A MULTIDIMENSIONAL CONCEPTUALIZATION OF ENVIRONMENTAL VELOCITY

1. Academy of Management Review 2010, Vol. 35, No. 4, 604–626. A MULTIDIMENSIONAL CONCEPTUALIZATION OF ENVIRONMENTAL VELOCITY IAN P. MCCARTHY THOMAS B. LAWRENCE BRIAN WIXTED BRIAN R. GORDON Simon Fraser University Environmental velocity has emerged as an important concept but remains theoreti- cally underdeveloped, particularly with respect to its multidimensionality. In response, we develop a framework that examines the variations in velocity across multiple dimensions of the environment (homology) and the causal linkages between those velocities (coupling). We then propose four velocity regimes based on different patterns of homology and coupling and argue that the conditions of each regime have important implications for organizations. Environmental velocity1 has become an im- integration (Smith et al., 1994); rapid organiza- portant concept for characterizing the conditions tional adaptation and fast product innovation of organizational environments. Bourgeois and (Eisenhardt & Tabrizi, 1995); and the use of heu- Eisenhardt (1988) introduced this concept to the ristic reasoning processes (Oliver & Roos, 2005). management literature in their study of strate- More generally, research on velocity has shown gic decision making in the microcomputer in- that it affects how managers interpret their en- dustry. They described this industry as a “high- vironments (Nadkarni & Barr, 2008; Nadkarni & velocity environment”— one characterized by Narayanan, 2007a), further highlighting the ef- “rapid and discontinuous change in demand, fects of environmental dynamism on key orga- competitors, technology and/or regulation, such nizational members (Dess & Beard, 1984). that information is often inaccurate, unavail- A common feature of the treatment of environ- able, or obsolete” (Bourgeois & Eisenhardt, 1988: mental velocity in the literature has been the 816). From the perspective that the environment use of singular categorical descriptors to char- is a source of information that managers use to acterize industries—most typically as “low,” maintain or modify their organizations (Aldrich, “moderate,” or “high” velocity (e.g., Bourgeois & 1979, Scott, 1981), velocity has important impli- Eisenhardt, 1988; Eisenhardt, 1989; Eisenhardt cations for organizations. Studies have found, & Tabrizi, 1995; Judge & Miller, 1991; Nadkarni & for example, that success in high-velocity indus- Narayanan, 2007a,b). Although Bourgeois and tries is related to fast, formal strategic decision- Eisenhardt (1988) defined environmental veloc- making processes (Eisenhardt, 1989; Judge & ity in terms of change (rate and direction) in Miller, 1991); high levels of team and process multiple dimensions (demand, competitors, technology, and regulation), research on velocity has tended to overlook its multidimensional- We are grateful to associate editor Mason Carpenter and ity, instead assuming that a single velocity can three anonymous reviewers for their helpful and construc- be determined by aggregating the paces of tive comments. The development of this paper also benefited from comments from Joel Baum, Danny Breznitz, Sebastian change across all the dimensions of an organi- Fixson, Mark Freel, Rick Iverson, Danny Miller, Dave zation’s environment. This assumption over- Thomas, Andrew von Nordenflycht, Mark Wexler, Carsten looks the fact that environmental velocity is a Zimmermann, and seminar participants at Simon Fraser vector quantity jointly defined by two attributes University and the 2008 INFORMS Organization Science Pa- (the rate and the direction of change) and that per Development Workshop. We are also grateful to the Canadian Social Sciences and Humanities Research Coun- organizational environments are composed of cil for funding that supported this research. multiple dimensions, each of which may be as- 1 To increase the paper’s readability, we use the terms sociated with a distinct rate and direction of environmental velocity and velocity interchangeably. change. 604 Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s express written permission. Users may print, download, or email articles for individual use only.

2. 2010 McCarthy, Lawrence, Wixted, and Gordon 605 In this paper we aim to advance understand- studies characterizing velocity as simply high or ing of environmental velocity by developing a low. Specifically, we explain how the multidi- theoretical framework that articulates its multi- mensionality of velocity can affect the degree to dimensionality and by exploring the implica- which an organization’s activities will be en- tions of this framework for understanding the trained and adjusted over time. We then high- organization-environment relationship. We ar- light how these implications apply to two pro- gue that while there may be cases in which cesses that have been central to prior research organizational environments can be accurately on velocity: strategic decision making and new specified by a single descriptor (e.g., high veloc- product development. ity), a multidimensional conceptualization Our exclusive focus on environmental velocity opens up a number of opportunities. First, it differs from prior research that has sought to provides a basis for more fine-grained descrip- characterize organizational environments in tions of the patterns of change that occur in terms of a set of core properties—most com- organizational environments. An understanding monly some variation of complexity, dynamism, of a firm’s environmental velocity as composed and munificence (Aldrich, 1979; Dess & Beard, of multiple, distinct rates and directions of 1984; Scott, 1981). In pursuing this aim, we rec- change across multiple dimensions allows us to ognize the trade-offs among generalizability, ac- move beyond characterizations of industries as curacy, and simplicity (Blalock, 1982) inherent in high or low velocity and the assumption that all examining one aspect of the environment in dimensions change at similar rates and in sim- depth while bracketing other important environ- ilar directions (Bourgeois & Eisenhardt, 1988; mental dimensions. Research focused on the Eisenhardt, 1989; Judge & Miller, 1991; Smith et general organizational environment has strived al., 1994). Perhaps most important in this regard, for “high levels of simplicity and generalizabil- a multidimensional conceptualization allows ity, with a corresponding sacrifice of accuracy” for an examination of the relationships among (Dess & Rasheed, 1991: 703). This approach has the dimensions of velocity, which we argue can been characterized as “collapsing” the hetero- have a profound impact on organizations. geneity of the environment into a more parsimo- Second, a multidimensional conceptualiza- nious set of properties (Keats & Hitt, 1988). In tion of velocity offers a foundation for more con- contrast, we focus on a single specific aspect of sistent operationalizations of the construct, environmental dynamism—velocity—and ex- which would help improve the reliability and plore in detail its dimensions, how the velocities validity of research that employs it. Our review of these dimensions vary and interact, and the of the environmental velocity literature indi- consequences of those differences and interac- cates a reliance on singular descriptors of ve- tions. Our approach follows other studies that locity, which has led to inconsistent operation- have examined specific environmental con- alizations of the construct. Thus, while it has structs, such as uncertainty (Milliken, 1987) and sometimes been claimed that people can recog- munificence (Castrogiovanni, 1991). An impor- nize a high-velocity environment when they see tant consequence of focusing on a single aspect one (Judge & Miller, 1991), the different ways that of the environment is that any normative or pre- the velocity of the same industry has been cat- dictive claims we make must be made with ce- egorized by different researchers would seem to teris paribus restrictions placed on them. This, indicate otherwise. Such inconsistencies may be of course, complicates the application of such due to focusing on one or two particularly sa- claims in research or practice but also allows a lient velocity dimensions or to combining data deeper examination of specific phenomena (Pi- for multiple velocity dimensions without consid- etroski & Rey, 1995). ering the aggregation errors that can occur if the We present our arguments as follows. First, dimensions do not perfectly covary. we review the concept of environmental velocity Finally, by understanding that the environ- as it has been developed in management re- ments of organizations have multiple, distinct search, focusing on the opportunities that this velocities, it is possible to identify different pat- work presents for developing a multidimen- terns of environmental velocity whose condi- sional conceptualization. Second, we present tions affect organizations in ways that go be- our framework by defining several fundamental yond the insights that have emerged from dimensions of the organizational environment

