2.1 Evolutionary Processes: Variation, Selection, and Retention

Variation, selection, and retention are the key processes of evolutionary organization theory. According to Reference AldrichAldrich (1999), variation “is a departure from routine or tradition, and may be intentional or blind.” Intentional variations result from the deliberate attempts of people or organizations to generate alternative opportunities and seek solutions to problems. Blind variations, on the other hand, “result from accidents, chance, luck, conflict, malfeasance, creative exploration, and so forth.” Aldrich defines selection as generated by forces “that differentially select or selectively eliminate certain types of variations .” Selection occurs within organizations as well as at the organization and population (e.g., ecosystem) levels. Retention “occurs when selected variations are preserved, duplicated, or otherwise reproduced so that the selected activities are repeated on future occasions or the selected structures appear again in future generations.” Finally, competition is defined as a struggle for scarce resources that occurs between strategic initiatives within organizations, between organizations, and between populations.

These are generic processes that are not necessarily biological in nature (e.g., Reference Boyd and RichersonBoyd & Richerson, 1985; Reference DawkinsDawkins, 1986) and can be fruitfully applied to organizational evolution. The central axiom of evolutionary organization theory is that these key processes are necessary and sufficient for organizations to emerge and evolve. They occur simultaneously at the organization, population, and community levels of analysis. Strategy-making cast in terms of intraorganizational variation, selection, retention, and competition adds an additional level of analysis, which can be studied in its own right, as well as in relation to the other levels. Purposeful variations at the intraorganizational level, for instance, may sometimes look like emergent variations at the organization level, and intraorganizational selection processes can lead to organization-level adaptation (or maladaptation).

2.2 Three Conceptual Tools of Strategy-Making

In the realm of theory, the processes of variation, selection, and retention are general, abstract, and nonexperiential. In the realm of practice, on the other hand, common language describing strategy-making is particular, concrete, and experiential. Business leaders do not use the terminology of pure theory to think and talk about strategy-making, and find it difficult to relate to its formal mathematical and statistical models. Scholars, on the other hand, find it difficult to gain deeper insight when limited to everyday language, and like to do more than produce a coherent and complete narrative of best practice. Conceptual frameworks help bridge the gap between pure theory and best practice (see Reference Winter and TeeceWinter, 1987). They are specific, substantive, and suggestive. The boxes-and-arrows charts used to represent them can be more readily understood and related to by business leaders and scholars alike, and we use these throughout the Element in the various tools and frameworks we present.

Three tools form the backbone of identifying and analyzing the role of strategy-making in firm evolution at three interrelated levels of analysis: (1) Tool I (also called the strategy diamond): company–environment interface level; (2) Tool II: company level; and (3) Tool III: intracompany level. The evolutionary research lens has strong “zooming” capabilities: from environment and company interface-level forces and their interactions to company-level induced and autonomous strategy processes and their evolving balance, and then to the intracompany-level details of strategic leadership activities involved in these processes. Figure 2 presents these tools.

Figure 2 Three interrelated conceptual tools of an evolutionary research lens.

Source: Reference BurgelmanBurgelman (2002b: 9)

2.2.1 Tool I: Dynamic Forces Driving Firm Evolution (“Strategy Diamond”)

The framework of dynamic forces driving organizational evolution – Tool I in Figure 2, also referred to as the strategy diamond, and Figure 3 – serves to analyze industry-level and company-level processes of variation, selection, and retention and their interplay throughout organizational evolution (see Reference BurgelmanBurgelman, 1994).Footnote ⁵ This framework synthesizes the strategic positioning (Reference PorterPorter, 1980), resource-based view, and dynamic capabilities perspectives (Reference BarneyBarney, 1991; Reference TeeceTeece, 2007) in strategic management, and the formulation and execution aspects of the strategy-making process (Reference Mintzberg and WatersMintzberg & Waters, 1985).

Figure 3 Dynamic forces driving firm evolution.

Source: Reference BurgelmanBurgelman (1994)

Tool I comprises five dynamic forces that drive organizational evolution. The basis of competitive advantage in the industries in which the company chooses strategic product-market positions is determined by the interplay of external forces that constitute the firm’s business ecosystem. Distinctive competence encompasses the firm’s differentiated skills, resources, and capabilities. Corporate strategy reflects CEO and top management beliefs about how the firm should position itself in its business ecosystem and use its distinctive competences to occupy, defend, and leverage these strategic positions. Strategic action concerns the firm’s consequential actions (whether taken or not taken). The internal selection environment mediates – mostly through resource allocation to different business models and product offerings – the links between corporate strategy and strategic action, as well as the coevolution of industry-level sources of competitive advantage and firm-level sources of distinctive competence. These five forces can be aligned or misaligned.

Strategic inflection points. When the normative, technological, economic, and cognitive rules (among others) that determine the basis of competitive advantage in the industry and/or broader ecosystem change (Reference Burgelman and GroveBurgelman & Grove, 2007), the internal selection environment may become misaligned with the external selection environment, and the company risks being selected out and replaced by new competitors. For instance, ecosystem interdependencies can influence a company’s ability to respond to new technologies (Reference Kapoor and LeeKapoor & Lee, 2013) or react to latent disruptive threats from complementors (Reference Adner and LiebermanAdner & Lieberman, 2021). In such situations, the company faces a strategic inflection point (SIP). Confronted with a SIP, the company is entering (or has already entered) a period of crisis during which, if it does not change its strategy, it faces a material threat to its profitable growth prospects and potentially even its demise as an independent entity. If it can change its strategy in time, however, it may be able to get onto a new profitable growth trajectory. Figure 4 illustrates the concept of SIPs.

Figure 4 Strategic inflection points and changing the rules of the game.

Source: Adapted from Reference Burgelman and GroveBurgelman and Grove (1996: 11)

As suggested in Figure 4, the emergence of a SIP is indicated by growing misalignment between the basis of competitive advantage and the company’s distinctive competence, and between the corporate strategy and strategic action within the “strategy diamond.” This is typically related to the “strategic dissonance” among the organization’s managers (Reference Burgelman and GroveBurgelman & Grove, 1996). One option available to the company is strategic exit or abandoning a given business and moving to a different growth trajectory or even a new ecosystem, where alignment becomes important once again. For instance, this is what Intel eventually did (Reference BurgelmanBurgelman, 1994), exiting the DRAM industry and committing to the microprocessor industry (Section 3.2). Another option is the corporate split, which is particularly useful under conditions of ecosystem bifurcation, as evidenced by the split of HP in 2015 by its CEO Meg Whitman (Section 4.2).

2.2.2 Tool II: Evolutionary Framework of the Strategy-Making Process

Zooming in on the strategy-making process, the framework of induced and autonomous strategy processes – Tool II in Figures 2 and 5 – further illuminates the role of internal variation, selection, and retention.

Figure 5 Evolutionary framework of the strategy-making process.

