Strategy Formation in an The widely accepted view equates strategy making with Adhocracy planning, assuming that strategies are "formulated" be- fore they are "implemented." Based on the detailed track- ing overtime of the actions of a single project organization, Henry Mintzberg strongly resemblant of an ideal type called "adhocracy," and this paper shows that strategies can "form" in a variety of Alexandra McHugh different ways: from the precedents set by individual oper- ators, from thin streams of activity that eventually pervade an organization, from spontaneous convergence in the behavior of a variety of actors, and so on. The paper first identifies and tracks the strategies pursued by this orga- nization across almost four decades of its history and then presents the history of the organization in terms of strategic periods. These findings are then interpreted in conceptual terms, focusing on three themes*, the emergent nature of the organization's strategies and the difficulties of identifying intention in a collective context; the cycles of behavior that resulted from attempts to reconcile the con- current needs for convergence and divergence; and the organization's quest for adhocracy and the problems this posed forthe exercise of formal leadership. The paper concludes with a "grass roots" model of strategy formation.* "One best way" thinking has pervaded the field of manage- ment since Frederick W. Taylor (1947) coined the term early in this century. It has been particularly InfluerMial in two spheres — the designing of organizational structures and the making of organizational strategies. At least until recently, the underlying assumptions of organiza- tional design have been that organizations require articulated objectives, sharp divisions of labor, clearly defined tasks, well- developed hierarchies, and formalized systems of control. In fact, this configuration of elements — close to the ideal-type, machine-like bureaucracy Weber (1958) first described — appears to remain the predominant conception among practi- tioners in government, mass production, and the consulting profession: to many of them, "machine bureaucracy" is not just one alternate form of structure, it is structure. Strategy making still tends to be equated with planning — with the systematic "formulation" and articulation of deliberate, premeditated strategies, which are then "implemented." The traditionai precepts of "making strategy explicit" (Tilles, 1963), formulating before implementing (Andrews, 1980), and design- ing structure to "follow" strategy (Chandler, 1962) remain intact in almost all of the literature. This view of strategy making, however, is unnecessarily restrictive; it is inconsistent with more contemporary forms of structure and sometimes with the conventional forms as well. © 1985byComenUnrversity. One important contemporary form is project structure, or 0001-8392/85/3002-0160/$1.00. "adhocracy," (Bennis and Slater. 1964; Toffler, 1970). As described by Mintzberg (1979), this configuration includes the The authors are deeply indebted to William following elements: (11 The organization operates in an en- Litwack, who initiated tfie cdfection of the data and established the basic framework vironment that is both dynamic and complex, demanding in- in which it was analyzed, and to the Social novation of a fairly sophisticated nature. Each output tends to Science and Humanrties Research Coun- cit, which funded the study. be unique (e.g., a film, the prototype for a new product). (2) The 160/>WministrativeScienceQuarteHy,30(19K): 160-197 production of complex, unique outputs forces the organization to engage highly trained experts and to combine their talents in multidisciplinary teams. (3) These experts are housed m spe- cialized units, for administrative and housekeeping purposes, but are deptoyed in temporar/ teams to work on their projects; the structure thus takes on the form of a matrix. (4) Because of the complex and unpredictable nature of its work, the organiza- tion relies largely on mutual adjustment for coordination, which is encouraged by semiformal structural parameters such as liaison personnel and standing committees. Coordination by direct supervision and standardization are discouraged, as are the more formalized aspects of structure that support them, such as hierarchy, performance controls, and rules. And (5) the organization is decentralized "selectively"; power over differ- ent decisions is diffused in uneven ways, subject to the availability of information and expertise needed to deal with the issue at hand. While adhocracy is here characterized as an ideal type, most or all of these elements have, in fact, been described in a number of empirical studies — in Woodward's (1965) description of the process producers,"" Burns and Stalker's (1966) description of electronics firms, Galbraith's (1973) study of the Boeing Com- pany, Lawrence and Lorsch's (1967) study