Categorical evolution in nascent market spaces can be divided into two major phases. In an initial phase of divergence, an increasing number of categories are introduced and used; in the ensuing phase of convergence, a few categories start to win favor among market stakeholders, while many others are gradually abandoned.
Initial divergence: the creation of categories
At the beginning of their life cycles, new industries are seldom well defined, and so early entry occurs in a context of great uncertainty with regard to the meaning, boundaries, and even the very existence of the industry (Aldrich and Fiol, 1994; Lounsbury et al., 2003; Santos and Eisenhardt, 2009). Customer needs remain fluid (Clark, 1985), and so does the understanding of the industry by stakeholders—e.g., customers, producers, critics, and regulators (Grodal, 2007; Kennedy, 2008). During this time, the meaning of specific categories and the elements that belong to them are being negotiated (Granqvist et al., 2013). Because categories are still fuzzy, a given category might mean different things to different stakeholders (Vygotsky, 1987). Similar products might be positioned in different categories, whereas dissimilar products might claim membership to the same, still-fuzzy category. Rao (2008: p. 19) demonstrates these dynamics vividly in his description of the early automobile industry. As he notes, during the industry's very early phase, stakeholder confusion resulted in the automobile being “variously referred to as the ‘velocipede,’ ‘motorcycle,’ ‘locomobile,’ ‘electric runabout,’ ‘electric buggy,’ ‘horseless carriage,’ ‘automobile,’ and ‘quadricycle.’”
At such early stages of an industry, differences in categorical positioning often reflect basic technological differences among products or differences in how various stakeholders understand the new industry and its products (Kaplan and Tripsas, 2008). In the previous automobile example, for instance, “electric buggy” and “horseless carriage” are both examples of how underlying technological characteristics prompted the initial meaning of these two categories.
Alternatively, a novel categorical positioning can reflect the envisioned use of the product or correspond to deliberate attempts by firms to differentiate their products from their competition (Rao, 2008). The “Walkman,” a category introduced by Sony, did not explicitly reference technological characteristics, but rather the freedom of movement the new device allowed. By creating a new category, a firm can attempt to become the nascent market's cognitive referent or steer the whole market in a particular direction (Santos and Eisenhardt, 2009). For example, the “tablet PC” category was introduced by Bill Gates in 2002 in order to create a distinct identity for Microsoft's products and reignite the (at that time) still-fuzzy “pen computing” market.
Categories, however, can also be created by stakeholders other than firms in their attempt to understand and classify objects in an emerging market space (Kennedy and Fiss, 2013). Users, industry observers, analysts, or even visionaries can play an active role in introducing new categories and influencing the emergence of a collective conceptual schema of the industry. The “mountain bike” category, for example, was created by a bicycle aficionado, whereas a fiction writer coined the “robot” category.
The introduction of new categories by producers, users, and other stakeholders during the early stages of an industry results initially in a rapid increase in the number of categories in use. However, the fluid state of the various stakeholders' understandings implies that no one particular category is likely to gain much of a following. Uncertainty and confusion reign, and early entrants' attempts to predict which category and which understanding of the industry will eventually achieve dominance are prone to error.
Convergence on the dominant category
The initial phase of categorical divergence is generally followed by a phase of convergence during which the number of categories in use declines. As experience and familiarization with the new product class increase, stakeholders become better able to assess which of the product's features, uses, and connotations are relevant and important. Categories that best capture and express such relevant dimensions are preferred and gain traction over alternatives that are gradually abandoned (Kennedy, Lo, and Lounsbury, 2010).
Repeated use and increasing experience offer stakeholders greater certainty about the meaning of the different categories that remain in use—i.e., about the features that products belonging to any particular category are expected to possess (Hannan et al., 2007; Zerubavel, 1997). Categorical boundaries become better defined, and categories gradually assume meanings of their own; i.e., rather than being defined by the products that claim membership in them, they increasingly dictate the characteristics that a product needs to possess in order to be seen as a legitimate member. Subsequent product offerings are judged against these categorical requirements or “categorical test codes” (Hannan et al., 2007: p. 79): products that comply with the requirements of the category to which they claim membership are seen as legitimate members, whereas products that do not comply see their categorical claims questioned (Hannan et al., 2007). Stakeholders might even classify or reclassify a product irrespective of the producer's categorical claims if they find that it complies better with the requirements of another category (Bowker and Star, 2000).
The convergence of understandings and crystallization of categories in an emerging industry not only constrain the ability of producers and other stakeholders to introduce new categories, but also limit their need for doing so. Shared understandings about the meaning of categories and about product characteristics that are deemed relevant and valuable leave less room for the introduction of new categories. Moreover, firms may find it more attractive to position their products in a well-established category because doing so allows them to enter the consideration set of a large pool of customers (Zuckerman et al., 2003). Producers that initially positioned their products in some of the categories that failed to gain traction feel the pressure to reposition them, and producers that stick to categories that are faltering run the risk of being marginalized and seeing their market performance and survival chances suffer. As a result, the rate at which new categories are introduced to the industry declines, and the total number of categories decreases as some of them are abandoned.
Convergence toward a specific perception of the industry leads both producers and other stakeholders to gravitate toward the category that best captures this perception. The preferred category thus gradually gains dominance and becomes the referent for the emerging industry and its products. We define the dominant category as the conceptual schema that most stakeholders adhere to when referring to products that address similar needs and compete for the same market space. It follows that,
Proposition 1: The number of categories in a new market space has an inverted-U relationship with time.
