaluation (such as monthly cash flows). In one case, the new product facilitated the cure of cancer, and thus was not fully developed. It had been sold to several research labs and was generating modest revenue, however. Given its “market” status, management placed the product in a %ormal business” mode, forcing the director to focus on cash flows rather than on finishing development: My boss is in favor of this project but he is an entrepreneur in the business sense, not the technology sense. He looks at the project like a businessman would. Here I am in day 59 of my new business and we don’t even know if the technology works. But we worry about whether we will meet the profit objectives.. . Another culture clash is that they are applying management by objectives and quantifying me by the numbers on a monthly basis. His marketing director also described the failure of standard budgeting processes to take into account the emergent nature of their market:
68
THE JOURNAL OF HIGH TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. l/ 1995
We do the budgeting annually. In fact, we are now in the process of coming up with the numbers for next year. It is far too early for a start-up venture like this to develop a budget for next year. If we were on our own we would not do this until October, probably November, [7 months hence] because we don’t know where we are going to be in the market. When we started this last year the funding [Federal] for this approach to cancer treatment was on the rise so we made certain assumptions and forecasts. When the source [shifted its priorities] our market expectations had to be revised. Top management didn’t understand, and still holds us to those numbers.
Some innovators broke away from the pattern of detached judgement to monitor their efforts more realistically. In this example they relied on one another’s expertise and judgement to “calibrate” their knowledge of the product’s progress: I don’t see how we could have to be able to share it to run experiments in the guy at the plant needs
ever do a new product without everyone all together. You information between all of the functions.. . What good is the lab in test tubes when you really need to know what to produce in volume?. . . In the beginning when you don’t
know anything there is a long period where you rely strictly on baloney. By everyone working together and asking each other what they know, we get calibrated on where the venture is and what the potential is. The other way, when you do the product one step at a time, you spend $5 million and you don’t have anything.
His marketing counterpart described how their interactive perspective deviated from the rest of the division:
judgements
and longer term
It is important that even though this [new product organization] is set up under the auspices of corporate R&D, the manager also has responsibilities for sales, marketing, R&D, and manufacturing.. . We have many people in the group with a good mix of backgrounds to draw on. I know very little about making a sale, so I can go and ask the salesmen here about techniques. The boss is a territory manager and knows how to run new businesses.. . The rest of the division is set on the next quarter, so every time we go up to management we have to fight that, and prove that our products are economical and that the risk is manageable.
Those who got around the limits of detached judgement created an alternate pattern of thought and action which enabled them to develop useful heuristics to evaluate their progress. Staffing by Dislocated Responsibility Burns & Stalker (1966) argued that innovation requires a deeper commitment than routine work. The boundaries of responsibility must be broader and more inclusive in the rapidly changing, ambiguous conditions of innovation. Individuals need to see themselves as fully implicated in the discharge of the innovation task and as working together for the common purpose of the firm, so that they can attend fully to the “perpetual canvas of ideas and information outside the limits ordinarily set.. .” (Burns & Stalker, 1966, p. 89). When we considered how innovators or others understood their responsibility, we found that accountability and responsibility for innovations seemed
69
Bureaucracy and Product Innovation
dislocated and diffused, since no group of people legitimately held them. In a bureaucracy, accountability and responsibility are ideally located in an office, not an individual. Ironically, since innovation did not fit, responsibility for it became lodged outside the usual system of authority, in individuals rather than in offices. This dislocation meant that innovators could not rely on the established system to secure broader commitment or acquire resources. Instead, they had to rely on personal charisma and informal mechanisms. This pattern of thought and action affected innovation in three different ways. First, to take on a role appropriate to innovation as Burns & Stalker describe it requires that one step outside the bounds of legitimate authority. This not only made innovators uncomfortable, but also prompted others on whom innovators relied to refuse to cooperate. A planner who worked on a new plastic explained how people in his firm worked hard to avoid taking personal risk, and in so doing perhaps wasted far more money: People here who have a culture of “don’t screw up” don’t like to make mistakes. They have a culture where they pull in their horns. They’ll spend 14 to 16 million dollars a year with huge staffs on a new product and try to overwhelm any problems that come up with people. Another working on the same project described risk taking was simply not rewarded:
a widely shared attitude that individual
It is easier for people working here to ride the wave than to stick their necks out. If you do stick your neck out it will be either positive or negative. You will either succeed or fail. If you succeed, you are not much better off than if you hadn’t done anything. If you fail, you lose-you are out.