Source: Reference BurgelmanBurgelman (2002a)

Autonomous strategic action refers to action taken without top-down direction from management by organizational members that can result in new initiatives, such as the emergence of microprocessors at Intel or printers at HP. Autonomous strategic action is typically the source of internal variation in strategic initiatives and varied experiments.

Induced strategic action refers to action taken by organizational members following top management direction – for instance, pursuing specific R&D projects (such as incremental product innovation) aligned with the structural organizational context. Both induced and autonomous strategic actions can result in new initiatives that are selected and retained (or not) through the organizational structural and strategic context. The good functioning of the internal selection environment (i.e., producing sufficient variety of new initiatives) is important to ensure organizational renewal, particularly when the external environment might be evolving.

We also propose that as an autonomous initiative gains impetus in the strategy-making process, a critical top management role is to evaluate first, the extent to which the autonomous opportunity has been validated (through the process of strategic context determination), and second, the extent to which available cash reserves are sufficient to protect the company from disaster in case the scaled-up autonomous initiative ultimately fails (see Table 2). This suggests four possible strategic choices: (1) “safe bet”: validated opportunity and sufficient cash reserves; (2) “bet the company”: validated opportunity but insufficient reserves; (3) “wait to bet”: not-yet-validated opportunity and sufficient cash reserves, which also raises the question how long to wait; and (4) “desperate bet”: not-yet-validated opportunity and insufficient cash reserves. Having to “bet the company” and “desperate bet” are the strategic choices to avoid.

Table 2 Critical role of top management related to autonomous strategic opportunities

		Autonomous opportunity
		Validated	Not yet validated
Cash reserves	SufficientFootnote a	Safe bet	Wait to bet: but how long?
	Insufficient	Bet the company	Desperate bet

^a To protect the company from disaster in case scaled-up autonomous initiative fails.

Source: Reference Burgelman and GroveBurgelman and Grove (2007).

Over time, the outcomes of the combined internal and external selection processes operating on induced and autonomous variations can either ensure renewal or lead to organizational demise. Like the role of founding and disbanding as sources of population-level change in organizational ecology, entry into new businesses and exit from existing ones are sources of corporate-level transformation in the internal ecology of strategy-making (Reference BurgelmanBurgelman, 1991).

2.2.3 Tool III: Process Model of Strategy-Making (Internal Corporate Venturing)

The process model – Tool III in Figures 2 and 6 – helps zoom in further on the patterns of detailed strategic leadership activities involved in the internal variation, selection, retention, and competition processes associated with entering into new businesses.

Figure 6 Process model of strategy-making (internal corporate venturing).

Source: Reference BurgelmanBurgelman (1983c)

This tool highlights the simultaneous strategic actions of multiple actors distributed throughout the organization that together contribute to strategy-making. The research of Burgelman at Intel, for instance, showed that coordinating the simultaneous actions of multiple strategic actors into a coherent pattern of strategy-making, especially in the face of environmental change, was a major challenge. During the tenure of Gordon Moore as CEO, coordination was achieved primarily through the internal selection environment. During Andy Grove’s tenure as CEO, Intel achieved coordination through strategic direction set by top management.

Coordination through internal selection is a powerful manifestation of self-organization, which depends on “mutually dependent criteria of action (utility functions)” (Reference PrigoginePrigogine, 1980). Such criteria seem closely related to the idea of “focal points,” which is used by some scholars to give operational meaning to culture (cf. Reference SchellingSchelling, 1963).Footnote ⁶ Andy Grove strongly focused Intel’s induced strategy process on the core microprocessor business and concentrated strategic decision-making at the top, thereby limiting the role of lower levels mostly to the execution of the official corporate strategy. While tightly coordinated and extremely successful in executing the core business strategy, Intel’s strategy-making process also became severely constraining. The core business strategy and associated structural context at Intel had an overwhelming influence on the definition of new business opportunities. The ICV process-model tool and several derivatives are discussed in further detail in Section 3.

Taking a broader view, there are clear links between process modeling and ecological notions of self-organization (Reference Thiétart and ForguesThiétart & Forgues, 1995). Ilya Reference PrigoginePrigogine (1980), Nobel laureate in Chemistry, uses the Lotka–Volterra equations of prey–predator competition, which are part of the conceptual foundation of organizational ecology, in a discussion of self-organization to examine “structural stability.” According to Reference PrigoginePrigogine (1980: 109), structural stability “seems to express in the most compact way the idea of innovation, the appearance of a new mechanism and a new species, which were initially absent in the system.” He views “structure” as resulting from the actions (choices) of multiple actors, having “in part at least mutually dependent criteria of action (utility functions)” (p. 126). Referring to the “baroque of the natural world,” he posits:

ecosystems contain many more species than would be “necessary” if biological efficiency alone were an organizing principle. This “over creativity” of nature emerges naturally from the type of description being suggested here, in which “mutations” and “innovations” occur stochastically and are integrated into the system by the deterministic relations prevailing at the moment. Thus, we have in this perspective the constant generation of “new types” and “new ideas” that may be incorporated into the structure of the system, causing its continual evolution.

(p. 128)

Keeping in mind the pitfalls of the fallacy of unwarranted analogy (cf. Reference MertonMerton, 1957), there seems nevertheless to be a potentially interesting isomorphism between Prigogine’s analysis and the analysis of the role of the induced and autonomous processes in strategy-making. Autonomous initiatives generate innovations (new ideas) in the organization, but they are viable only if they can be integrated into the corporate strategy and eventually become part of the induced strategy process. Strategic context determination is the part of the process through which this integration is attempted. How strategic context determination works, and what determines its success, may be the central issue in a theory of corporate entrepreneurship (Reference BurgelmanBurgelman, 1983a).

2.3 The Crucial Role of the Internal Selection Environment

The insights generated with the three tools of the evolutionary lens for studying strategy-making processes indicate the importance of the internal selection environment for maintaining organizational adaptation and longevity. They also shed light on several important dilemmas occurring during the dynamics of organizational adaptation that raise further questions: How much of a company’s resources should be devoted, over time, to the induced and autonomous strategy processes, respectively? How to manage the difficulties of balancing opposing orientations (toward variation or selection) within the strategy-making process? These questions have potentially important theoretical and normative implications for designing the internal ecology of strategy-making as an adaptive organizational capability.

2.4 Conclusion

The intraorganizational ecological perspective on strategy-making proposes the need to balance the variation-reducing (induced) and variation-increasing (autonomous) processes in the organization’s internal selection environment. One process leads to relative inertia and incremental adjustments, while the other expands the organization’s domain and renews its distinctive competence base, countering inertia and serving adaptation through strategic renewal. The balancing of these processes constitutes the fourth level in the hierarchy of an ecological system, which comprises organization, population, and community levels. The effectiveness of the internal selection environment may be at the heart of strategy-making as an adaptive organizational capability.