The concept of the dominant category is related to, but fundamentally different from the established concept of the dominant technological design. The dominant category is a sociocognitive construct that is triggered primarily by the need of stakeholders to communicate meaningfully with other stakeholders regarding their activities in the emerging industry. Successful categories facilitate information exchange between disparate parties by conveying both the novelty of the category and its relationship to preexisting categories. Emerging categories thus have to resolve the paradox of being simultaneously (1) distinctive enough that they convey the novelty of the underlying product and attract the attention of stakeholders, and (2) familiar enough to be easily comprehensible (Bingham and Kahl, 2013).
This inherent tension between novelty and familiarity is often resolved by recombining existing categories in original ways (Berger and Heath, 2005). For example, the category “voice mail” combines the known element of mail with the novel aspect of the message not being delivered in a written form; likewise, the category “smartphone” invokes the familiar category of “phone” but simultaneously stresses the enhanced capabilities of the new product class by adding the qualifier “smart.” However, the success of a category in gaining traction will depend not only on its ability to manage the inherent tension between familiarity and novelty but also on its endorsement by prominent stakeholders (Negro, Koçak, and Hsu, 2011b) and its alignment with contemporaneous cultural trends (Bingham and Kahl, 2013; Wry et al., forthcoming).
In contrast to the dominant category, the dominant design is materially constituted and arises through technological experimentation (Suarez and Utterback, 1995), path dependence (Anderson and Tushman, 1990), and investments in process R&D and economies of scale (Klepper, 1997). Unlike the case of novel categories for which practically any recombination of elements of the cognitive space is possible, the materiality of technological designs limits the experimentation preceding a dominant design to what is technologically feasible. For each product class, there is only a limited set of possible technological trajectories that are feasible. Moreover, once a firm decides to invest and progress along a specific trajectory, technological “path dependence” imposes strong restrictions on what can be done and undone in the design. It is often technologically impossible to go back up the design hierarchy in order to reverse previous decisions (Clark, 1985). In contrast, while switching categories can incur significant costs for firms whose identities become associated with specific product categories (Tripsas, 2009), repositioning a product in a different category always remains a feasible option.
Combining our conceptualization of the dominant category with the industry life-cycle theory, we develop a theoretical framework that depicts the dynamics of a new market space (Figure 1). In line with received theory, we depict the number of firms (density) in the new space as increasing over time, reaching a peak, and then decreasing to stabilize at a relatively low level. The point at which the density curve peaks is often associated with the emergence of a dominant design (Utterback and Suarez, 1993) or, in platform-mediated industries, with the emergence of a dominant platform (Eisenmann et al., 2006; Suarez, Cusumano, and Kahl, 2013; Suarez and Cusumano, 2010).
Figure 1. Theoretical framework: dominant category and dominant design during the industry life cycle. During the industry life cycle, the number of categories will increase before the number of firms increases. The emergence of the dominant category occurs as the number of categories begins to decrease. This point in time marks the opening of the window of opportunity for entry, whereas the emergence of the dominant design marks the closing of the window of opportunity
Download figure to PowerPoint
We complement this accepted model of the industry life cycle by adding the dynamics associated with the emergence of the dominant category. The broken line in Figure 1 plots the number of categories in an emerging market space. During the early divergence phase, the number of categories increases as producers and other stakeholders introduce new categories (Santos and Eisenhardt, 2009). As the industry develops further, a process of convergence unfolds, in which some categories win favor and others recede (Glynn and Abzug, 2002). This process leads to the emergence of a dominant category to which the remaining and future products and producers generally adhere, resulting in a decline in the number of categories in the industry.
Time to the dominant category's emergence
As a collective conceptual schema, the dominant category emerges though iterative interactions among industry stakeholders. Initially disparate understandings or visions of the industry need to converge until a high degree of agreement is reached on the meaning of the industry. The duration of this convergence process that culminates in the emergence of the dominant category depends on the number of categories that populate the sociocognitive space, and their distances from one another.
In the first case, the use of a multitude of categories, many of which might be similar and overlapping, contributes to increased confusion among customers and other stakeholders. The use of different categories to refer to similar products makes communication more difficult and hinders understanding (Clark and Wilkes-Biggs, 1986). The emergence of shared notions about the industry is thus delayed, making it harder for any particular category to become dominant.
In the second case, the distance between competing categories implies the presence of different perceptions about the industry and the envisioned uses of its products. The larger the distance between two focal categories, the lower the conceptual overlap between them because they invoke meanings with few connotations in common (Kennedy et al., 2012). For example, at the initial stages of what is now known as the smartphone industry, producers introduced products for which they claimed such disparate categories as “handheld PDA,” “multimedia device,” “business tool,” “gaming device,” and “handheld computer.” The connotations associated with these categories were considerably different. The category “gaming device,” for instance, suggested that the device was to be used primarily for entertainment purposes. In contrast, positioning the device as a “handheld computer” invoked meanings of a sophisticated technology that could be used for a range of activities, including writing and reading documents and sending and receiving e-mail.
The presence of many categories, particularly if they are distant from one another, implies that a dominant category can only emerge through the gradual marginalization and abandonment of some of the competing perceptions about the industry and the corresponding categories. The more distant competing categories are from one another, the more exchange and debate is required for stakeholders to reach consensus on the core connotations of the industry, and the longer it takes for a dominant category to emerge. Therefore, we propose the following:
Proposition 2a: The time required for the dominant category to emerge in the early phases of an industry increases with the number of categories.
Proposition 2b: The time required for the dominant category to emerge in the early phases of an industry increases with the distance between categories.