of plastics firms, and in Chandler and Sayles' (1971) study of NASA. In fact, consistent with Stinchcombe's (1965) proposition that struc- tures tend to reflect the age of founding of their industry, all of these organizations function in industries that developed only well into this century. Adhocracy seems to be the structure of the industries of our age. This paper presents an in-depth description of strategy making in one project organization, to accomplish three objectives: to develop the concept of strategy, to present a very different view of how it can form {without necessarily being formulated) in an organization, and to associate this view especially, but not exclusively, with the configuration we are calling adhocracy. RESEARCH ON STRATEGY FORMATION This study is part of a research project that began in 1971 with the intention of opening up the strategy-making process con- ceptually. Based on the definition of strategy as a pattern in a stream of decisions or actions,^ it has sought to track ar>d study strategies as they form and disappear in organizations While Woodward (1965) equates process over long periods of time. production with such a configuration, be- cause in her view the automation of ^e If strategy is defined only with regard to intention, the re- operating core renders the administrative competent of an organization a project searcher is reduced to studying perceptions, devoid of be- structure, there is evidence that process havior. Defining strategy with respect to realization, however, producers with large, nonprofessional enables the researcher to track the rise and fall of strategies in woric forces (such as steel connpantes) tend to resemWe nriachine bureaucracies empirical terms. Moreover, the most interesting issues seem (seeMintzberg, 1979: 257-^261.264-266. to revolve around the interplay of these two definitions so that, 458-45^. as shown in Figure 1, deliberate strategies (intentions realized) can be distinguished from emergent strategies (patterns real- ized despite or in the absence of intentions). The original definition (Mint^wrg, 1972) included only dectsic«is; for a number of The definition of strategy as pattern in action may appear reasons, not tf>e least being that decisions, too, represent intentions and that we controversial: the term is almost always defined formally with necessarily ended up studying actions, the regard to intention. We would argue, however, that the defini- definition was recerrtiy modified to include tion "in use" is often based on observed consistency in actions. This point is ^abwated upon m a note available from the first author. behavior, that is, pattern in action, as when a manager infers Strategy Formation Deliberate Strategy Intended \ y/^ Realized Strategy \ / Strategy Unreali»d Emergent Strategy Strategy Rgura 1. Basic forms of strategy. the "strategy" of a competitor (to which he or she might, of course, impute intentions). The concept of emergent strategy may also seem controver- sial — the notion that organizations can pursue "strategies" without intending them. After all, the term strategy has always been associated with voluntarism and free wilt, with what Dill (1979: 48) has called "machismo management." Were "de- liberate" and "emergent" strategies two distinct phenomena in practice, then there might be merit in applying the label strategy only to the former. We contend, however, that the two represent the end points of a continuum, indeed that there may be no such thing as a purely deliberate strategy (intentions realized precisely) or a purely emergent one (the total absence of intention, despite pattem in action). Moreover, how are we to determine intentions in a collective context, that is, the intentions of an organization? To assume that the intentions of the leadership are the intentions of the organization may not be justified, since others can act contrary to these intentions. Presumably, to be "organizational," intentions must be broadly shared — truly collective. And yet, stated intentions can belie the real ones — even the most detailed plans may be fanta- sies, or simply ptoys to fool the competition. Thus, to the extent that ostensibly deliberate strategies contain emergent elements, the two concepts deserve to share the same label, if only to force practitioners and researchers alike to recognize them as part of the same phenomenon (Mintzberg and Waters, 1985). The formulation of deliberate strategies would seem to accord well with the configuration of machine bureaucracy, which requires tight coordination in order to mass-produce its prod- ucts or services. It seems to fit with "simple structure," too, in which an entrepreneurial leader in personal control of an organization guides it according to his or her vision (Mintzberg, 1979), although vision need not be articulated formally and so is less "planned" and more liable to emergent shift. Our findings to date bear this out in part (Mintzberg, 1978; Mintz- berg and Waters, 1982,1984; Brunet, Mintzberg, and Waters, 1986), although it would seem to be more accurate to describe machine bureaucracies as elaborating, extrapolating, or copying existing strategies rather than formulating radically new ones. In contrast, the formation of emergent strategies would seem to be most compatible with the adhocracy con- figuration, in which many people — operating personnel, ex- perts and advisors of all kinds, managers at all levels — are potentially involved in the establishment of precedents and, so, the strategies. 