The intraorganizational ecology perspective also extends Reference MintzbergMintzberg’s (1978) strategy-making framework by documenting some of the sources of emergent strategy more explicitly, by further elucidating the organizational decision processes through which emergent strategies become part of realized strategies (Reference Mirabeau and MaguireMirabeau & Maguire, 2014), and by identifying simultaneous and sequential feedback mechanisms between realized and intended strategy.

3.1 Within the Firm: The Internal Corporate Venturing Process Model

This model (for the original study see Reference BurgelmanBurgelman, 1983c) investigates the processes through which a diversified major firm transforms R&D activities at the frontier of corporate technology into new businesses through ICV. The research was carried out in one large, USA-based, diversified high-technology firm. The case firm had traditionally produced and sold various commodities in high volumes, but had also tried to diversify through the internal development of new products, processes, and systems. In doing so, it hoped to get closer to the final user or consumer and capture a greater portion of the total value added in the chain from raw materials to end products. A longitudinal processual approach (Reference PettigrewPettigrew, 1990) was adopted, where the ICV process was studied exhaustively in one setting. Data were collected on six ongoing ICV projects that were in various stages of development. The historical development of each case was traced and the progress of each case during a fifteen-month research period was observed and recorded. These materials formed the basis for a comparative analysis of the six projects.

Figure 7 maps the activities involved in ICV onto the B–B process model. It shows how the strategic process in and around ICV is constituted by a set of key activities (the shaded area) and by a set of more peripheral activities (the nonshaded area). These activities are situated at the corporate, middle, and operational levels of management. Figure 8, which can be superimposed on Figure 7, shows how these different activities interlock with each other, forming a pattern of connections. The relative importance of activities is indicated by the different types of line segments. The data also suggested a sequential flow of activities in this pattern, as indicated by the numbers in Figure 8.

Figure 7 Key and peripheral activities in a process model of ICV.

Source: Reference BurgelmanBurgelman (1983c: 230)

Figure 8 Flow of activities in a process model of ICV.

Source: Reference BurgelmanBurgelman (1983c: 230)

This process model shows that ICV is primarily a bottom-up (autonomous) process and depicts the key role performed by middle management. Entrepreneurial activities at the operational and middle levels (1, 2, 3) interact with the selective mechanisms of the structural context (5). These selective mechanisms can be circumvented by activating, through organizational championing (6), the strategic context, which allows successful ICV projects to become retroactively rationalized by corporate management in fields of new business delineated by the middle level (7, 8). These parts of the pattern, represented by the solid line segments in Figure 8, constitute the major forces generated and encountered by ICV projects.

The finely dotted lines in Figure 8 (4, 9) represent the connections between the more peripheral activities in the ICV process and their linkages with the key activities. Corporate management monitor resource allocation to ICV projects. Middle-level managers manage these resources and facilitate collaboration between R&D and businesspeople in the definition of new business opportunities; however, these activities support the definition process rather than driving it. In the same fashion, authorizing further development is clearly the prerogative of corporate management, but this is a result, not a determinant, of the impetus process. In the strategic context determination process, gatekeeping, idea-generating, and bootlegging activities by operational-level participants were all found to be important in developing a basis for further definition processes, but are a result of the process rather than a determinant. In the process of structural context determination, questioning of the structural context by operational-level participants and efforts by middle managers to negotiate changes in it are reactive rather than primary.

The broken line segments in Figure 8 (10, 11) indicate two important delayed effects in the ICV process. First, the successful activation of the process of strategic context determination encourages further entrepreneurial activities at the operational level, thus creating a feedforward loop to the definition process (10). Second, corporate management attempts to influence the ICV process, primarily through its manipulations of the structural context. These manipulations appear to be in reaction to the results of previously authorized ICV projects. This creates a feedback loop (11) between the core and overlaying processes.

One important finding of this study is that the motor of corporate entrepreneurship (or even strategic innovation more broadly) resides in the autonomous strategic initiatives of individuals at the operational levels in the organization. However, because of their very nature, autonomous initiatives are likely to encounter serious difficulties in the diversified major firm. Their proponents must often cope with problems of resource allocation because they attempt to achieve objectives that are beyond the corporation’s horizons. Because such initiatives require unusual, even unorthodox, approaches, they create managerial dilemmas that are temporarily resolved through the deliberate neglect of administrative issues during the early stages.

This study also elucidates the key role of middle-level managers in the strategy-making process. The middle manager performs the crucial role of linking successful autonomous strategic behavior at the operational level with the corporate concept of strategy. Both the continuation of the impetus process and the change of the corporate strategy through the activation of the process of strategic context determination depend on the conceptual and political capabilities of managers at this level. Relatedly, corporate management’s role seems to be limited to the retroactive rationalization of autonomous strategic initiatives that have been selected by both the external environment and the internal environment. When they do intervene, top management’s direct influence is through the manipulation of structural context.

This study can also be related to the validity of rational versus natural selection models to explain organizational growth and development (Reference AldrichAldrich, 1979; Reference Pfeffer and SalancikPfeffer & Salancik, 1978; Reference WeickWeick, 1979). The multilayered picture of the strategic management process presented here suggests that these strategic choice processes, when exercised during innovation initiatives, take the form of experimentation and selection, in line with organizational evolutionary theory rather than the strategic planning or rational actor approach. Relatedly, this study also illustrates that large, complex business organizations have separate variation and selection mechanisms. Previously unplanned, radically new projects at the product-market level are generated from the relatively unique combination of productive resources of such firms. Not all of these projects are retained – not so much because the market may turn out to be unreceptive, but because they must overcome the selection mechanisms in the internal selection environment of the firm, which reflect, normally, the current strategy of the corporation, that is, the retained wisdom of previously selected strategic behavior. Thus, the experimentation and selection model draws attention to the possibility that firms may adopt externally unviable projects, or fail to adopt externally viable ones, and provides a clue as to why firms occasionally produce strange innovations.

The insights generated by the present study also have some implications for further research on the management of the strategy-making process in general. Comparative research studies of a longitudinal-processual nature, carried out at multiple levels of analysis, are necessary to document and conceptualize the multilayered, loosely coupled network of interlocking, simultaneous, and sequential key activities that constitute the strategy-making process. Following Reference BowerBower (1970), the study found it useful to focus the research on particular strategic projects rather than on the strategy-making process in general. This is consistent with Reference QuinnQuinn’s (1980: 52) observation that top managers “deal with the logic of each subsystem of strategy formulation largely on its own merits and usually with a different subset of people.” A concrete focus, it would seem, is more likely to produce data on the vicious circles, dilemmas, paradoxes, and creative tensions that are embedded in the strategy-making process.