3. 606 Academy of Management Review October and defining the key aspects of velocity—the dustry context, the level (high, moderate, or low) rate and direction of change—for each dimen- of velocity considered, and the measures em- sion. Third, we examine the potential relation- ployed (if any). Looking across these studies, we ships among velocity dimensions (such as prod- identify three themes that characterize much of ucts and technology) by introducing three the existing research in the area and provide the concepts: (1) “velocity homology,” which is the motivation for the theoretical framework that we degree to which velocity dimensions have sim- develop. ilar rates and directions of change at a point in First, existing studies have predominantly fo- time; (2) “velocity coupling,” which is the degree cused on high-velocity environments, with lim- to which the velocities of different dimensions ited attention to other potential patterns of ve- affect one another over time; and (3) “velocity locity. Consequently, we know relatively little, regimes,” which represent patterns of velocity for instance, about the velocity-related chal- homology and velocity coupling. Fourth, we ex- lenges faced by firms operating in low-velocity plore the implications of our framework for or- environments, where the slow pace of change ganization-environment relationships and for may be associated with protracted development strategic decision making and new product lead times, long decision horizons, and rela- development. tively infrequent feedback. Also, and more generally, the focus on high-velocity environments may be a significant factor in the treatment of ENVIRONMENTAL VELOCITY IN velocity in terms of singular categorical descrip- MANAGEMENT RESEARCH tors; the term high-velocity environment itself In physics, velocity refers to the rate of dis- seems to imply that multiple dimensions of the placement or movement of a body in a particular environment (e.g., products, markets, technol- direction. Thus, it is a vector quantity jointly ogy) combine nonproblematically to produce a defined by two distinct attributes: the rate of single, cumulative, high level of velocity. While change and the direction of change. The defini- this may be true in some cases, it is not clear tion of high-velocity environments articulated that it applies broadly across firms and by Bourgeois and Eisenhardt (1988) captured industries. these two attributes, referring to rapid and dis- Second, high-velocity environments are often continuous change in multiple dimensions of presented as synonymous with high-technology the environment, such as demand, competitors, industries, perhaps because Bourgeois and technology, and regulation. The notion of high Eisenhardt’s initial study focused on the early velocity provided an evocative way to charac- microcomputer industry. Industries have been terize the fast-moving, high-technology industry categorized as high velocity simply because that was the context of their studies, and it com- they are technology intensive (Smith et al., 1994) plemented a number of similar but conceptually or are built around an evolving scientific base distinct environmental constructs, including dy- (Eisenhardt & Tabrizi, 1995), regardless of namism (Baum & Wally, 2003; Dess & Beard, whether other environmental dimensions ex- 1984; Lawrence & Lorsch, 1967), turbulence (Em- hibit low or modest rates of change or relatively ery & Trist, 1965; Terreberry, 1968), and hypertur- continuous directions. Judge and Miller (1991), bulence (McCann & Selsky, 1984). More recently, for instance, identified the biotechnology indus- environmental velocity has been used in con- try as high velocity, despite its relatively long junction with or as a synonym for other related product development lead times and product environmental constructs, such as “clockspeed” life cycles (both ten to twenty years). (i.e., the speed of change in an industry; Fine, Finally, existing research tends to lack an ex- 1998; Nadkarni & Narayanan, 2007a,b) and hy- plicit measurement model or justification for the percompetition (Bogner & Barr, 2000; D’Aveni, categorization of specific organizational con- 1994). texts or industries. Instead, researchers declare Table 1 lists some of the major studies in stra- that they are studying high-velocity environ- tegic management and organization theory in ments and reiterate Bourgeois and Eisenhardt’s which the concept of environmental velocity (1988) original definition without significant ex- plays a central role. For each study the table planation or direct evidence (the studies by delineates the phenomenon of interest, the in- Judge and Miller [1991] and Nadkarni and Barr

4. 2010 McCarthy, Lawrence, Wixted, and Gordon 607 TABLE 1 Environmental Velocity in Management Research Management/Organization Level of Velocity Velocity Measures Example Studies Phenomena (Industry Context) Conceptualization of Velocity Used Bourgeois & Pace and style of strategic High (microcomputer Uniform change in the rate and Illustrative statistics Eisenhardt (1988) decision making industry) direction of demand, and examples competition, technology, and regulation Eisenhardt & Politics of strategic decision High (microcomputer As per Bourgeois & Eisenhardt Illustrative statistics Bourgeois (1988) making industry) (1988) and examples Eisenhardt (1989) Rapid strategic decision making High (microcomputer As per Bourgeois & Eisenhardt Illustrative statistics industry) (1988) and examples Judge & Miller Antecedents and outcomes of High (biotechnology), Aggregation of industry growth Industry data and (1991) decision speed medium (hospital), and perceived pace of survey data from and low (textile) technological, regulatory, firms and competitive change Smith et al. (1994) The effect of team demography High (informational, Rate of change in product, Illustrative statistics and team process electrical, demand, and competition biomedical, environmental) Eisenhardt & Rapid organizational adaptation High (computer) As per Bourgeois & Eisenhardt Illustrative statistics Tabrizi (1995) and fast product innovation (1988) and examples Brown & Continuous organization change High (computer) As per Bourgeois & Eisenhardt Illustrative statistics Eisenhardt (1997) (1988) and examples Stepanovich & Strategic decision-making High (health care) Rate of change in demand, An illustrative Uhrig (1999) practices competition, technology, and example regulations Bogner & Barr Cognitive and sensemaking High (IT) A form of hypercompetition None (2000) abilities Oliver & Roos Team-based decision making High (toys and IT Rate of change and the time None (2005) tools) available to make decisions Brauer & Schmidt Temporal development of a High and low A form of dynamism and Industry market (2006) firm’s strategy (industries not volatility returns data implementation specified) Davis & Shirato A firm’s propensity to launch High (computer), The number of product lines R&D expenditure/ (2007) World Trade Organization medium (auto), and and the rate of product total revenue actions low (steel) turnover Nadkarni & How cognitive construction by High (computers, Rate of change (clockspeed) for Industry clockspeeds Narayanan firms drives industry velocity toys) and low product, process, and organi- (2007a) (aircraft, steel) zational dimensions Nadkarni & Relationship between strategic High (computers, The rate of industry change Industry clockspeeds Narayanan schemas and strategic toys) and low (clockspeed) (2007b) flexibility (aircraft, steel) Nadkarni & Barr How velocity affects managerial High (semiconductor, As per Bourgeois & Eisenhardt A review of existing (2008) cognition, which in turn cosmetic) and low (1988) literature and affects the relationship (aircraft, matching using between industry context and petrochemical) industry attributes strategic action Davis, Eisenhardt, The performance and structural High and low The speed or rate at which new A Poisson & Bingham implications of velocity (conceptual opportunities emerge in the distribution of new (2009) simulation model) environment opportunities [2008] representing notable exceptions). This has equated velocity with the speed at which variation in the extent to which velocity has new opportunities emerge (Davis, Eisenhardt, & been operationalized has resulted in some coun- Bingham, 2009). terintuitive and inconsistent categorizations of Looking across these themes, we see that re- industry velocity. Studies of health care, for in- search on environmental velocity has provided stance, have labeled those environments as interesting and influential insights, particularly both high velocity (Stepanovich & Uhrig, 1999) into the nature of organizational processes op- and moderate velocity (Judge & Miller, 1991). erating in fast-changing, high-technology in- Furthermore, our understanding of velocity and dustries. We suggest, however, that the con- its effects across industry contexts has largely struct itself requires a more fine-grained focused on only one attribute of velocity—the examination, since existing research tends to rate of change—since prior research has tended assume that it can be adequately represented to use measures associated with the clockspeed by an aggregation of the rates of change across of an industry (e.g., Nadkarni & Narayanan, different environmental dimensions or by a fo- 2007a; Oliver & Roos, 2005; Smith et al., 1994) or cus on change in only one dimension of the