162/ASaJune198B THE NATIONAL FILM BOARD OF CANADA AS AN ADHOCRACY The National Film Board of Canada (the NFB) was founded in 1939 as an agency of the federal government of Canada to produce and distribute films that would interpret Canada to Canadians and to people abroad. It has since developed an International reputation for the quality of its documentary film- making and for its innovations in film content, process, and technology. In 1975, the NFB had a permanent staff of approx- imately 950 people, with an annual budget of $23.5 million (Canadian). The NFB is an organization quite unlike any other: it is a film company with full production and marketing functions, concentrating on short, documentary films aimed at predomi- nantly nontheatrical markets, yet it falls under full state ownership. The NFB seems to fit the description of an adhocracy. Each of its films is unique, and many are distinguished for their sophis- tication and innovation. Each requires a distinct but temporary project, which draws together experts from a variety of func- tional departments. This suggests matrix structure, as is indi- cated in Figure 2, the NFB's organigram {area 1975), in which the filmmakers are shown suspended under a structure of studios and specialized functions, as well as regions, in fact, the absence of connections between the filmmakers and the Figure 2. Organigram of tiw National Rim Board of Canada (ca. 1975). Commissioner Planning & Research Public Relations Deputy Commissioner Director of Director of English Programming French Programming 1 1 English French Technical Ottawa ^ministration Distribution Production Production Services Bureau {similar to English Production) Financial Production Management Coordinator Research Technical Coordinator Coordinator 1 Studio A Studio B Studio C Studio 0 Studio E Studio G Studio H Animation Fiction Documentary Documentary "Challenge Multimedia Environment 'or Change" Regional Program* Regional Program Regional Program British Columbia Prairies Maritimes Filmmakers Filmmakers Filmmakers Filmmakers Filmmakers Filmmakers Filmmakers *No lines shown on original organigrann connecting Regional Programs to Studios or Filmmakers. 163/ASQ, June 1986 S&ategy Fonnation Structure indicates the relative weakness of authority and hierarchy in the organization. Controls exist in the NFB and attempts are made at formal planning, but most of the real coordination has to be achieved through mutual adjustment. Selective decentralization is perhaps best illustrated by the process surrounding the approval and funding of film projects. On average, the NFB has made about 75 films per year. Were it structured as a machine bureaucracy, word would presum- ably come down from on high, dictating subject matter, length, style, etc., resulting in a stable and rather deliberate film- content strategy. The facts here are rather different. Aside from the specific films commissioned, or "sponsored," by other government departments, all the others draw on the general budget of the organization, supplied by the govern- ment and supplemented by revenues from the sale and rental of films. Film ideas generally originate with a filmmaker in consultation with an executive producer and are eventually proposed to a standing committee — the Program Committee — which consists of representatives elected by the film- makers, appointees of the Distribution Branch (marketing), the Director of Production, and the Director of Programming. As chief executive officer, the Film Commissioner must approve this committee's choices, and almost inevitably does. Thus, control over film choices is distributed across al! levels of the organization and among various functions. On first impression, then, this project-by-project working rhythm would seem to epitomize adhocracy. As we probe more deeply into NFB history, we shall elaborate on this conclusion, showing how its structure is in fact an evolved compromise, a balancing act worked out in response to the demands of the individual project, on the one hand, and the need for a certain order, on the other. RESEARCH METHODOLOGY To track strategies as we have defined them requires a longitu- dinal approach; the researcher either has to five with an organization over time (through direct observation or periodic interviews) or has to reconstruct its behavior after the