In another firm, a person was to innovate, described
who worked in the venture unit, and whose job officially how she still felt marginal and unwilling to take risks:
My job, in a way, is to try and thread through all of the constraints [that people in other departments put on a new product]. Sometimes I think I would like a job where you just do the same thing everyday. Here, you have to do something new and different every day. Since I have to work with people in other departments, I get the sense that every time I call them they are thinking: “Oh God! Now what does she want?!” Also, I will not develop a product idea that may cost a great deal but bomb. I don’t have the courage or the will to push an idea that may bomb.
Second, both successes and failures were seen as the result of individual action, not the result of normal organizational functioning. The innovators themselves were seen in heroic terms; as champions when they succeeded or, all too often, as evil incompetents when they failed. For example, managers in one firm attributed the success of a product to the ability of the project leader to guide the product through the organizational labyrinth. Two years later, a new CEO was hired to “straighten out the firm,” and he cancelled all innovations, including this one. The project leader was still seen as personally responsible, this time for the product’s failure. In another firm, people from several departments formed a “board of directors” for a venture, while a marketing
70
THE JOURNAL OF HIGH TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. l/ 1995
person was charged with day-today operations. When the product did not sell well everyone (except the marketing person) attributed failure to the marketing person’s lack of selling and interpersonal skills. They did not consider poor planning or organization as causes for failure. Third, dislocated accountability kept people from working effectively across both lateral and hierarchical lines, since jobs were understood in narrow, pre-set ways. This manufacturing manager described how established roles kept people from taking on the broad responsibility a problem required: When you run into a problem you say that it is someone else’s responsibility, so you don’t ask the next question. You wait for an answer instead. In a major corporation like this one you have professionals in the functions, and we are intimidated by knowing that there are better experts for certain problems.. . We had problems getting people to coordinate between engineering, research, and suppliers on the packaging because of some deference to other people’s expertise. They assumed it was another person’s job. It’s a turf issue, to not take another person’s job.
Relations with senior management were equally constrained by this dislocated view of one’s “proper” place, as in this description of how expectations precluded a frank discussion with senior managers: We view senior managers as adversaries so we hide our real issues from them. The idea is, let me prove that I know more than anyone else. We will not reveal or share our concerns since no one reveals or shares theirs, and we assume that management won’t like it. But I really ought to say: “Here are the milestones, here is the meter on the investment so far, and here are the three things that keep me awake at night.” It would be nice to take advantage of their experience, but we don’t.
In another example, a woman described a meeting between the development group and senior management. When a vice president asked an unanticipated question, she said everyone in the room “shut up.” She was the only one brave enough to answer “I don’t know,” she said, because getting fired did not scare her. We did not talk to these senior managers, and cannot say if they saw themselves as uncompromising ogres-we suspect not. However, people who have years of experience honestly felt that they could not openly discuss the uncertainty and ambiguity of innovation across vertical levels in their firms. People did not have the language or the framework to discuss innovation as it really is, and felt pressured to distort their experience by presenting it as routine to senior management. Some innovators overcame the dislocation of responsibility by setting up direct, personal relationships with senior managers and making public commitments to meet certain goals that had been negotiated and agreed to by all parties. Others overcame it by setting themselves apart in special “project groups” with broadened charges. By stepping out of the established pattern for responsibility, these innovators were able to create more collective commitment to the product without getting caught up in concerns over risk or over proper boundaries to the expectations that were placed on a job.