5. 608 Academy of Management Review October environment to the exclusion of others. In con- In order to describe the direction of change in trast, we believe that a multidimensional con- a way that allows comparison across industry ceptualization of velocity would provide a dimensions, we follow Bourgeois and Eisen- stronger foundation for clarifying and opera- hardt (1988), who suggest that the direction of tionalizing its characteristics and for better change varies in terms of its degree of continu- understanding its diversity and impacts on ity-discontinuity. They argue that continuous organizations. change represents an extension of past development (e.g., continuously faster computer technology), whereas discontinuous change repre- ENVIRONMENTAL VELOCITY AS A sents a shift in direction (the move from film to MULTIDIMENSIONAL CONCEPT digital photography, or the shifts that occur in The core understanding of environmental ve- fashion industries). Discontinuities, therefore, locity that we propose is that organizational en- can be represented by inflection points in the vironments are composed of multiple dimen- trajectories that describe change in a dimension sions, each of which is associated with its own over time (e.g., technology price-performance rate and direction of change. This simple notion, curves or demand curves for specific products). we argue, has profound effects on how we un- To more fully articulate a continuum of con- derstand and research velocity and on the or- tinuous-discontinuous change, we draw on ganizational reactions to velocity we expect and Wholey and Brittain’s (1989) three-part concep- prescribe. In this section we begin to construct tualization of environmental variation, arguing our theoretical framework, first by defining the that the direction of change is discontinuous to basic concepts of rate of change and direction of the extent that shifts in the trajectory of change change as they apply to the organizational en- are more recurrent, with greater amplitude and vironment in general, and then by describing with greater unpredictability over a period of how these basic concepts apply to some primary time. This approach helps us distinguish be- dimensions of the organizational environment. tween relatively regular, predictable (e.g., seasonal) variations in environmental velocity and irregular types of change that are more difficult The Rate and Direction of Change to predict and, consequently, more challenging Environmental velocity is a vector quantity in terms of organizational responses (Milliken, defined by the rate and direction of change ex- 1987). We suggest that such variations in the hibited by one or more dimensions of the orga- continuity-discontinuity of a velocity dimen- nizational environment over a specified period. sion’s trajectory allow for the use of structural The rate of change is the amount of change in a equation modeling (Kline, 2004) and difference dimension of the environment over a specified scores (Edwards, 1994) to produce growth models period of time, synonymous with such concepts that measure transitions in change over time as pace, speed, clock rate, or frequency of (Bliese, Chan, & Ployhart, 2007; Singer & Willett, change. The direction of change, while often 2003). mentioned in studies citing Bourgeois and Furthermore, to operationalize the rate and Eisenhardt’s (1988) definition, has attracted rel- direction of change of each velocity dimension, atively little attention beyond that. One possible we suggest that the measures will require scale reason for this is the relative difficulty of de- uniformity to allow the relative differences be- scribing the direction of environmental change. tween the dimensions to be compared and cor- Whereas the velocity of a physical object can be related (Downey, Hellriegel, & Slocum, 1975; Mil- described simply as moving eastward at 50 km/ liken, 1987). To achieve this, we suggest that the hr, similarly straightforward descriptions of the rate and direction of change will be some form direction of change of an organizational envi- of scalar measure (e.g., change/time). Therefore, ronment are not so obvious. This is particularly even though what is changing will vary for each the case when we consider the direction of of the dimensions, their relative rates and direc- change across different industry dimensions, tions of change can be determined and com- such as products, technology, and regulation, pared by using the same period of time for the the direction of each of which could be de- different dimensions (i.e., new products per year scribed in numerous distinct ways. and changes in product direction per year).

6. 2010 McCarthy, Lawrence, Wixted, and Gordon 609 Dimensions of Environmental Velocity duction processes and component technologies that underlie a specific industrial context, such The second way in which we break down the as float glass technology in glass manufactur- concept of environmental velocity is in terms of ing, genetic engineering in the biotechnology the dimensions of the organizational environ- industry, and rolling mills in metals processing. ment that are changing. While the dimensions of the environment that are salient for any par- See Table 2 for a summary of the definitions for ticular study will vary according to the specifics each of the velocity dimensions on which we of the research project, there are several that focus. have been widely used in prior research on or- The rate of technological change is the ganizational environments. We use the four di- amount of change in those technologies over a mensions suggested by Bourgeois and Eisen- specific time period, including the creation of hardt (1988)— demand, competitors, technology, new technologies, the refinement of existing and regulation—and to this list we add a fifth technologies, and the recombination of compo- dimension—products. We do this because prior nent technologies. The rate of technological research on environmental velocity has tended change varies dramatically across industries. to merge the technology and product dimen- Drawing on patents as an indicator of the rate of sions, and we argue that they often have dissim- technological change, one can argue, for in- ilar rates and directions of change, which makes stance, that the electronics industry exhibits a separating them important for our purposes. more rapid rate of technological change than Archibugi and Pianta (1996) point to the impor- does the oil industry. In 2006, rankings for the tance of this distinction when they argue that number of patents granted in the United States product changes need not be technical but can showed that the top five positions were held by also include changes in the aesthetic, branding, electronics companies, whereas the oil industry or pricing features of a product. Our discussion firms Shell and Exxon occupied positions 126 of environmental dimensions is not meant to be and 139, respectively (IFI, 2008). Although some exhaustive; rather, it is meant to highlight the technological change is either not patentable or heterogeneity of environmental dimensions that not patented for strategic reasons, the rate of motivates our development of a multidimen- patenting can nevertheless provide a useful in- sional conceptualization of velocity. dication of the technological rate of change Technological velocity. Technological velocity since it is a relatively direct and publicly avail- is the rate and direction of change in the pro- able indicator of the proprietary technological TABLE 2 Environmental Velocity: Dimension Definitions and Example Measures Definition/Example Measures Technological Product Demand Regulatory Competitive Velocity dimension The rate and direction The rate and direction The rate and direction The rate and direction The rate and direction definition of change in the of change in new of change in the of change in laws of change in the production product willingness and and regulations that structure of processes and introductions and ability of the market affect an industry competition within component product to pay for goods an industry technologies that enhancements and services underlie a specific industrial context Example measures The number of new The number of new The change in The number of new The change in of the rate of patents and products introduced industry sales in a and amended laws industry population change in the copyrights granted in a given period given period and/or regulations size and density dimension in a given period (i.e., product introduced in a (i.e., number and clockspeed) given period size of firms) in a given period Example measures The changes in the The change in the The change in the The change in the The change in of the direction direction of the nature of product trend (e.g., growth nature and scope of industry growth of change in the relationship features as versus decline) and the control provided trends (e.g., growth dimension between the price perceived by the nature (e.g., by new laws and versus decline) in a and technical market in a given personal versus regulations in a given period performance of period impersonal) of given period technology in a demand in a given given period period

7. 610 Academy of Management Review October outputs of an industry (Archibugi & Pianta, 1996; tion points with respect to price (in 1981 and Griliches, 1990). 1988) and no major inflection points with respect The direction of technological change refers to to performance. In contrast, fashion products, the trajectories along which technological ad- such as clothing, music, and travel, all change vancements take place (Abernathy & Clark, frequently through the creation of new products 1985; Dosi, 1982; Tushman & Anderson, 1986). Dis- and the transformation and repackaging of ex- tinguishing between continuous and discontin- isting ones. Such variations in product change uous directions of technological change is most across industries are associated with differ- easily understood in terms of performance/price ences in the complexity, risk, and impact of the curves. Continuous technological change in- product change. volves a series of improvements that enhance While velocity research has often lumped to- the performance of the technology (e.g., ad- gether product and technological velocities, our vances in photographic film technology focused definitions of their rates and directions of on improving contrast quality, light sensitivity, change illustrate the importance of distinguish- and speed). Such changes move a technology ing between them. Over the past several de- smoothly along a performance/price curve, usu- cades, for example, the underlying materials ally at a decreasing rate, thus creating a con- and production processes in the automobile cave downward performance/price curve. In industry have changed more rapidly and dis- contrast, discontinuous technological change continuously than have the end products them- involves “architectural” (Henderson & Clark, selves. In contrast, textile production technolo- 1990) or “radical” innovations that “dramatically gies have changed more slowly and advance an industry’s price vs. performance continuously than the fashion products they are frontier” (Anderson & Tushman, 1990: 604). These used to create. innovations temporarily alter the shape of the Demand velocity. Demand velocity is the rate performance/price curve such that it becomes and direction of change of the willingness and concave upward until the immediate benefits of ability of the market to pay for goods or services, the innovation are exhausted. including changes in the number and types of Product velocity. This dimension is the rate transactions and market segments. The rate of and direction of new product introductions and change in demand varies tremendously across product enhancements. We define products as industries, with some experiencing rapid any combination of ideas, services, and goods growth or decline and others facing steady offered to the market (Kotler, 1984). The rate of growth for years. Such variance is influenced by change in products can vary tremendously a wide range of factors, including changes in across industries and across market segments taste, new rival products, substitutes, comple- within an industry. In terms of the former, Fine ments, changes in relative prices, business cy- (1998) and Nadkarni and Narayanan (2007a,b) cle fluctuations, and switching costs. Empirical show that the movie, toy, and athletic footwear research has used summary industry sales fig- industries have relatively high rates of product ures as an indicator of the rate of change in change (new products launched every three to demand (e.g., Bourgeois & Eisenhardt, 1988). six months), whereas the aircraft, petrochemi- The direction of change for demand is contin- cal, and paper industries have low rates of prod- uous when there is a steady progression of in- uct change (new products launched every ten to creasing or decreasing sales to a consistent set twenty years). of consumers. Conversely, change in the direc- The direction of change for products can be tion of demand is discontinuous when there are described as continuous when new product infrequent, significant, unpredictable shifts in the troductions represent improvements on previ- growth, decline, or steady state of demand, or a ously important product attributes, and discon- radical change in the segments that compose tinuous when the new products introduce the overall market. For example, demand veloc- fundamentally new attributes for consumer ity in the U.S. restaurant industry from 1970 to choice. Adner and Levinthal’s (2001) study of the 1995 was relatively continuous, with sales gains personal computer industry between 1974 and made nearly every year during that period (Har- 1998 provides an example of relatively continu- rington, 2001). In contrast, the demand for com- ous product change, with only two major inflec- modities, such as copper and gold, can be