fact. The former approach is discouraged by our finding that strategies can remain stable over very long periods of time (Mintzberg and Waters, 1982). An historical reconstruction is, therefore, more feasible if not, in principle, more desirable. The method- ology accordingly proceeded through the following four basic steps: Step 1: Collectior) of basic data. The study began in the archives, with the search for traces of decisions and actions taken by the organization, sorted into various strategy areas, as well as traces of external trends and events in the environment and indicators of performance. Sources for this data included film catalogs, annual reports, in-house documents, and books and articles on the organization, all supplemented by inter- views to fill in gaps. Step 2: Inference of strategies and periods. The data on decisions and actions were arranged in chronological order, when possible plotted on common time scales, and were analyzed to infer patterns or consistencies over time, i.e., strategies. The strategies were then represented in symbolic 164/ASQ,June198S form, lined up on a common time scale, and scanned to infer distinct overall periods in the history of the organization. Step 3: Analyses of each period. At this point, the character of the research changed, from the systematic collection of "hard" data to the more intensive investigation of more qual- itative data. Interviews were conducted with key people who were present, and reports of the period were studied to explain the major changes in strategies. A number of thirty- to forty-year NFB veterans were accessible, as were several reports, books, and theses on or related to the organization, including the Massey Commission Report (1951), McKay (1965), James (1968). Gray (1973), Mclnnes (1974), Jones (1976). and Hardy (1979). Posing very specific questions in the interviews, based on the data gathered, aided recall and helped to avoid distortion. Step 4: Theoretical analysis. The researchers then brain- stormed around a number of theoretical questions to try to interpret each historical period, as welt as the entire study, in conceptual temis. Questions ranged across a variety of issues, including the patterns of strategic change, the relationships between deliberate and emergent strategies, the interplay of environment, leadership, and organization, and the relationship between strategy and structure. The study of the NFB from 1939 to 1975 took place on site over a number of years and eventually resulted in a 383-page document based on an extensive amount of data. In particular, each of the 2,839 original films completed between 1939 and 1975 were categorized along a number of dimensions and then were plotted and analyzed in various ways. The report that follows reflects the basic steps of the research. The first section describes the various realized strategies that we inferred in the behavior of this organization across thirty- seven years. After using these to identify distinct periods in its history, we then describe each period in the second section, seeking to put some flesh on the skeleton of realized strategies and to consider the origin of these strategies as deliberate or emergent. The final section develops a concep- tual interpretation of the entire study. STRATEGIES OF THE NFB, 1939-1975 The NFB, of necessity, took a large number of actions over the course of thirty-seven years. Discussion with its members, as well as our own overall assessment of its activities, suggested the following areas to be of central importance; first, the films produced, which could be analyzed in terms of various charac- teristics — the number made, their duration in minutes, their content, the original language of production (English or French), whether sponsored or not. in black and white or color, and in 16 or 35 mm — then in terms of finance, staffing, distribution (nnarketing), and the internal structure, A full presentation of all the strategies identified, together with the data and text that support their inference. Is well beyond the space limitations of this paper (but is available from the authors). We therefore present here only those strategies 165/ASa June 1986 Strategy FonnatkHi most revealing of the course of the organization over time and related material, in graphic ar)d symbolic form with a minimal amount of textual material, except for the area of film content (because of its importance and what it reveals about our method of inferring strategies). Finance. Since most of the NFB's financing arrived as direct grants from the government, and some from sales revenues, the area of finance is not only one of strategy (mternal actions} but also of environment and performance (see Mintzberg et al., 1986, for a discussion of these ambiguities). Figure 3 shows the total annual revenues of the NFB, by source. Of particular importance were the sudden, temporary dips in appropriations (notably im 947-1949,1958,1970-1971) known to em- ployees as periods of "austerity." Figure 3. Sources of National Film Board's revenues. 