In the bud? Disk array producers as a (possibly) emergent organizational form.

In 1986, a little-known Dutch company named Twincom introduced a software product designed to manage "disk drive arrays," which are data storage subsystems linking several (or many) hard disk drives. In the following year, disk array products were introduced by an additional seven companies: 1776, Atlantic Microsystems, Core International, Ford/Higgins, Maximum Strategy, Thinking Machines, and Micropolis, a disk drive manufacturer. Little over a decade later, disk arrays had a well-established world market and were widely used: over $12.6 billion in disk array products were sold in 1998 by 130 different producers. In a very different domain, namely beer, another new market was developing around the same time. In 1977, the New Albion Brewing Company opened in Sonoma, California. It joined the existing (but recently transformed) Anchor Brewing Company of San Francisco in offering heavier, full-flavored malt beverages (e.g., ales, porters, stouts) not found in the American market for beer. In subsequent years, others followed; by 2000, the economic contribution of mainly small "craft" breweries was estimated at $11 billion. The market includes scores of producers such as Anchor, commonly known as "microbreweries," as well as hundreds of other producers known as "brewpubs." The brewpub also makes full-flavored malt beverages using craft techniques but serves them at the site of production, usually in conjunction with food. Founded in 1987 in Hopland, California, the Hopland Brewery is widely recognized as the first brewpub in the U.S. since Prohibition.

Chances are that the microbrewery and brewpub terms will be familiar, while the disk array producer term (or anything similar that would describe specifically the firms in this market) will not be. Chances are also that it would not take much thought to name a microbrewery or brewpub, but it would take some research to name a disk array producer other than one of those listed above. The comparison thus raises two questions. First, does it matter that in the one market we have readily accessible descriptive labels to classify and distinguish participating firms while in the other we do not? Second, presuming it does matter, how can we account for the difference in public cognitive status of the various organizations?

Our answer to the first question is that having descriptive labels does matter. In our view, the ready accessibility of classificatory terms for types of organization derives directly from whether or not a particular type of organization constitutes an organizational form, defined as a recognizable pattern of activity, that takes on rule-like standing, which Polos, Hannan, and Carroll (2002) called a code. The term code here denotes and connotes both cognitive recognition and imperative standing. By this definition, a form is an external identity code, meaning that it is the perceptions and opinions of "outsiders" that matter. The external identity code possesses rule-like status, so that its observable violation is negatively sanctioned--it causes outsiders to drop discontinuously their valuation of the entity to which it is applied.

Our concept of organizational form implies legitimation or social-taken-for-grantedness, sometimes called constitutive legitimation. Much contemporary organizational theory treats legitimation as both privileging and constraining. An organization possessing a legitimated organizational form appears unproblematic and can be interacted with and regulated unambiguously; accordingly, it typically benefits from greater access to resources, more protection from authorities, and higher visibility--all provided that the organization does not violate any of the form-specific rules constraining its appearance and behavior (Meyer and Rowan, 1977; Zuckerman, 1999). We view establishment of legitimation as a process in which positive returns potentially increase from the first appearance of a potential organizational form up to a subsequent ceiling, signaling the organizational form's establishment. These positive returns of the form-establishment process involve ease of organizing, resulting in higher rates of organizational founding and enhanced life chances for the organizations using the potential form.

Accepting our answer to the first question means that the second question actually asks, When and where will a new organizational form emerge? In an exploratory case study of the disk array market, McKendrick and Carroll (2001) examined arguments drawn from organizational theory and juxtaposed them with basic facts of the situation. They found that the disk array organizational form had not developed, despite the presence of formal institutions representing collective action and ecological processes often associated with form emergence. They speculated that the reason the form had not crystallized lies in organizational diversity: the heterogeneous set of origin industries spawning and still supporting disk array producers (i.e., continuing to provide the bulk of many firms' revenue) makes it difficult for the disk array producer organizational form to gain perceptual recognition and take hold. The difficulty arises because form establishment is essentially about identity formation: if many firms in the market derive their primary identities from other activities and there are few firms deriving their primary identity from disk arrays, then the disk array producer identity will likely not be readily perceived by outsiders, thus impeding its coherence into a code or form.

McKendrick and Carroll's (2001) case materials suggest a reformulated specification, which we develop below, of the density-dependent process commonly thought to generate a legitimated organizational form. More precisely, an organizational form likely emanates from the density of producers with perceptually focused identities in a market, rather than the total density number typically used by ecologists, and the perceptions of outsiders will be more focused when the identities arise from (1) de novo entrants and (2) entrants that are concentrated in geographic locations that possess related identities. We used firm-level event-history data collected for every producer to enter the market for disk arrays worldwide, from Twincom's initial introduction to the end of 1998, to test these arguments in analyses of organizational entry and exit.

PRIOR THEORY AND RESEARCH ON FORMS