8. 2010 McCarthy, Lawrence, Wixted, and Gordon 611 highly volatile owing to a wide range of macro- ment (Bowman & Gatignon, 1995). Such mea- economic influences, exemplifying the case of a sures describe the overall pace at which the discontinuous demand velocity. Similarly, the competitive conditions that define an industry Nintendo Corporation created discontinuous are changing—a factor that has been shown to change in the demographics of demand since its influence firm performance across a wide range Wii games console appealed to nontraditional of industries, including the automotive (Hannan, market segments, such as families, women, and Carroll, Dundon, & Torres, 1995), computer (Hen- older people. derson, 1999), and insurance (Ranger-Moore, Regulatory velocity. We define regulatory ve- 1997) industries. locity as the rate and direction of change in the The direction of change in competitive struc- regulations and/or laws that directly affect the ture involves continuity-discontinuity with re- firm or industry under consideration. This in- spect to the value chain in an industry (Jaco- cludes government action (e.g., changes in laws, bides & Winter, 2005), the nature of rivals (Porter, regulations, and polices) and industry self- 1980; Schumpeter, 1950), or changes in market regulation (e.g., voluntary standards and codes). contestability (Hatten & Hatten, 1987). Change in It is a dimension that can open or close markets, competitive structure is continuous to the de- present organizations with compliance costs, gree that these characteristics remain constant and necessitate strategic shifts in practices. The and stable over time. Conversely, the change in rate of regulatory change is a function of the direction in competitive structure is discontinu- creation of new laws or regulations, or changes ous to the degree that industry value chains are to existing laws or regulations, in a time period. in flux (Jacobides, 2005) and existing bases of It can vary greatly across industrial, national, competition are challenged by firms introducing and historical contexts, and it often depends on new products, pioneering new markets or other factors, such as technology (e.g., regula- sources of supply, or implementing new means tions for stem cell research), business scandals of production (Schumpeter, 1950). (e.g., the Enron scandal), health and safety issues (e.g., mad cow disease), and demographic RELATIONSHIPS AMONG VELOCITY shifts (e.g., an increase in the retired DIMENSIONS: VELOCITY HOMOLOGY, population). VELOCITY COUPLING, AND The direction of change in regulation is con- VELOCITY REGIMES tinuous to the degree that new regulations re- semble the old in scope, form, or substantive An important benefit of a multidimensional areas of concern, and it is discontinuous to the conceptualization of environmental velocity is degree that they address new issues, focus on the potential it provides to examine the differ- different kinds of behaviors, or employ new prin- ences and relationships among the velocities of ciples. For example, the U.S. airline industry different dimensions. To that end, we introduce from 1938 to 1975 experienced changes in regu- three concepts: (1) velocity homology—the rela- lations that were relatively continuous, in that tive similarity among the rates and directions of the Civil Aeronautics Board (CAB) restricted change of different dimensions; (2) velocity cou- prices, flight frequency, and flight capacity pling—the degree to which the velocities of dif- (Vietor, 1990). Then, in 1975, the direction of reg- ferent dimensions are causally connected; and ulatory change changed as the CAB began ex- (3) velocity regimes—the different patterns of perimenting with limited deregulation, and in environmental velocity that emerge from varia- 1978 the industry was completely deregulated tions in velocity homology and velocity and the CAB abolished. coupling. Competitive velocity. Competitive velocity is the rate and direction of change in the structural Velocity Homology determinants of industry profitability (Barney, 1986; Porter, 1980). Its rate of change is, in part, a The term homology was coined by the paleon- function of the entrance and exit of industry tologist Richard Owen (1843) to explain the mor- rivals (Hannan & Carroll, 1992), as well as the phological similarities among organisms. It has speed with which firms respond to competitors’ been used by management scholars to describe strategic moves or other shifts in the environ- the degree to which two phenomena are similar

9. 612 Academy of Management Review October (Chen, Bliese, & Mathieu, 2005; Glick, 1985; Han- Shook, 1996), factor analysis (Segars & Grover, lon, 2004) and is consistent with the homogene- 1993), and multidimensional scaling (Cox & Cox, ity-heterogeneity aspect of environmental com- 2001), all of which are considered suitable for plexity (Aldrich, 1979; Dess & Beard, 1984). In our assessing interdimension similarity in construct framework, velocity homology is the degree to composition (Harrison & Klein, 2007; Law, Wong, which the rates and directions of change of dif- & Mobley, 1998). ferent dimensions are similar to each other over An assumption of a highly homologous set of a period of time. Thus, “high homology” de- velocities typified much of the early work on scribes a condition in which the velocities of high-velocity environments, in which industries different dimensions in a given environment ex- such as microcomputers were characterized by hibit relatively similar rates and directions of “rapid and discontinuous change” across multi- change, whereas “low homology” describes rel- ple dimensions (Bourgeois & Eisenhardt, 1988: atively dissimilar rates and directions of 816). An assumption of high homology carried change. over to subsequent studies, with limited consid- To help explain velocity homology, we present eration of the degree to which homology might a map of the velocities of different dimensions, vary across firms and industries. Most studies with the rate of change and the direction of seem to have aggregated the velocities of differ- change on each axis (see Figure 1). With this ent dimensions, regardless of the variance image of velocity (based on the fashion apparel among these dimensions, thereby assuming industry example we present in the following similarity (i.e., high homology in our terms) sections), homology is represented by the close- among the velocities of different environmental ness of the points. Thus, low homology (as is the dimensions. Consequently, we know relatively case in Figure 1) is represented by relatively little about the conditions and effects of low- spread out points, and high homology would be homology environments, where the velocity represented by relatively tightly clustered properties of a firm’s multiple environmental di- points. To operationalize this concept of homol- mensions are highly dissimilar. ogy, we suggest using distance measures and To illustrate and clarify the concept of homol- methods, such as cluster analysis (Ketchen & ogy, we present the example of the apparel in- FIGURE 1 Fashion Apparel Industry Example Discontinuous Demand Competitive Product Direction of change Technological Regulatory Continuous Low Rate of High change Key: The solid lines indicate tight coupling and the dashed lines loose coupling.