3.2 Within the Firm: Strategic Business Exit Process Model

This model (see Reference BurgelmanBurgelman, 1996 for the original study) is based on the Intel Corporation’s SBE in the mid-1980s from the dynamic random access memory (DRAM) business, which had been its core business during the 1970s. The data on Intel’s exit from DRAMs were part of a longitudinal field study of the evolution of its corporate strategy. The study was based on a longitudinal, two-stage, nested case study design within one corporate setting (e.g., Reference Leonard-BartonLeonard-Barton, 1990; Reference YinYin, 2018). While concentrating on one organization with more than twenty years of continuity in leadership limits the generalizability of the findings, it also made it possible to gain access to sources with intimate knowledge of the details of the firm’s evolution.

The construction of the SBE process model involved an iterative approach of qualitative pattern-matching, moving back and forth between the B–B process model and the DRAM exit data. Qualitative pattern-matching suggested that the overall structure of the ICV process model was useful to conceptualize the strategy-making involved in SBE. Qualitative pattern-matching also revealed, however, that the managerial activities involved in SBE, and the linkages between them, were quite different. This effort yielded new categories to conceptualize the managerial activities involved in SBE. These are shown in Figure 9.

Figure 9 Managerial activities in a process model of SBE.

Source: Reference BurgelmanBurgelman (1996: 197)

Figure 9 shows that each part of the SBE process model involved top, middle, and operations-level managerial activities. The relations between the operational and middle levels are viewed in terms of responsibilities rather than hierarchy. Intel was a multibusiness firm, but its organizational structure maintained a strong functional dimension, with managers of the different product divisions needing to obtain R&D and manufacturing support for their products from the functional units. This is reflected in Figure 9: operational managers were business managers who interacted with the product-market environment and were primarily concerned with business strategy-level issues, while middle managers were functional managers primarily concerned with maintaining and developing the firm’s distinctive technological competencies and with resource-allocation issues. Top managers, on the other hand, were primarily concerned with corporate strategy-level issues. Some activities – shown in bold type – drove the SBE process more strongly than others.

Figure 10, which can be superimposed on Figure 9, shows how these different activities interlock with each other, forming a pattern of connections. The arrows indicate the dominant direction of influence among the key managerial activities. The relative importance of activities is indicated by the different types of line segments. The rough sequential and simultaneous flow of the activities in this pattern is indicated by the numbers.

Figure 10 Flow of activities in a process model of SBE.

Source: Reference BurgelmanBurgelman (1996: 198)

The solid lines in Figure 10 identify the activities that were strongest in driving the SBE process. Inertial competence deployment of middle-level managers (1) led to the unlinking of market needs and Intel DRAM products. Operational and middle-level managerial activities at the business level (3, 4) responded to the selective pressures of the structural context (1’) favoring other businesses, and to DRAMs falling behind (2) in the market. At the corporate level, the undermining of the DRAM business (7) resulting from resource shifts, combined with the continued selective pressures from the structural context (8), set the stage for middle managers’ technological uncoupling activities and top management’s strategic recognition and rationalization of the exit from DRAMs (9, 10). The undermining of the DRAM business was unsuccessfully countered by the efforts of some DRAM middle-level managers to negotiate changes in Intel’s structural context (7’).

The finely dotted lines in Figure 10 represent the set of activities that were weak drivers of the SBE process. At the business strategy level, top management’s equivocation about the importance of DRAMs for Intel (5) led them to continue to invest heavily in DRAM R&D, which reinforced the inertial competence deployment of middle-level managers (definition). On the other hand, top management authorized (6) incremental shifting of scarce manufacturing resources from DRAMs to other businesses (impetus). At the corporate strategy level, operational managers associated with the DRAM business questioned Intel’s business strategy in DRAMs (5’) (strategic context). Operational-level managers also questioned some aspects of the structural context, in particular the resource allocation rule, as shifts in resource allocation were made and authorized (6’).

The broken line segments in Figure 10 indicate two delayed effects in the SBE process. First, a last-ditch, unsuccessful effort was launched by a DRAM business manager to change the strategic context of the DRAM business in 1985 (11). Second, the decision to exit from DRAMs led top management to reconsider some aspects of Intel’s structural context (12).

This study shows how process models of substantive areas of strategy-making such as SBE provide windows into the “black box” of strategy-making in complex organizations. The approach is positive descriptive, and it helps to identify and explain paradoxes, vicious circles, dilemmas, and tensions in the strategy-making process that derive from the activities of managers who are differentially situated in the organization and respond to different external and internal pressures. For instance, the process model of SBE shows how the process technology competence that allowed Intel to be the first successful mover in DRAMs paradoxically also bore the seeds of its later failure in DRAMs. It also explains how a vicious circle (“death spiral”) in resource allocation to DRAMs resulted from some middle-level managers’ inertial deployment of process technology competence, which failed to meet competitive challenges in DRAMs, while others responded to the internal pressures of the structural context by moving scarce manufacturing capacity away from DRAMs.

This study also shows how developing a grounded process model for a substantive area such as SBE helps produce categories and concepts that, while somewhat rudimentary and evocative, are closely tied to the phenomenon and enrich the repertoire for conceptualization. In this respect, the study identified unlinking and repositioning; resource shifting and technological uncoupling; and strategic recognition and structuring as key categories of managerial activity, performed by different levels of management, that shaped the SBE process. The study also showed how these activities brought about the dissolution of the strategic context of a core business of the firm. Linked together in the process model, these categories and concepts depict the complex pattern of managerial activities and organizational forces associated with the SBE phenomenon in a parsimonious way and show the underlying order and sources of rationality in a process that, on the surface, looks chaotic.

Relatedly, a comparison of the process model of SBE to process models of other substantive areas of strategy-making provides additional insight. The SBE process differed in several respects from the process of ICV (Reference BurgelmanBurgelman, 1983c). New business development through ICV was structurally separated from mainstream operations using a new venture division. This purposefully protected new businesses from internal competition for resources for a while, and provided some leeway for operational-level product championing and middle-level organizational championing activities. Strategic context determination for a new venture was expected to involve a positive, upward spiral of growth. SBE, in contrast, took place in an integrated organizational structure where the DRAM business could not escape internal competition for resources with other businesses. Strategic context dissolution of memory businesses involved negative, downward spirals of decline. This made product championing and organizational championing for memory businesses impossible tasks, like swimming against an overwhelming tide.

The process model of SBE supports theory that views the notion that top management defines a strategy that is then expeditiously implemented as an incomplete representation of strategy-making and enactment processes (e.g., Reference MintzbergMintzberg, 1978; Reference Mintzberg and WatersMintzberg & Waters, 1985; Reference QuinnQuinn, 1989; Reference WeickWeick, 1979) and of organizational change more broadly defined (Reference Grinyer and McKiernanGrinyer & McKiernan, 1990; Reference LevinthalLevinthal, 1991; Reference PettigrewPettigrew, 1990). The SBE process model underscores the difficulty of establishing who the relevant “actor” is (Reference Allison and ZelikowAllison & Zelikow, 1999) when considering strategic interaction involving complex organizations, and it provides additional insight into the internal sources of inertia in strategic response and the myopia of organizational learning (Reference Levinthal and MarchLevinthal & March, 1993). The model also helps to show why the application of formal theories of competitive interaction, such as game theory (Reference CamererCamerer, 1991; Reference SalonerSaloner, 1991), is problematic when complex organizations are involved.