Bureaucracy
and Product Innovation
TOWARD A THEORY: HOW BUREAUCRACY
71
IS BAD FOR INNOVATION
The second research question required that we stand back from the descriptive findings and develop a conceptual frame to explain them. From all these data, our first inference is to agree with Weber, March & Simon, Bums & Stalker, and Schon-that an underlying worldview or interpretive system was at play, not only structure. The illustrations above indicate that people were aware that they (or others) approached problems with innovation in a “wrongheaded~ manner. They could articulate that their usual approaches to thinking and acting did not allow them to make sense of customers, react creatively to unexpected problems, determine if a project was going well, and staff the effort, even though they could not necessarily say why that was so. The trouble with innovation in our data was pervasive, intractable, and affected people’s very day-today action. Therefore, our first proposition for a theory of innovation in bureaucratic firms: PI:
Bureaucratic patterns of thinking and acting through which people make sense of customers and understand how to organize to develop a product inhibit their ability to innovate effectively. We predict that among organizations with similar structures, those in which these bureaucratic patterns do not prevail are more adept at product innovation. Similarly, organizations that are more adept at product innovation are less likely to have these non-innovation supporting patterns of thinking and acting.
This proposition does not deny that structure also affects a firm’s capability to innovate. It highlights the role of shared interpretations and understandings. We suggest that patterns of thinking and acting are at least as important as structure in understanding what makes an organization capable of effective product innovation. Exactly how, when, and under what conditions structural characteristics such as formalism, centralization, etc. affect the activities of innovation or reinforce a bureaucratic worldview is a matter for empirical research. Ident~ying these bureaucratic patterns of thinking and acting also remains an important research question. Our description of these patterns is robust across our data, we feel, but their final determination requires additional testing and clarification. However, by connecting the four bureaucratic patterns of thinking and acting directly to the activities of product innovation, we can explain the persistence and intractability of problems with innovation, and provide a common thread among them. These patterns are both consistent with and integrate a vast, disparate literature. They explain why innovators fail to do market analyses (Cooper & Kleinschmidt, 1986), or learn effectively (Van de Ven & Polley, 1992), or fail to break out of established technology “architectures”(Henderson & Clark, 1990), but do fall into incremental “strategic drift” (Johnson, 1988). To prompt research that will test and clarify our patterns: P2:
The four particular patterns described above are what hinder effective product innovation in bureaucracies: 1) defining a product from an inward, technology orientation; 2) organizing innovation steps and activities in a linearly progressive, incremental fashion;
12
THE JOURNAL
3) 4)
OF HIGH TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. l/ 1995
evaluating progress through criteria and rules of thumb that are detached from the specific reality; and staffing the effort by dislocating responsibility from the organization to the individual.
All else being equal, we predict that failed innovation efforts are more likely to be managed with these particular patterns of thinking and acting, white successful innovation efforts are more likely to be managed with alternate approaches. A third important question is where do these patterns come from? We have suggested that bureaucracy as an organizing form induces, maintains, and reinforces these kinds of practices rather than those more conducive to innovation. Indeed, at a general level, our four patterns of thought and action are consistent with the dysfunctions of bureaucracy described by Merton (1957), where procedural regulations become ends in themselves, and the instrumental, formalistic aspects of jobs become more impo~ant than substantive ones like good service to customers. However, our patterns connect such generic dysfunctions in a detailed way directly to product innovation. How bureaucracy engenders these patterns remains an important but unanswered question. March & Simon’s (1958) early work suggested that this kind of limited, segmented, abstracted approach to work would always occur, regardless of organizational form, because humans are inherently limited. However, Weber (1947), Burns & Stalker (1966), and Schon (1967) would argue that an interpretive system like instrumental rationality creates these patterns of thinking and acting. Instrumental rationality is a limited, narrow kind of rationality. The instrumentally rational organization raises precision, speed, unambiguity, routinization, and knowledge of the files to the optimum (Meyer 1990). Our four patterns are consistent with instrumental rationality’s focus on well-connected means/ends relationships and readily calculable activities, its privileging of the labelling, measuring, and evaluating of activities in the same way so they can all be compared, and its reliance on rational-legal authority, which presumes universal rules that extends to individuals only in so far as they occupy a legitimate office (Weber 1947; Blau & Meyer, 1987). Inward attention to technology, for example, is more amenable to “scienti~c” reason than are the shifting needs of customers, linear progression fits the methodical perspective of instrumental rationality, detached evaluation fits its universal rules, and dislocated responsibility fits with its segmentation of work. To challenge others to investigate these possibilities: P3:
When it comes to product innovation, an interpretive system of instrumental rationality is what is bad about bureaucracy. Instrumental rationality emphasizes conformity with rules, hierarchical control, and specialization. It orients people’s day-to-day sensemaking toward compliance, measuring activities precisely against standards, proceeding in a methodical, regimented manner, and simpl~ying complex problems into discrete parts. Concomitantly, instrumental rationality orients sensemaking away from the ability to evaluate ambiguous conditions, respond quickly to the unexpected, and notice possibilities not bounded by prevailing expectations. We predict that organizations operating with an instrumentally rational interpretive system or worldview will generate all four anti-innovative
73
Bureaucracy and Product Innovafion
patterns of thinking and acting, regardless of their structure.