10. 2010 McCarthy, Lawrence, Wixted, and Gordon 613 dustry and focus on the industry segment in- over the past twenty or so years—toward greater volved in the design and supply of seasonal automation and efficiency in textile manufac- fashion apparel. This includes brands sold pri- turing, more rapid response to customer de- marily through own-brand stores (e.g., Gap, mands, and more efficient communication and Zara, and American Apparel) and brands sold coordination in fashion design and retailing through a mixture of own-brand stores and in- (Doeringer & Crean, 2006; The Economist, 2005). dependent stores (e.g., Armani, Benetton, and In contrast to product and technological veloc- Levi’s). We chose this industry because aca- ities, regulatory change in this industry has, for demic studies and business reports suggest that the past two decades, occurred relatively slowly from 1985 through 2005 the velocities of different and continuously. The regulation that affects dimensions in this industry spanned a diverse this industry most significantly is directed at the range of rates and directions of change (Djelic & manufacture of clothing and the protection of Ainamo, 1999; The Economist, 2005; Jacobides & consumer rights, both of which have changed Billinger, 2006; Taplin & Winterton, 1995). slowly over that period. With respect to the man- Beginning with the product dimension, this ufacture of garments, the Multi Fibre Arrange- segment of fashion retailing is associated with ment (MFA) was introduced in 1974 as a short- a relatively high rate of change and a moder- term measure to govern world trade in textiles ately discontinuous direction. This is illustrated and garments, imposing quotas on the amount by the operations of Zara, one of Europe’s lead- developing countries could export to developed ing fashion brands. Zara launches some 11,000 countries (Spinanger, 1999). This regulation un- new products annually, most of which are com- derwent only minor modifications until it ex- pletely new products as perceived by the cus- pired in 2005 (Audet & Safadi, 2004). National- tomer and typically take only five weeks from level regulation tends to focus on labor and design to retail store (The Economist, 2005). Even employment standards. In response to the shift casual fashion houses, such as Sweden’s of clothing manufacturing from developed to Hennes & Mauritz (H&M) and the American emerging economies, the governments of West- chain Gap, roll out between 2,000 and 4,000 prod- ern nations have been reluctant to further regu- ucts each year. Moreover, the rate of change in late (and potentially stifle) clothing manufactur- products has increased, with the emergence of ing, much of which occurs as home-based work “fast fashion” as a dominant strategy for mass (Ng, 2007). market designers/retailers (Doeringer & Crean, Change in demand for fashion apparel has, 2006). We argue that the direction of product for the past twenty years, occurred moderately change is moderately discontinuous, because slowly, with a high degree of discontinuity. Re- although these firms launch many new prod- searchers argue that the fashion industry is ucts, they represent a mix of new items and characterized by low to moderate levels of pos- extensions of existing products. This view is itive sales growth each year (Nueno & Quelch, consistent with studies of the rate and direction 1998), with occasional major demographic and of change in women’s formal wear (Lowe & lifestyle shifts and changes in customer prefer- Lowe, 1990). ences (Danneels, 2003; Siggelkow, 2001). Al- The technologies that underpin the fashion though the direction of change in demand for industry have been changing rapidly over the fashion has oscillated between relative stability past twenty years (cf. Richardson, 1996) but at a and discontinuity over the last 150 years (Djelic relatively slower rate than changes in fashion & Ainamo, 1999), the past 20-year period has products. Although manufacturing technology been associated with customers becoming more in the apparel industry has remained stable for demanding, arbitrary, and heterogeneous nearly a century (Audet & Safadi, 2004), there (Djelic & Ainamo, 1999; The Economist, 2005). have been advances in the manufacture of tex- The competitive velocity of the fashion indus- tiles, as well as in communication and informa- try has long fascinated observers. In recent tion technologies, that have facilitated the move years it has altered as increased cost pressures to quick response (Forza & Vinelli, 1997) and fast have led firms to engage in rapid-fire attempts fashion (Doeringer & Crean, 2006) strategies in to source the lowest-cost materials and to move fashion design and retailing. The direction of labor-intensive aspects of the value chain to these changes has been relatively continuous countries with lower costs. The industry has also

11. 614 Academy of Management Review October experienced constant shifts in the major centers change in the velocity of one dimension causes of production (Dosi, Freeman, & Fabiani, 1994). a change in the velocity of another. By way of example, U.S. employment levels in Weick (1976) defined loosely coupled systems this sector in 2002 were a third of what they were as those in which the properties of constitutive in the early 1980s (Doeringer & Crean, 2006). The elements are relatively independent, whereas intersection of cost pressures and the increasing the properties of elements in tightly coupled sys- rate of change in consumer preference and de- tems are strongly mutually dependent. Weick mand has led to significant shifts in firms’ strat- (1982) further argued that loose coupling in- egies, particularly speeding up the supply chain volves causal effects that are relatively periodic, (Richardson, 1996) and altering organizational occasional, and negligible, whereas tight cou- structures and boundaries (Djelic & Ainamo pling involves relatively continuous, constant, 1999; Jacobides & Billinger, 2006; Siggelkow, and significant causal effects. Thus, we de- 2001). Such conditions characterize change that scribe the velocities of different dimensions of a is both moderately rapid and continuous in firm’s environment as loosely coupled when nature. changes in the velocity of one dimension (e.g., The fashion industry points to two important technology velocity) have relatively little imme- issues with respect to understanding homology diate, direct impact on the velocities of other among environmental velocity dimensions. dimensions (e.g., product velocity), and we de- First, it highlights that the organizational envi- scribe them as tightly coupled when the rela- ronment is composed of a number of distinct tionship between the velocities of different di- dimensions, each of which is defined by its own mensions involve significant immediate, direct rate and direction of change— or velocity. Sec- causal effects. To determine the degree of cou- ond, we see that there are significant differ- pling between velocity dimensions, we suggest ences in the rates and directions of change (low using structural equation modeling (Kline, 2004), homology) across the five dimensions that we which is recommended for operationalizing co- have considered. This makes the idea of de- variance between construct variables (Law et scribing the industry as having a single veloc- al., 1998). ity, whether based on an “average” across di- Although coupling and homology both de- mensions or on the velocity of whichever scribe the relationships among velocity dimen- dimension might be considered most important, sions, they are separate, distinguishable as- misleading both to researchers attempting to pects of those relationships. The velocities understand the industry and to managers need- of different dimensions can have high levels of ing to make strategic decisions. interdependence (coupling), regardless of whether they exhibit similar rates and directions of change (homology). Homology is a first- order property of velocity, describing the simi- Velocity Coupling larity among velocities over a period of time. In A second important aspect of the relationship contrast, coupling is a second-order property, between velocity dimensions is the degree to describing the degree to which changes in the which and the ways in which they interact over velocity of a dimension affect the velocity of time. We examine these interactions through the another dimension over the same specified pe- concept of coupling. This is the degree to which riod of time. The distinction between homology elements of a system, including product compo- and coupling is observable in the biotechnology nents (Baldwin & Clark, 1997; Sanchez & Ma- industry, which experiences high rates and dis- honey, 1996), individuals (DiTomaso, 2001), or- continuous directions of technological change ganizational subunits (Meyer & Rowan, 1977; but relatively slow, continuous regulatory and Weick, 1976, 1982), and organizations (Afuah, product velocities (Zollo, Reuer, & Singh, 2002). 2001; Brusoni, Prencipe, & Pavitt, 2001), are caus- While these dimensions have very different ve- ally linked to each other (Orton & Weick, 1990; locities (low homology), there is evidence to sug- Weick, 1976). In our framework velocity coupling gest that they are relatively tightly coupled. This is the degree to which the velocities of different is illustrated by the impacts of the 2001 U.S. dimensions in an organizational environment regulation on stem cell research, which re- are causally connected—the degree to which a stricted research to twenty-one stem cell lines (a