Another insight from this study is that the process model documents matrixed levels of management by levels of strategy-making, which helps to elucidate the interplays between business- and corporate-level strategy-making. Conceptualizing the complete process, depicting simultaneous as well as sequential managerial activities and contextual forces, also helps to illuminate the difficulties in establishing when exactly the strategic exit decision is made, and when SBE occurs. As a diagnostic tool, it may draw top management’s attention to business-level strategic activities that are already paving the way for an exit, but of which the corporate-level strategic implications have not yet been fully realized, and vice versa. It may also help top management to assess more carefully the evolving links between business-level product-market issues and corporate-level competence issues.

This study also reveals that generic strategy may be a more enduring feature of a company than its substantive strategy. Intel’s generic strategy based on differentiation and product leadership endured in DRAM for several product generations in the face of the changing basis of competitive advantage in the industry, which favored the cost leadership-based generic strategy of new entrants. After the exit from DRAM, the differentiation-based generic strategy remained in force, having found a new substantive expression in microprocessors (Reference Burgelman, Cogan, Graham, Chesbrough and BurgelmanBurgelman et al., 1997).

Finally, by elucidating internal selection processes, process models of strategy-making also contribute to the debate concerning the prevalence of punctuated equilibrium models of strategic change (e.g., Reference GersickGersick, 1991; Reference Romanelli and TushmanRomanelli & Tushman, 1994; Reference Tushman and RomanelliTushman & Romanelli, 1985). Intel’s transformation from a memory into a microcomputer company seems at first to have been a rather abrupt organization-level change taking place in the 1984–1985 time frame. A closer look at the strategy-making processes that played out in the late 1970s to mid-1980s time frame, however, reveals that the organization-level change was the culmination of gradual replacement of memory by microprocessors as Intel’s core business – a change that took top management a relatively long time to come to grips with. Figure 11 shows this.

Figure 11 Intel’s transformation from memory company to microcomputer company.

Source: Reference BurgelmanBurgelman (2002b: 122)

The process model thus calls into question the universal applicability of the punctuated equilibrium model, and may help to establish more precisely how and when strategic change takes place. In other words, strategic change may take place, or at least begin, before it is recognized or acknowledged as such by top management due to the activities by the operational and middle-level managers.

3.3 Within the Firm: The Corporate Venture Capital Process Model

As noted in Section 2, in highly dynamic technological environments, companies may face insufficient variation despite having a highly active autonomous strategy process. Adopting CVC as an innovative capability has many features that make it potentially effective for combining internal sources of innovation with external innovation: (1) it helps to keep pace with the innovation happening externally without trying to internalize everything; (2) it helps in exploring far-out bets with lower capital requirements than outright acquisition; and (3) it eases the strain on operational resources by leveraging start-ups’ resources to multiply corporate innovation and R&D efforts.

Based on several months of research into JetBlue Technology Ventures (JTV), the CVC arm of JetBlue Airways, including interviews with top executives and public archival material (Reference Burgelman and SridharanBurgelman & Sridharan, 2021), we constructed a B–B process model of JTV’s strategy-making. The construction of the CVC process model involved an iterative approach of qualitative pattern-matching, moving back and forth between the B–B process model and the JTV CVC data. Qualitative pattern-matching suggested that the overall structure of the B–B process model was useful to conceptualize the strategy-making involved in CVC. It also revealed, however, that the managerial activities involved in CVC, and the linkages between them, were quite different. This effort yielded new categories to conceptualize the managerial activities involved, shown in Figure 12.

Figure 12 Process model of CVC.

Source: Reference Burgelman and SridharanBurgelman and Sridharan (2021: 33)

Based on our data, the CVC process model shown in Figure 12 reflects that JetBlue’s core business strategy was clearly defined and well understood throughout the organization; in other words, the strategic context part of the process was relatively fixed, and did not need to be changed to integrate the innovations brought by JTV’s start-up innovations. The structural context for CVC-corporate relations was also robustly designed. Hence, Figure 12 shows how the overlaying processes provided unequivocal feed-forward corporate management guidance to the core definition and impetus parts of the B–B process model to highlight CVC-driven external innovation activities, and the feedback the core processes provided to reinforce the corporate management guidance. This model shown in Figure 12 applied to the short- and mid-term strategic time horizons of JetBlue: 40 percent of investments focused on near-term benefits (zero to eighteen months) and 50 percent focused on benefits within a medium time horizon (two to five years). Note that an additional application of the B–B process model could have focused on the last 10 percent of the investments, which was focused on innovations that could impact JetBlue in the long term (seven years and beyond). In that case, however, the corporate strategy related to new business opportunities would be less well defined (as it is in ICV), and consequently the strategic context part of the process model would be less tightly determined and would have to allow for CVC initiatives to drive changes to the corporate strategy. This application of the B–B process model was not pursued in the JTV research.

Figure 12 shows that each part of the CVC process model involved top, middle, and operations-level managerial activities. The relations between the operational and middle levels are viewed in terms of responsibilities rather than hierarchy. In the definition part of the process, operational managers were JetBlue managers who interacted with the CVC personnel to perform need linking of the core business to potential external innovative solutions; middle managers were JetBlue functional managers primarily concerned with stewardship of resource allocation in developing the firm’s distinctive technological competencies in collaboration with CVC management; and top managers were primarily concerned with monitoring the core business-CVC collaboration from the corporate strategy-level perspective.

In the impetus process, top management’s main role was authorizing the investments made by the CVC; the middle management of the CVC played an important role in ecosystem building for the start-ups it invested in and for broadening the potential product-market scope of JetBlue, as it intended to gradually redefine itself as a “travel” company rather than an airline services company; at the same time, JetBlue operational managers collaborated with CVC personnel and the start-up venture personnel in securing deployment activities. As Figure 12 shows, some managerial activities – shown by larger arrows – were stronger in driving the CVC process than others.

This model provides a link with the research on open innovation that promulgates connections between the firm and its ecosystem (Reference Bogers, Chesbrough and MoedasBogers et al., 2018; Reference Chesbrough and AppleyardChesbrough & Appleyard, 2007) and sets the stage for our next model, which takes us beyond the boundaries of the firm.

3.4 Beyond the Firm: The Business-Model Disruption Process Model

This process model (for the original study, see Reference Snihur, Thomas and BurgelmanSnihur et al., 2018) tracks processes unleashed by a disruptive innovator in its ecosystem, using the B–B process model to theorize these processes beyond the firm. The underlying research investigated the early years of Salesforce (1999–2006), during which it emerged as the leader in on-demand customer relations management software and instigated the emergence and development of the cloud ecosystemFootnote ⁷ around the concept of software-as-a-service (SaaS). Through an iterative process of coding the available historical information, analyzing the existing literature, and broader conceptual reasoning about temporal and evolutionary dynamics, the study develops a parsimonious process model of the role of framing and adaptation in business-model disruption.