30 - Total 28 - Parliament appropriation Sales to government departments 26 - Rental royalties & other sales 24 - Note: Due to the tack of consistent records, the data in the early 22 - years are incomplete. 20 - REVENUE 18 - ($ MILLION) 16 - 14 - 12 - 10 - 8 - 6 - 4 - 2 - 0 ^ 7 1939 '40 '45 '50 '55 '60 '65 '70 '75 YEAR Staffing. Figure 4 shows the number of persons on regular staff since 1945, as wet! as temporary employees ("on con- tract") since 1957. The strategies inferred (for the most part, directly from these data) are represented symbolically below, in the Staffing diagram. In this representation of strategies, and all that follow, the vertical dimension is used to symbolize certain characteristics of strategies but does not represent any specific scale. 166/ASa June 1^5 Figure 4. Staffing lev^. 1000 - CO oc 850 - LU Regular Staff plus / 00 S 750 - Contract People / 650 - 550 - (Estimated) 450 - 350 - 250 ^ t 1939 '40 '45 '50 '55 '60 '65 70 '75 YEAR STMTING I '46 '60 Distribution. Distribution, the NFB's name for marketing, can be divided into three main channels — theatrical, nontheatrical (e.g., schools, church groups), and television, A variety of distribution strategies were inferred from a number of sources, notably archival records on the media employed in different periods and a statistical analysis of the duration of films by year. The latter, for example, revealed a heavy emphasis on films for television f ronn 1953 to 1957 (sharp rise of films in the 12-15 minute and then 26-30 minute range to fill slots in the newly created Canadian Broadcasting Corporation network), and the beginnings of a feature film strategy in the mid-1960s (first significant appearance of films longer than 60 minutes). These two strategies, as well as a number of others inferred (to be discussed later), are shown symbolically below, in the Distribution diagram. DISTRBiniON C TyFocu« \ t5min.«)min. \ 60min. + more30min. Taimision Distributic TamatiyBFeaturw Theatrical Distribution Rurat, I Nontheatrical Library ]t[ Union CiroiiU Aborttve Rvntal Scbems Panial Shiftto sProHkct Oriwitation 167/ASa June 1965 Strategy Fornnation Structure. Of particular interest in the structuring of this orga- nization are the forms of departmentalization used in the production function over the years, as well as the separatbn of French and English filmmaking (always a contentious issue). The various forms of structure are shown symbolically below, in the Production Structure diagram. Move to Montreal Informal Units 12 Units 4 Units 5 Engiish Units Pool System 7 English Studios 2 French Units (English & French} 3 French Studios f TV Unit" French Production Split Science Unit Ragionali- Ution I 1 1939 '75 Nutnber and length of films. An analysis and comparison of several types of records in the NFB archives uncovered 2,839 original films completed between 1939 and 1975. A plot of their number per year, as well as of their average duration in minutes, led to the inference of the strategies shown in the Film-Number-and-Length diagram. FILM MiMaER AM) LfNGTH Sharp Cycles around Higher Plateau Rapid (approK. 85film6/200Q min.l Growth PlBtBBU? -p. —1 '55 '75 Language of films. Plots of the number and duration of films in the French and Engiish languages revealed sharply different patterns for each, shown as strategies in the diagram below. iANGUAGEOFFIUWS ENGLISH FILMS Slow long-tarm increase in number lapprox. 2/year) and length laf^rox. 20 min./year) of gnglish films FRENCH RLMS StMdy growth in minimum number of French films, with sharp peaks (rHSraasing (ensth of Fmnch films —r 1939 '66 168/ASa June 1985 Spor)sorship of films. Films made by the NFB under direct contract to other government departments (e.g., training films for the RCMP) accounted for exactly one-quarter of the total over the years, and 21 percent of all minutes of films produced. Figure 5, which shows the proportion of minutes of sponsored films by year, reveals one strong surge after World War !l, which is depicted as a strategy in the diagram below. Rgure 5. fhvportion of minutes of sponsored fHms. 75 SPONSORSHfP r 1940 '45 '55 '60 '65 '70 '75 '50 Content of films. Finally, and perhaps most importantly, is the actual content of the films made by the NFB. The 2,839 films were each assigned to one of 37 content or related categories, chosen partly with regard to the categories actually used in NFB film catalogs. The data for three years in the 1970s was coded by two people independently, with almost complete agreement. We found it most useful to analyze each of the 37 categories over time in terms of three basic flows: • trickle — defined as a stream of five or fewer films per year. • blip — defined as one or two years in which production increased by an increment of five, from a trickle before and after. 169/ASQ, June 1986