12. 2010 McCarthy, Lawrence, Wixted, and Gordon 615 family of constantly dividing cells) and, in turn, in their study of the U.S. fashion apparel indus- limited the rate and direction of U.S. stem cell try in the 1980s, Abernathy, Dunlop, Hammond, research activity (i.e., technological velocity) rel- and Weil (1999) explain how changes in demand ative to other countries. In 2009 this regulation led to “lean retailing,” which, in turn, required was overturned, permitting research on up to firms to drastically alter their information and 1,000 new stem cell lines, allowing “U.S. human production technologies to enable new working embryonic stem-cell research to thrive at last” practices. In contrast, there is little evidence to (Hayden, 2009: 130). suggest that changes in the velocity of technol- We again draw on the fashion apparel indus- ogy for the fashion industry will affect or are try to illustrate the idea of coupling among ve- affected by changes in the velocities of compe- locity dimensions. Beginning with products, tition or regulation. changes in the velocity of this dimension have In this illustrative example (see Figure 1), we been attributed to increases in the adoption of argue that seven of ten possible dyadic connec- new communications, design, and manufactur- tions among velocity dimensions are relatively ing technologies, suggesting a relatively tight tightly coupled (designated by solid lines) such coupling between product and technological ve- that changes in the velocity of one dimension locity dimensions. Perhaps most significant, will affect the velocity of another. We have ar- changes in the direction of technology have im- gued that the three other connections are proved the ability of fashion apparel firms to loosely coupled, as indicated by the dotted gather market feedback and, thus, to develop lines. Thus, although not all of the velocity di- new product offerings at a faster rate (Jacobides mensions of the fashion industry exhibit strong & Billinger, 2006; Kraut, Steinfield, Chan, Butler, causal connections to each other, we suggest & Hoag, 1999; Richardson, 1996). Similarly, the that this industry can be described as a rela- velocity of demand has been tightly coupled to tively tightly coupled environment. Any assign- product velocity over the past two decades: in- ment of such a category is somewhat arbitrary dustry observers argue that the perceived new without a formal measurement of coupling, so arbitrariness of customer demand has forced for now we follow work on modular (loosely cou- fashion organizations to frequently engage in pled) and integrated (tightly coupled) organiza- large-scale market explorations (Cammet, 2006; tional forms that suggests that when at least 50 Jacobides & Billinger, 2006). In contrast, there is percent of the system elements are tightly cou- little evidence of a strong relationship between pled to each other, the system can be considered product velocity and competitive velocity. Prod- tightly coupled (Schilling & Steensma, 2001). uct velocity appears to be primarily driven by changes in market demand and the product innovation programs of existing organizations ex- Velocity Regimes ploiting those changes, as opposed to a flow of new entrants (Cammet, 2006). We propose the concept of a velocity regime In terms of the velocity of regulation in this as a way to describe the pattern of velocity ho- industry, there is evidence that it is tightly cou- mology and velocity coupling within an organi- pled to the velocities of competition, demand, zational environment. Although both these char- and products, with changes in international acteristics of velocity vary continuously, we trade regulations (Spinanger, 1999) and domes- focus on combinations of high or low homology tic labor standards (Ng, 2007) leading to increas- and tight or loose coupling to more clearly illus- ing imports from developing economies, both trate how they vary and the effects of these creating and satisfying the demand for cheaper variations. The result is a typology (see Figure 2) fashion products. Similarly, the velocities of with four distinct velocity regimes that repre- competition and demand appear to be tightly sent ideal types, rather than an exhaustive tax- coupled, with firms in this industry attempting onomy of velocity conditions. To illustrate and to predict and adapt to what Siggelkow (2001) visualize the degrees of homology and coupling calls “fit-destroying changes” that can signifi- that characterize each regime, we have embed- cantly alter their competitive positions. There is ded a variation of Figure 1 into each cell of also tight coupling between the velocity of tech- Figure 2. Like Figure 1, these embedded figures nology and the velocity of demand. For example, present illustrative sets of velocities, the rela-

13. 616 Academy of Management Review October FIGURE 2 Environmental Velocity Regimes Conﬂicted velocity regime Integrated velocity regime R D Tight R D Direction Direction C C of change of change T P T P Velocity Rate of change Rate of change coupling Divergent velocity regime Simple velocity regime R D Direction R D Direction of change C of change C Loose T P T P Rate of change Rate of change Low Velocity High homology Key: T technological velocity, R regulatory velocity, D demand velocity, C competitive velocity, and P product velocity. The solid lines indicate tight coupling and the dashed lines loose coupling. tive positions of which indicate their rates and An example of a simple velocity regime is the directions of change for different dimensions. U.K. tableware industry from the mid 1950s to The first velocity regime in our typology oc- the late 1970s. During this period, this industry curs when environmental dimensions are highly was exposed to changes in regulations, de- homologous and loosely coupled to each other. mand, product, technology, and competition that We call this the “simple velocity regime” be- were all relatively slow and continuous in na- cause it has similar rates and directions of ture (Imrie, 1989; Rowley, 1992). At the same time, change across all dimensions. Thus, regardless this industry had relatively loose coupling of whether these dimensions are all changing among velocity dimensions. For example, when slowly and continuously or rapidly and discon- change did occur in the velocity of the product tinuously, we argue that it is the relative unifor- dimension during the 1970s, due to an increase mity of the change in strategic information that in the rate at which product variety and customi- makes the environment relatively analyzable zation changed, the only other velocity dimen- (Daft & Weick, 1984). Furthermore, because the sion to be affected was technology, whereby velocities of the multiple dimensions are loosely changes in the flexibility of production machin- coupled, they are free to vary independently so ery altered at a similar rate (Carroll, Cooke, that changes in the velocity of one dimension Hassard, & Marchington, 2002; Day, Burnett, For- are unlikely to affect the velocities of other rester, & Hassard, 2000). This combination of dimensions. high homology and loosely coupled dimension

14. 2010 McCarthy, Lawrence, Wixted, and Gordon 617 velocities created an environment that analysts products were developed, which, in turn, af- and scholars described as being uniformly sta- fected the rate at which new market segments ble, consistent, and regular in nature (Imrie, were created (Bresnahan & Malerba, 1999; Lan- 1989). glois, 1990). This coupling among dimensions The second environmental velocity regime in also brought about the wholesale change in the our typology occurs when the velocities of dif- velocities that occurred around 1995 as the in- ferent dimensions are highly homologous and dustry began its fourth era—the age of the net- tightly coupled. This creates what we call an work (Malerba et al., 1999). “integrated velocity regime.” This regime is in- The third velocity regime, which we call the tegrated in two senses: the velocity attributes “divergent velocity regime,” has a set of dissim- of each dimension (i.e., rates and directions of ilar and loosely coupled velocities, so firms face change) are very similar, and the velocities of diverse and possibly contradictory environmen- the dimensions are highly interdependent on tal conditions. This potentially makes the envi- each other for a period of time. The tight cou- ronment more difficult to analyze, because some pling differentiates this regime from the simple dimensions change slowly and continuously— regime, presenting managers with the complex generating modest amounts of information— task of monitoring and responding to causally while other dimensions change rapidly and dis- connected changes in a velocity. This is what continuously—producing large quantities of Aldrich (1979: 77) calls the “everything’s related information that quickly becomes inaccurate or syndrome,” where a change in the velocity of obsolete. This set of dissimilar velocities pre- one dimension reverberates throughout the ve- sents diverse temporal demands on the informa- locities of other dimensions. Together, these tion processing and sensemaking abilities of conditions create an environment that is best managers. The relatively loose coupling among understood as having, at least for a time, a sin- these dissimilar velocities, however, somewhat gle overarching velocity. Moreover, if all the di- lessens the challenge of monitoring and re- mensions are changing rapidly and discontinu- sponding to environmental conditions, because ously, this situation will be exemplified by the changes in the velocities of different dimensions “high-velocity” industries that have dominated are relatively independent, limiting the poten- research on environmental velocity. tial for rapid, widespread change in the flows of Consequently, an example of an integrated strategic information. velocity regime is the global computer industry An example industry of this regime would be from approximately 1982 to 1995. During this pe- the U.S. flat glass manufacturing industry from riod, which is known as the third era of the 1955 to 1975. During this period, the environmen- industry, the microprocessor and personal com- tal dimensions for this industry had very differ- puter were invented (Malerba, Nelson, Orsenigo, ent and unconnected velocities. The technology— & Winter, 1999), and most of the environmental float glass production methods—that was dimensions were changing rapidly and in a dis- developed to produce flat glass was adopted continuous direction. Firms were frequently en- relatively quickly during this period compared tering and exiting the industry, as well as form- to other process technology innovations (Teece, ing and breaking alliances with each other 2000). It was also a discontinuous change that (Bresnahan & Malerba, 1999; Langlois, 1990). revolutionized how flat glass was made, with Technological substitution in hardware and productivity gains approaching 300 percent as software was a frequent occurrence, resulting in the need for grinding the glass was eliminated regular product innovations (Bourgeois & Eisen- (Anderson & Tushman, 1990). This led to signifi- hardt, 1988; Brown & Eisenhardt, 1997). While cant price/performance improvements so that Eisenhardt and colleagues clearly argued that float glass products replaced existing flat glass such conditions equated to multiple velocities products in a relatively rapid and continuous undergoing similar “rapid and discontinuous fashion, rising from 30 million square feet per change,” we suggest there was also a signifi- year of glass in 1960 to 1,730 million square feet cant level of interdependence among the veloc- per year of glass in 1973 (Bethke, 1973). Because ities of these dimensions. For example, studies this change in demand was generated by exist- have explained how the velocity of competition ing producers for existing automotive and con- affected the rate at which new technologies and struction customers, the pace and direction of