In this case, the B–B model was adapted to study ecosystem-level outcomes, in particular the ecosystem value proposition (Reference Thomas, Autio and GannThomas et al., 2022). As structures of value codiscovery and coproduction, ecosystems feature relationships that are “not fully hierarchically controlled” (Reference Jacobides, Cennamo and GawerJacobides et al., 2018: 2264) and which are instead informal, using mechanisms such as mutual understandings of participant relationships, role definitions, and technological architecture (Reference AdnerAdner, 2017; Reference Jacobides, Cennamo and GawerJacobides et al., 2018; Reference Thomas and RitalaThomas & Ritala, 2022). As such, by disentangling the pattern of activities carried out by various ecosystem actors, such as the disruptor, incumbent, customers, partners, and the media, the B–B process model was able to provide insight into the emergence of the Salesforce cloud ecosystem.

Transposing the B–B process model from the intraorganizational to the ecosystem context helped elucidate the generative mechanisms of ecosystem evolution by showing how the activities of ecosystem actors combine to generate strategic outcomes at the level of the ecosystem (rather than the organization). Thus, in contrast to the previous B–B studies, the level of analysis is the ecosystem, and the unit of analysis is the process (or processes) associated with business-model disruption. The process model helps examine the interactions, alignment, and misalignment of distinct ecosystem actors during disruption, its key advantage being the depiction of simultaneous as well as sequential activities that might result in nonlinear (e.g., circular or iterative) dynamics. This is particularly important in the case of disruptive innovation, where a time lag between the actions of the disruptor and the incumbent is often documented (Reference ChristensenChristensen, 1997; Reference Christensen, Raynor and McDonaldChristensen et al., 2015).

Given that the process model considers how the processes initiated by the business-model innovator (i.e., a pioneering cloud-based SaaS provider) interact with the incumbent and other actors, such as customers, partners, media, and so forth, this study reconceptualizes the business-level processes to consist of the disruptor’s and the incumbent’s framing. In the original model, the contextual overlay processes consisted of the strategic and structural context, which make up the context in which change occurs inside the organization (e.g., Reference GilbertGilbert, 2006; Reference Pratap and SahaPratap & Saha, 2018). As this study lifts the level of analysis from organizational to ecosystem, there are two contextual aspects: the ecosystems of the disruptor and the incumbent, which are originally separate from each other, but over time coevolve in an interlocking fashion as customers and partners migrate from the incumbent’s ecosystem to the disruptor’s. Figure 13 maps the activities involved in business-model disruption onto the B–B process model.

Figure 13 Ecosystem activities in a process model of business-model disruption.

Source: Reference Snihur, Thomas and BurgelmanSnihur et al. (2018: 1294)

Figure 14 illustrates our process model of business-model disruption by depicting the main constructs and the links between them. The first link, a feed-forward process representing the disruptor’s framing, connects the disruptor’s original business model and the participation of customers, partners, media, and analysts in the emergence of a new, disruptor-centered ecosystem. The second link represents the feedback relationship between the new ecosystem stakeholders’ response to the disruptor’s framing and the disruptor adapting its business model to the needs of the new ecosystem. This led to an important insight – namely, that the disruptor’s framing to the new ecosystem stakeholders and the adaptation of its business model in response to their needs constituted a virtuous framing-adaptation cycle that sustained the increasing adoption of the disruptor’s business model. The third link represents the incumbent’s framing: a feed-forward process directed to the incumbent’s ecosystem stakeholders. The fourth and final link represents the incumbent’s maladaptation to the changing ecosystem needs due to constraints from the incumbent’s (previously successful) business model. The interplay between the incumbent’s framing and the maladaptation of its business model to the changing ecosystem needs constitutes a vicious framing-maladaptation cycle that sustains the increasing abandonment of the incumbent’s business model.

Figure 14 Process model of ecosystem-level business-model disruption dynamics.

Source: Reference Snihur, Thomas and BurgelmanSnihur et al. (2018: 1292)

The study of Salesforce’s framing-adaptation cycle suggests that ecosystem-level interactions play an important role in better capturing business-model disruption dynamics. In doing so, it extends the findings of earlier research that has examined the role of customers (Reference Christensen and RosenbloomChristensen & Rosenbloom, 1995) and partners (Reference AdnerAdner, 2017) by focusing intensely on the interactions of various actors (Reference Gray, Purdy and AnsariGray et al., 2015) during ecosystem evolution. The data suggest that the feed forward and feedback for the disruptor (Figure 14) were amplified by the media, which produced additional raw material, such as various awards, that the disruptor subsequently leveraged to substantiate its leadership.

Researchers have begun to point out the role of external actors, such as media (Reference Pollock and RindovaPollock & Rindova, 2003; Reference RaoRao, 1994) or analysts (Reference Beunza and GarudBeunza & Garud, 2007), as contributors to legitimizing new business models and the new ecosystems formed around them. Their role is important because media and analysts can be “active builders of frames, rather than passive classifiers” (Reference Beunza and GarudBeunza & Garud, 2007: 22). In our case, it is the disruptor’s ability to sustain positive media attention during its early years that might have fueled the dynamics of business-model disruption. Explicitly including these actors in the business-model disruption process model (Figure 14) helps us understand why the disruptor can initiate a virtuous cycle that is difficult for the incumbent to reverse due to the interaction with a variety of actors.

Our process model of evolving ecosystem dynamics (Figure 14) also highlights the importance of framing – effective on the part of the disruptor, ineffective on the part of the incumbent – as an important force in shaping ecosystems in addition to patents or licensing fees (Reference GawerGawer, 2014), standardization of procedures and interfaces (Reference Wareham, Fox and Cano GinerWareham et al., 2014), the oversight of the quality and number of complementors (Reference Mantovani and Ruiz-AlisedaMantovani & Ruiz-Aliseda, 2016), or dynamic control (Reference Dattée, Alexy and AutioDattée et al., 2018). Yet, the framing process contains the seeds of its own destruction: strong framing helps the disruptor shape the ecosystem around its business-model innovation, but over time it might degenerate into a vicious cycle if not accompanied by continuous business-model adaptation by the ecosystem-central firm. For instance, without constant adaptation, firms with central roles in the ecosystem can find themselves diminished or displaced by other ecosystem participants (Reference Adner and LiebermanAdner & Lieberman, 2021). This suggests a somewhat different view of the disruption process than, for instance, the insights of Reference Anderson and TushmanAnderson and Tushman (1990) on technology cycles, which emphasize the importance of the emergence of a dominant design and awareness of technological discontinuities rather than a focus on continuous adaptation and progressive ecosystem shifts.