and
innovate
less successfully,
Finally, drawing on the idea that a particular worldview causes the problems with innovation suggests a more complete theory of how to organize for innovation. The theory that instrumental rationality engenders anti-innovation behavior in large bureaucracies provides an explanation for why people who are not individually incompetent can nonetheless collectively work in a way which undermines the very goal they are working toward. Innovation is a complex and uncertain process. It is hard to make sense of new market opportunities or unfamiliar needs, because the information about them is unarticulated and ambiguous. It is hard to coordinate large numbers of people with diverse expertise. It is hard to evaluate the potential of unfamiliar ideas and to tailor that potential to the capabilities of the firm. And it is hard to define people’s roles and engender their commitment to such a task. Our analysis indicates that as innovators sought to innovate, they and the people around them invoked the only collectively shared patterns for thought and action they had-patterns derived from instrumental rationality to foster routinization and methodical control. The inherent difficulties of new product development were thus exacerbated at every turn by bureaucratic instrumental rationality. In addition, by bringing Weber and other early theorists back into the picture, we can develop a more complete theory of the innovative organization and suggest at the same time how large firms might start becoming more capable at innovation. If instrumental rationality creates the anti-innovative orientation, then perhaps substantive rationality would create an innovative one. Our data suggest that an orientation to ends rather than instrumental means, and more value-laden goals such as quality and customer satisfaction, helped innovators work on the ambiguous, hardto-measure activities of innovation. These innovators controlled the ambiguous task of innovation, not by chopping the work up into separate bits, but by framing the overall domain of action and then working more freely within that domain. Scholars such as Schon (1983) Perlmutter (1984), and Etzioni (1988) have described how substantive rationality is manifest in management, so that reflective, professional, and ethical practice can operate. In addition, Simon (1979) advocates the development of what he calls procedural rationality, which would enable people to deal with complex problems which lack information. However, we do not suggest that bureaucratic organizations should drop one narrow kind of rationality for another. Rather, we propose that they learn to broaden their rationality to include both kinds. Becoming more substantively rational would be easier for the large bureaucratic firms than becoming nonbureaucratic, we think. A more broadly rational system is still premised on “rationality,” which has meaning to managers. As Simon (1979) points out, the dictionary definition of “rationality” refers to being sensible and systematic, not only to measurement and precision. Burns & Stalker (1966) argued that most firms needed characteristics of both the mechanistic and organic form. Ritzer & LeMoyne (1990) also argue that Japanese management ideology combines instrumental and substantive rationality, and it is the combination which enables them to be both more productive and more innovative. In an interesting analysis of the NUMMI plant, Adler (1993) argues that a bureaucratic efficiency can be implemented in a way that enables learning, provided the social arrangements are
74
THE JOURNAL
OF HIGH TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. l/ 1995
premised on common goals, participation in defining key policies and standards, trust and respect, and a balance of power. These, too, are substantively rational elements of work organization. Weber (1947), however, argued that the two forms of rationality could not coexist, since instrumental rationality would drive out substantive concerns. Research on our last proposition would find out if Weber is still right (if ever)? Or has the world changed significantly? P4:
An interpretive system or worldview that combines instrumental rationality and substantive rationality would enable large, bureaucratic firms to be efficient and adept at innovation at the same time.