15. 618 Academy of Management Review October competitive change remained relatively slow the implications of velocity homology and veloc- and continuous in nature. The only significant ity coupling in terms of their general impacts on regulatory event for this industry was that the organizing and on the processes of strategic de- U.S. Tariff Commission and Treasury more fre- cision making and new product development. quently cited foreign producers for dumping flat glass on the U.S. market at prices lower than Implications of Velocity Homology those in their own markets (Bethke, 1973). This link between the rate of government action and We argue that the notion of velocity homol- the increase in production capacity from the ogy significantly affects how we need to think new technology appears to be the only major about the relationship between an organization interdependency between the different veloci- and the temporal characteristics of its environ- ties of the dimensions for this industry during ment. The dominant notion that has emerged this period. over the past two decades in the velocity litera- The final velocity regime we propose is com- ture, and more broadly in research on time and posed of dimensions whose velocities are rela- organizations, has been the importance of orga- tively dissimilar and tightly coupled. We call nizations operating “in time” with their environ- this the “conflicted velocity regime,” since orga- ments and in synchrony across their subunits nizations operating with such a regime will ex- and activities. This is the view of research on perience diverse and potentially contradictory organizational “entrainment” (Ancona & Chong, velocities that are also highly interdependent. 1996; McGrath, Kelly, & Machatka, 1984; Perez- ´ As in the case of the divergent regime, the low Nordtvedt, Payne, Short, & Kedia, 2008), which level of homology among velocity dimensions in argues that “functional groups not only must be the conflicted velocity regime leads to condi- [internally] entrained with each other for the tions that are, as a whole, inconsistent and rel- organization to work, there must also be exter- atively unanalyzable. However, the tight cou- nal entrainment, at both the subsystem and sys- pling among these heterogeneous velocities tem levels, to ensure adaptation to the environ- increases the difficulty associated with track- ment” (Ancona & Chong, 1996: 19). The impact of ing, understanding, and responding to changes external entrainment on performance is echoed in the conditions of this regime, because the in research on high-velocity industries, which causal variation makes the environment rela- argues that organizational performance in such tively unstable over time. Although neglected in environments is associated with rapid decision the velocity literature, we believe that this kind making (Eisenhardt, 1989) and fast new product of velocity regime may be quite common. Our development (Eisenhardt & Tabrizi, 1995; example of the fashion industry since the mid Schoonhoven, Eisenhardt, & Lyman, 1990). In 1980s illustrates the dynamics associated with their discussion of “timepacing,” Eisenhardt and the conflicted velocity regime. We argued that Brown (1998) provide examples of the impor- the rates and direction of change in this industry tance of external entrainment, including the span a diverse range. We further argued that household goods manufacturer that timed its this industry’s environmental dimensions are product launch cycles to key retailers’ shelf relatively tightly coupled. Such conditions de- planning cycles and, thus, was able to win more fine an environment with a set of dimensions shelf space. that are not only changing dissimilarly but are Our multidimensional conceptualization of also highly interdependent. velocity suggests that temporal alignment between an organization’s operations and its environment is critically important but that varia- ORGANIZATIONAL AND tions in homology create significant limits to the STRATEGIC IMPLICATIONS synchronization of activities within firms (inter- The importance of environmental velocity is nal entrainment). If the velocities associated due to the impacts it has on key organizational with different environmental dimensions are and strategic processes. Thus, in this section we similar, as in our high-homology regimes (sim- examine how a multidimensional conceptual- ple and integrated), then it is appropriate to ization of environmental velocity would affect entrain the pace and direction of all organiza- our understanding of these impacts. We explore tional activities to this uniform environmental

16. 2010 McCarthy, Lawrence, Wixted, and Gordon 619 velocity. This will be a relatively simple situa- A second key strategic process that illustrates tion to manage. However, if the dimension ve- the implications of velocity homology is new locities differ significantly, as in our low- product development—the set of activities that homology regimes (conflicted and divergent), transforms ideas, needs, and opportunities into then the situation will be more difficult to man- new marketable products (Cooper, 1990). Previ- age. This is because the task of entraining or- ous research has shown the value of rapid new ganizational activities with dissimilar dimen- product development in high-velocity industries sion velocities will lead to heterogeneous sets of (Eisenhardt & Tabrizi, 1995) but leaves open the paces and directions of activities within firms. question of how this might change if we incor- Such differences create challenges for firms, in- porated a multidimensional conception of envi- cluding potential incoherence among subunits ronmental velocity. Although new product de- and activities, fragmented internal information velopment processes may seem to be primarily flows, and the breakdown of issue capture and linked to the product dimension of the organiza- analysis across intraorganizational boundaries. tional environment, they cut across a wide Furthermore, managers who understand that range of organizational functions, including re- changes in velocity homology conditions can be search, development, design, manufacturing, le- both endogenous and exogenous in nature will gal, marketing, and sales. Consequently, each have not only the option of reactively entraining of these different new product development ac- their organizations to their environment but also tivities collects, interprets, and applies relevant the option of trying to alter the speed and direc- information from different dimensions of the or- tion of change in specific environmental dimen- ganization’s environment. Thus, the contribution sions to suit their organization. Firms might, for of each function to new product development is example, lobby to influence the rate at and di- likely to be more effective when that function is rection in which legislators develop laws and entrained with the environmental dimension for regulations (i.e., shape what is regulated/ which it is more directly responsible. The ability deregulated in an industry and the pace at of marketing, for instance, to effectively contrib- which regulatory reform occurs), or undertake ute to the development of new products depends marketing activities to influence changes in on its being entrained with the velocity of de- demand. mand. This means that different new product A central theme of research on environmental development functions may need to operate at velocity has been its effect on strategic decision different speeds and in different directions in making—those “infrequent decisions made by order to ensure process-environment entrain- the top leaders of an organization that critically ment. Again, this can potentially create signifi- affect organizational health and survival” cant organizational challenges in terms of coor- (Eisenhardt & Zbaracki, 1992: 17). Following our dination and integration across the stages of the general argument regarding the impact of ve- new product development process. locity homology, we argue that variations in homology reward strategic decision-making activ- Implications of Velocity Coupling ities that are individually entrained with the velocity of their relevant environmental dimen- We argue that the notion of velocity coupling sion. Thus, more effective strategic decision significantly affects how we think about the sta- making in high-homology regimes (simple and bility of velocity conditions and impacts how integrated) will involve a set of activities with organizations coordinate changes in the pace similar paces and directions. Such internal con- and direction of their internal activities. Previ- sistency will provide benefits in terms of greater ous research has tended to treat environmental efficiency and lowered task conflict (Gherardi & velocity not only as a unidimensional concept Strati, 1988). In contrast, strategic decision mak- but as a relatively stable feature of organiza- ing in low-homology regimes (conflicted and di- tional environments. In contrast, we argue that vergent) will be more effective when the pace variations in velocity coupling will lead to im- and direction of strategic decision-making ac- portant differences in the stability of the velocity tivities are dissimilar, because they are tailored conditions of environments. For firms operating to their relevant but distinct dimension in tightly coupled environments, a change in the velocities. velocity of any one dimension (e.g., technology)