3.5 Conclusion

This section shows that the B–B process model continues to serve as a useful tool to conceptualize patterns of strategic managerial activities associated with the evolution of large, complex organizations as well as modern business ecosystems. In the evolution of such organizations, new forms labeled as “nonhierarchical,” “network,” “boundaryless,” “virtual,” and the like have emerged. Future research based on questions derived from the B–B process model may help shed light on the managerial activities that constitute strategy-making in these and other evolving novel organizational forms.

Analysis based on the B–B process model also suggests that in adaptive organizations, resource allocation and reallocation, and competence deployment and redeployment, are effectively governed by internal selection processes that involve the combined activities of differentially positioned managers. This is why evolutionary theories of the firm must take strategy-making processes seriously, and why resource-based views of the firm must take managerial action seriously. In light of this, network theory and research (Reference BurtBurt, 1992; Reference Hutt, Reingen and RonchettoHutt et al., 1988), for instance, could help measure strategic context determination and dissolution in terms of growing, or respectively, declining, networks of supporting senior executives in the company over time.

As the recent extension to the ecosystem context illustrates (Reference Snihur, Thomas and BurgelmanSnihur et al., 2018), the B–B process model is a potentially useful tool for extending theory about the role of internal selection in organizational evolution to different units and contexts of analysis, such as business models, networks, or ecosystems, where we are still at the start of academic inquiry into the underlying processes and activities (Reference Shipilov and GawerShipilov & Gawer, 2020; Reference Snihur and BockenSnihur & Bocken, 2022; Reference Snihur and EisenhardtSnihur & Eisenhardt, 2022; Reference Thomas, Autio and GannThomas et al., 2022).

4.1 The CEO and Organizational Evolution

To understand the role and impact of CEO strategic leadership in organization evolution, we begin with the basic idea that no CEO starts with a clean slate and instead must deal with the strategy-making legacy and unresolved strategic challenges of their predecessors. During our study of organizational evolution at HP, we found that successive CEOs shaped the internal selection environment to support the execution of their strategy-making and secure organizational evolution. Specifically, we uncovered three interacting CEO strategy-making practices in service of organizational evolution: (1) harnessing the organizational past; (2) differentially focusing on organizational fit with the current growth trajectories and associated ecosystems and the need (or not) for evolvability to an extended or new growth trajectory and associated ecosystem; and (3) driving of the organizational future with new strategic intent. The strategic leadership of successive CEOs helped shape HP’s organizational evolution, as manifested in changes in the firm’s business-model portfolio and distinctive competence over time. These insights are reflected in the framework presented in Figure 15, abstracted from the HP case below, showing CEO-level and organization-level dynamics. The connecting arrows reflect the intricate interplays between them.

Figure 15 Framework of CEO strategic leadership practices in relation to organizational evolution.

Source: Authors

Each CEO brought with them the experience and track record (at HP or elsewhere) that informed their view of which parts of the past would be important to keep for the future. Each CEO also brought a strategy for how to capitalize on the company’s existing growth trajectories (fit) and/or develop new ones (evolvability), and how to sustain the company’s process of evolution through the changes they intended to make in its existing internal selection environment.

At the CEO level of analysis, the above framework indicates how the differential focusing on fit and evolvability practice forms the bridge between the practices of harnessing the past and driving the future during each CEO’s tenure, and that all three practices interact. In other words, some CEOs start with a clear strategic intent about driving the future that provides insight into how they need to harness the past and differentially change the focus on fit and evolvability. Others start by harnessing the past, and differentially change the focus on fit or evolvability to drive the future.

Inter-CEO vicarious learning helps each successive CEO to exert strategic leadership through the three interacting strategy-making practices and creates interfaces between successive CEOs. The strategic actions of the preceding CEO enable inter-CEO vicarious learning by the succeeding CEO, who learns from the strategic action (or inaction) of their predecessor while developing their own strategic leadership. This inter-CEO-tenure learning is important for CEOs – particularly those who are new to a firm and lack the benefit of insider knowledge. Inter-CEO learning might help CEO adjust more quickly to job requirements in the contemporary times of contracting CEO tenures.

Inter-CEO vicarious learning contrasts with the intra-tenure CEO learning highlighted in previous research (Reference BurgelmanBurgelman, 1988; Reference Henderson, Miller and HambrickHenderson et al., 2006). While existing literature suggests the possibility of inter-CEO learning with executive turnover (Reference Virany, Tushman and RomanelliVirany et al., 1992), it does not systematically document its content. This may be partly because inter-CEO vicarious learning turns out to be a relatively weak process. Our findings indicate that it was not easy for new CEOs to grasp the strategic situation at a company as large and complex as HP. CEO Whitman, for instance, pointed out how difficult it was for her to understand the situation she inherited, despite being an HP board member. Nevertheless, studying inter-CEO vicarious learning sheds additional light on the extent to which successive CEOs read the not-blank slate inherited from their predecessors to inform their own strategic leadership to support continued organizational evolution.

At the organizational level of analysis, the framework shows how the changes to the internal selection environment made by the new CEO move the organization from what it had been at the time of CEO transition to what the new CEO wants it to become. To ensure that the internal selection environment served this execution purpose, successive HP CEOs made significant changes to it. These included: (1) centralizing or decentralizing strategic decision-making; (2) allocating resources to different existing and new businesses in the corporate portfolio; (3) emphasis on cost leadership versus spending on R&D for growth; and (4) altering the composition of the top and senior management team to secure alignment with the CEO’s strategic intent.

The framework above shows the cycle of CEO-level and organization-level processes of strategy-making and organizational evolution during the tenure of a single CEO. Extending the model into an inter-CEO tenure model turns “What we become” for the incumbent CEO into “What we have been and are now,” and therefore the object of the “harnessing the past” practice for the succeeding CEO, thereby starting a new CEO-tenure cycle of strategy-making and organizational evolution. Across these successive cycles, and the strategic legacies they create, each successive CEO challenges the past and envisages the future of the organization in the evolving ecosystem. They use these insights to harness the past pattern of privileging fit or evolvability in resource allocation and drive a new resource allocation pattern that affirms or changes the past one. These within-cycle interplays of CEO strategizing drive the successive modes of adaptation that shape the organization’s evolution.

To better reflect these evolutionary dynamics, we extrapolate the frameworks for each individual CEO to a framework incorporating successive CEOs, depicted in Figure 16. This extended framework depicts CEO strategic leadership, incorporating the strategy diamond of each successive CEO (middle of the figure, representing Tool I) as mediating between the evolving ecosystem (top) and the internal ecology of strategy-making (bottom). It reiterates that no CEO starts from a clean slate, but rather must deal with the strategic legacy and unresolved strategic challenges of their predecessors (“harness the past”). That legacy significantly determines the new CEO’s initial degree of freedom to develop a corporate strategy to drive the future and continue the process of organizational evolution. Each CEO, in turn, leaves unresolved challenges for their successors.