These propositions require additional study and analysis, especially the latter two. Moreover, labeling the anti-innovation force within bureaucracy is not enough for a practical theory. Research needs to explore the mechanisms by which instrumental rationality becomes so deeply and thoroughly embedded in order to develop a more complete theory about change. In conclusion, we suggest that the tenacious problems with effective innovation are deeply entrenched in the day-to-day culture of work and organizing in large bureaucracies. Becoming more adept with innovation requires that organizations alter their overall interpretive system along with their day-to-day practices. This suggestion is disturbing, since it contradicts the popular literature which presents the innovative organization as merely a matter of rearranging business units, adopting teams, or adding a venture unit. However, we think that by combining classical theory with modern ideas, we have developed some potentially useful reconceptions of bureaucracy’s relationship with innovation that may help resolve persistent problems both in theory and in the real world.
NOTE I.
Pairs of assistants coded subsets
disagreed
of interviews, with instructions to flag any problems over. They found only about 15 percent of the problems hard to code.
they
REFERENCES
Adler, P. (1993). The “learning bureaucracy”: New united motors manufacturing, inc. Research in organization behavior, 15(pp. 11 l-194). Greenwich, CT:JAI Press. Argyris, C., & Schon, D. (1978) Organizational learning: A theory of action approach. Reading, MA: Addison Wesley. Arthur D. Little (1991). World-wide survey on product innovation. Cambridge MA. Author. Bailyn, L. (1977). Research as a cognitive process: Implications for data analysis. Quality and Quantity, 11, 97-l 17. Berger, P., & Luckmann, T. (1967). 7%e social construction of reality. New York: Anchor Books. Blau, P., JL Meyer, M. (1987). Bureaucracy in modern society. New York: Random House. Burgelman, R. (1984) Managing the internal corporate venturing process, Sloan Management Review(Winter), 33-48.
Bureaucracy and Product Innovation
15
Burns, T., & Stalker, G. M. (1966). 7&e management of innovation (2nd ed.). London: Tavistock. Cohen, M., March, J., & Olsen, J. (1972). A garbage can model of organizational choice, Administrative Science Quarterly, 17, I-25. Cooper, R. (1983). A process model for industrial new product development. IEEE Transactions on Engineering Management, 30, 2-l 1. Cooper, R., & Kleinschmidt, E. (1986). An investigation into the new product process: steps, deficiencies, and impact. Journal of Product Innovation Management, 3,71-85. Daft, R. L., & Weick, K. E. (1984). Toward a model of organizations as interpretive systems, Academy of Management Review, 9,284-295. Day, D. (1994). Raising radicals: Different processes for championing innovative corporate ventures. Organization Science, 5, 148-172. Dougherty, D. (1992a). Interpretive barriers to successful product innovation in large firms. Organization Science, 3, 179-202. Dougherty, D. (1992b). A practice-centered model for organizational renewal through product innovation. Strategic Management Journal, 13,77-92. Dougherty, D., & Heller, T. (1994). The illegitimacy of product innovation in large firms. Organization Science, 5, 200-2 18. Dumaine, B. (1991, June 17). The bureaucracy busters. Fortune, pp. 36-50. Eisenhardt, K. (1989). Building theory from case study research. Academy of Management Review, 14, 532-550. Ettlie, J. E. (1988). Taking charge of manufacturing. San Francisco:Jossey-Bass. Etzioni, A. (1988). The moral dimension: Toward a new economics. New York: Free Press. Gagliardi, P. (Ed.). (1990) Symbols and artifacts: Views ofthe corporate landscape. Berlin: Walter de Gruyter. Galbraith, J. (1982) Designing the innovative organization, Organizational Dynamics, 10(3), 525. Hedberg, B. (1981). How organizations learn and unlearn. In P. Nystrom 8 W. Starbuck (Eds.), Handbook of organizational design(pp. l-27).. NY: Oxford University Press. Henderson, R., & Clark, K. (1990) Architectural innovation: The reconfiguration of existing product technologies and the failure of established firms. Administrative Science Quarterly, 35, 9-30. Johnson, G. (1988). Rethinking incrementalism. Strategic Management Journal, 9, 75-91. Kanter, R. (1983). 