17. 620 Academy of Management Review October will have a broad impact on the velocity condi- tively planned and shifted the velocities of their tions of the regime, through its effects on the research advocacy units to better link with the velocities of the other dimensions to which it is activities of patient advocacy groups. These coupled (e.g., products, demand, competition). changes helped the industry to garner the pub- This suggests that regimes with tight velocity lic support necessary to overturn regulations coupling (integrated and conflicted) will have (Campbell, 2009). relatively unstable velocity conditions. This ar- Achieving this sequenced change in the pace gument follows research on coupling in both and direction of organizational activities would organizational environments and organizations involve the use of time-based mechanisms. that has shown that tight coupling among ele- These include scheduling and project deadlines, ments of a system increases the instability of information technologies that align organization- that system (Aldrich, 1979; Dess & Beard, 1984; al activities, and resource allocation rules that Terreberry, 1968). An important facet of this in- specify the time to be spent on decision tasks stability is the rhythms through which it occurs. (McGrath, 1991). The impacts of changes in the velocity of one As with velocity homology, changes in veloc- dimension on the velocities of other dimensions ity coupling may stem from external conditions, are unlikely to occur instantaneously but, or it may be that managers are able to increase rather, over time, as the social and technologi- or decrease the causal connections among ve- cal mechanisms that connect the dimensions locity dimensions in order to create strategic are sequentially triggered and exert their advantage for their firms. One strategy to affect impact. velocity coupling is to alter the degree of mod- We argue that the environmental instability ularity in products (Baldwin & Clark, 1997; and sequencing of changes associated with Sanchez & Mahoney, 1996), technologies (Yaya- tight coupling provide an advantage to certain varam & Ahuja, 2008), organizations (Meyer & firms over others. In particular, tightly coupled Rowan, 1977; Weick, 1976, 1982), or interorgani- regimes (integrated and conflicted) will reward zational networks and supply chains (Afuah, firms that employ mechanisms that sensitize 2001; Brusoni et al., 2001). Such changes can af- them to velocity changes and allow them to rap- fect the overall coupling among environmental idly and effectively shift the paces of their inter- dimensions, particularly if they establish new nal operations. Typical mechanisms could in- competitive standards. Furthermore, such clude strategic scanning systems that managers changes can be hard to attain and therefore use to monitor and respond to changes in their difficult to imitate, thus creating a competitive environments (Aguilar, 1967; Daft & Weick, 1984) advantage. Shimano, for example, became the and “interactive control systems” (Simons, 1994) dominant supplier of bicycle drive train compo- to promote external reflection and internal com- nents (shifters, chains, derailleurs, etc.) by de- munication and action. These mechanisms are veloping high-performing, tightly coupled com- analogous to other traditional organizational in- ponent systems that changed the nature of the tegration (Lawrence & Lorsch, 1967) and bound- new product development and production func- ary-spanning (Galbraith, 1973) mechanisms, but tions for their customers, as well as the nature of with a focus on coordinating change in the pace end-user demand. Shimano’s strategy altered and direction of organizational activities to the pace and direction of multiple velocity di- match temporal instability in the environment. mensions for the bicycle industry and has been Moreover, sequenced changes in velocities credited with helping Shimano gain almost 90 provide an advantage to firms that recognize percent of the drive train market for mountain these causal connections and are consequently bicycles (Fixson & Park, 2008). able to anticipate sequences of velocity The effects of velocity coupling on how orga- changes. For example, increases in human ge- nizations coordinate their activities can also be netic engineering technology in the late 1990s illustrated by considering strategic decision led geneticists and government agencies to call making and new product development pro- for more regulation to control the development cesses. For strategic decision making, coordina- and application of this technology. Those firms tion is an issue of social cognition within top that anticipated the connection between techno- management teams (Forbes & Milliken, 1999), logical velocity and regulatory velocity proac- which we argue is significantly affected by the

18. 2010 McCarthy, Lawrence, Wixted, and Gordon 621 “temporal orientation” of a team. A temporal with changes in their paces and directions hav- orientation is a cognitive concept that describes ing limited impacts on each other. how individuals and teams conceive of time—as In contrast, “recursive” new product develop- “monochronic,” a unified phenomenon that mo- ment frameworks conceive of the process as a tivates attention to individual events in serial system of interconnected, overlapping activities fashion, or as “polychronic,” a heterogeneous that generate iterative and nonlinear behaviors phenomenon that necessitates simultaneous at- over time (McCarthy et al., 2006). These include tention to multiple events (Ancona, Okhuysen, & Kline and Rosenberg’s (1986) chain-linked model Perlow, 2001; Bluedorn & Denhardt, 1988; Hall, and Eisenhardt and Tabrizi’s (1995) experiential 1959). We argue that strategic decision making model, both of which, we argue, are suited to in tightly coupled regimes would benefit from a tightly coupled velocity regimes because they polychronic orientation on the part of top man- facilitate improvisation and flexibility. These agement teams so that team members share a capabilities help managers of the process to view of time as malleable and unstructured. focus on and accommodate both the greater in- This would help them to simultaneously coordi- stability and more turbulent information flows nate strategic decision-making velocities and to associated with these velocity regimes. pay continuous partial attention to a broad set of issues (Stone, 2007). In contrast, in loosely CONCLUSION coupled regimes the benefits of multitasking, monitoring, and simultaneously adjusting to the In the paper’s introduction we suggested that velocities of different dimensions are lower. a multidimensional conceptualization of environmental velocity presented three important Such situations, we argue, reward a mono- opportunities to advance research in the area. chronic temporal orientation that leads senior First, we argued that it would allow a more management teams to engage in strategic deci- fine-grained examination of environmental ve- sion making in a relatively independent man- locity so as to better understand the diversity of ner, focusing on one issue at a time. this construct across different organizational For new product development processes, the contexts. In our discussions of several indus- impact of velocity coupling rests on the ability of tries, including fashion, tableware, computers, firms to recognize and predict the conditions and flat glass, we have shown that characteriz- under which a new product will be launched. ing these environments simply as high or low The instability associated with tightly coupled velocity overlooks the fact that environmental regimes (integrated and conflicted) influences velocity is composed of multiple dimensions, the effectiveness of different process control each with a distinct velocity. frameworks that help ensure that the right type Second, we argued that a multidimensional of product innovation is launched at the right approach to velocity could lead to more reliable time (McCarthy, Tsinopoulos, Allen, & Rose- and, thus, more valid empirical research by of- Anderssen, 2006). “Linear” new product develop- fering a basis for more consistent operation- ment frameworks conceive of the process as a alizations of velocity. Consequently, with our series of relatively discrete, sequential stages, framework we have urged researchers to con- with team members at each stage making deci- sider both the rate and direction of change for sions (go forward, kill the project, put the project multiple pertinent dimensions of the organiza- on hold, etc.) about the progress and outputs of tional environment. This reveals homology and the process (McCarthy et al., 2006). These frame- coupling relationships among the velocity di- works include the waterfall model (Royce, 1970) mensions, which describe the different velocity and the stage-gate method (Cooper, 1990), which regimes we propose. These concepts provide a assume and impose structures or “scaffolds” basis to better specify environmental velocity that restrict the amount of iterative feedback. and use appropriate operationalizations to mea- We argue that such linear frameworks are best sure its diversity. This, in turn, helps avoid in- suited to new product development processes appropriate aggregations and inconsistent uses that operate in loosely coupled velocity regimes of the velocity construct. in which the activities within the new product Third, we suggested that a multidimensional development process are relatively discrete, conceptualization of environmental velocity and

A MULTIDIMENSIONAL CONCEPTUALIZATION OF ENVIRONMENTAL VELOCITY

You are reading a preview.

Check these out next

A MULTIDIMENSIONAL CONCEPTUALIZATION OF ENVIRONMENTAL VELOCITY

More Related Content

Viewers also liked (20)

Similar to A MULTIDIMENSIONAL CONCEPTUALIZATION OF ENVIRONMENTAL VELOCITY (20)

More from Ian McCarthy (20)

Recently uploaded (20)