7&e changemusters. New York: Simon and Schuster. March, J., & Simon, H. (1958) Organizations. New York: John Wiley. Merton, R. (1957). Social theory and social structure. New York: Free Press. Meyer, M. (1990). The weberian tradition in organization research. In C. Calhoun, M. Meyer, & W.R. Scott (Eds.), Structures ofpower and constrainl(pp. 191-217). Cambridge, UK: Cambridge University Press. Mintzberg, II., & McHugh, A. (1985). Strategy formation in an adhocracy. Administrative Science Quarters, 30, 160-197. Mohr, L. (1982). Explaining organizational behavior. San Francisco, CA: Jossey-Bass. Nelson, R. R., & Winter, S. (1977). In search of a useful theory of innovation. Research Policy, 6, 36-76. Nord, W., & Tucker, S. (1987). Implementing routine and radical innovations. Lexington, MA: Lexington Books. Pearce, J., & Page, R. A. (1988). Palace politics: Resource allocation in radically innovative firms. High Technology Mun~gement Research, I, 193-206. Perlmutter, H. (1984). Building the symbiotic societal enterprise: A social architecture for the future. World Futures, 19, 271-284. Perrow, C. (1986). Complex organizations: A critical essay (3rd ed.). New York: Random House. Pugh, D., Hickman, D., Hinings, C., & Turner, C. (1969). The context of organization structures. Administrative Science Quarterly, 14,9 l-1 14.
76
THE JOURNAL OF HIGH TECHNOLOGY
MANAGEMENT
RESEARCH
Vol. ~/NO. l/ 1995
Quinn, J. B. (1985) Managing innovation: Controlled chaos. Harvard Business Review, 3, 7884. Ritzer, G., & LeMoyne, T. (1990). Hyperrationality: An extension of Weberian andneo- Weberian theory. Working Paper, University of Maryland. Rothwell, R. (1977). The characteristics of successful innovators and technically progressive firms. R&D Management, 7, 191-206. Schein, E. (1979). Personal change through interpersonal relationships. In W. Bennis, J. Van Maanen, E. Schein, & F. Steele (Eds.), Essays in interpersonal dynamics. Homewood, IL: Dorsey Press. Schein, E. (1985). Organizational culture and leadership. San Francisco, CA: Jossey-Bass. Schon, D. (1967). Technology and change. Oxford: Pergamon. Schon, D. (1983). 77re rejlexivepractitioner: Howprofessionals think in action. New York: Basic Books. Simon, H. (1979). Rational decision making in business organizations. i%e American Economic Review, 69, 493-5 13. Sims, H., & Gioia, D. (1986). The thinking organization. San Francisco, CA: Jossey-Bass. Smircich, L., & Stubbart, C. (1985). Strategic management in an enacted world. 7’he Academy of Management Review, 10, 724-737. Souder, W. (1987). Managing new product innovations. Lexington, MA: Lexington Press. Starbuck, W., & Milliken, F. (1988). Challenger: Fine-tuning the odds until something breaks. Journal of Management Studies, 25, 3 19-340. Strauss, A. (1987). Qualitative analysis for social scientists. New York: Cambridge University Press. Van de Ven, A. (1986). Central problems in the management ofinnovation. Management Science, 32, 590-608. Van de Ven, A., & Polley, D. (1992). Learning while innovating. Organization Science, 3, 92116. Von Hippel, E. (1986). Lead users: A source of novel product concepts. Management Science, 32, 791-805. Weber, M. (1946). From Max Weber: Essays in sociology. In H. H. Gerth & C. W. Mills (Eds., trans.). New York: Oxford University Press. Weber, M. (1947). The theory of social andeconomic organization. In A. Henderson & T. Parsons (trans.). New York: Free Press. Weick, K. (1979) fie sociology of organizing (2nd ed.). Reading, MA: Addison-Wesley. Yang, E.A., & Dougherty, D. (1993). Product innovation: More than just making a new product. Creativity and innovation management, 2, 137-155. Yin, R. (1989) Case study research (2nd ed.). Beverly Hills; Sage. Zirger, B.J., & Maidique, M. (1990) A model of new product development: An empirical test. Management Science, 36, 867-883.