Figure 16 Framework of successive CEO strategic leadership.

Source: Adapted from Reference Burgelman, McKinney and MezaBurgelman et al. (2017: 357)

We represent the interplay between CEO tenures symbolically by showing the overlapping internal ecology, represented by the strategy diamonds (as derived from Tool I, introduced in Section 2) of successive CEOs. After reckoning with the legacy of their predecessors and opening the future with their own announced strategy, CEOs then close the future through the strategic actions that they take – or fail to take. These strategic actions then become part of the past that the succeeding CEO will have to harness.

Anticipating research into the role of CEO strategic leadership in the process of organizational evolution more than sixty years ago, Philip Selznick distinguished between the “organization” aspect (an expendable rational instrument to do a job) and the “institution” aspect (a responsive, adaptive organism) of social associations (Reference SelznickSelznick, 1957: 5–6). Consequently, he also distinguished human interaction-oriented interpersonal leadership from policy-oriented institutional leadership.

Reference SelznickSelznick’s (1957) seminal analysis of the nature of institutional leadership foreshadowed the important role of CEO strategic leadership for built-to-become organizations. While strategic management researchers have provided evidence of the impact of CEOs on firm innovation and performance, it remains less clear how successive CEO strategic leadership maintains or changes the distinctive character of an organization and its longevity. Our exploratory study of this important question in relation to HP’s evolution indicates that the strategic leadership of founders and successive CEOs is fundamental to sustaining the process of organizational evolution.

Interestingly, highlighting the role of strategic leadership in sustaining the process of organizational evolution also sheds additional light on the tension highlighted by Reference SelznickSelznick (1957) between instrumental and institutional views of an organization. This tension derives from the reality that a company founded with the intent of becoming a long-lived institution depends on its successive CEOs to secure the adaptive capacity to survive external context dynamics. A strong instrumental orientation, therefore, is necessary to support a lasting institutional orientation. David Packard clearly understood the strategic importance of adopting such an unsentimental perspective if HP was to remain successful in the unforgiving world of business. His efforts to continue to guide his CEO successors informed by this understanding is his lasting contribution to HP’s process of evolution.

Our study of HP CEOs also highlights a divergence between the fates of CEOs and organization-level evolution. While on the personal level, managing HP was rather difficult for successive CEOs, who were often fired or forced to retire, they collectively contributed to HP’s longevity. This suggests an underlying conflict between the forces associated with leaders’ agency and those associated with determinism and inertia.

4.2 The CEO and Organizational Splits

Long-lived multibusiness firms are complex systems that often englobe distinct business models that involve different ways to create and capture value, implying differences across value-chain activities, customers, and partners. The important question for corporate leaders remains how to correctly determine when the business-model portfolio might need to be reconfigured – for instance, due to changes in the business ecosystem where the firm operates. For instance, changes in the business ecosystem can be leveraged to facilitate the launch of new business models (Reference Hou, Cui and ShiHou et al., 2020) or can encourage market entry by competitors (Reference Snihur, Zott and AmitSnihur et al., 2021). To help answer this question and based on our analysis of HP, we have developed another framework that helps to elucidate the evolutionary forces driving the CEO’s decision to make a corporate split – that is, to transform the firm into multiple independent, smaller, but more adaptive public companies.

A corporate split results in speciation (Reference Garnsey, Lorenzoni and FerrianiGarnsey et al., 2008; Reference Lumsden, Singh and JitendraLumsden & Singh, 1990; Reference Singh and LumsdenSingh & Lumsden, 1990), or the creation of several new viable corporate entities from one originating corporation. Speciation, rooted in biology, examines the forces that drive previously integral populations to differentially split into new viable ones as the environment, including technological requirements, changes. Our data suggest that changes occurring in the business ecosystem drove HP’s PCs-and-printing and enterprise business models apart, and help explain why the transformational corporate split was necessary to enable the two new corporate “species” to adapt better to their specific business-ecosystem conditions. We thus suggest that incipient speciation can help a firm break out of the inertia of coevolutionary lock-in (Reference BurgelmanBurgelman, 2002a) and strategically renew itself. Building on the more proactive view of divestitures emerging in corporate strategy (Reference BerryBerry, 2010; Reference Karim and CapronKarim & Capron, 2016), studying splits as speciation allows a generative perspective on divestitures. In this view, the future of all the resulting entities is of concern – rather than conceptualizing divestitures as a way to correct past (over-diversification) mistakes.

We used HP as our primary example given the first author’s sixteen years of longitudinal research on the organizational strategy-making and evolution at the firm. Over time, HP successfully entered the product markets for computers, inkjet and laser printers, and networking, exploiting distinct business models. For example, it used the unit-based model for computers; the razor-and-blade model for printers and accessories (i.e., selling the printer with a low margin, while making money on the cartridges); and the consulting model to satisfy the information technology needs of business customers. These business models differ in terms of how revenue is generated, what value proposition is developed, and which customers and partners are involved (Reference Chesbrough and RosenbloomChesbrough & Rosenbloom, 2002). Corporate strategy-making in multibusiness firms involves decisions about managing such, more or less well-integrated, business-model portfolios (Reference Sabatier, Mangematin and RousselleSabatier et al., 2010; Reference Snihur and TarzijanSnihur & Tarzijan, 2018).

Our analysis of the HP split, announced in 2014 and executed in 2015 by its then-CEO Meg Whitman, suggests that two dynamic strategic forces propelled by external and internal developments shaped her decision to split based on the increasing cost (driven by rising complexity) and the decreasing benefit (driven by decreasing complementarity) of an integrated business-model portfolio. The exogenous evolution of the business ecosystem, typified by the increasing adoption of mobile communication, social networking, and cloud computing, as well as the increasing speed of execution demanded by customers (i.e., force multipliers), capital-market pressure, and the endogenous lateness of HP in responding to these changes, contributed to the timing of the split in the mid-2010s, when the CEO recognized and acted upon this pressure to split.

This framework also connects to the Figure 14 from the previous Section 3.4 on disruption as it suggests how an incumbent firm, such as HP, which managed several business models in its portfolio at the time, might react to the evolving ecosystem by implementing a radical strategy of portfolio split to preserve its evolvability and maintain ecosystem leadership rather than undergo maladaptation (as observed in the case of Siebel, see Reference Snihur, Thomas and BurgelmanSnihur et al., 2018).

The framework in Figure 17 illustrates two drivers of diverging business models at the organization that influenced CEO strategy-making: decreasing inter-business-model complementarity and increasing intra-business-model complexity. Both drivers are impacted by the evolving business ecosystem.

Figure 17 Framework of corporate split drivers.

Source: Reference Burgelman, Snihur and ThomasBurgelman et al. (2022b: 17)

The first dimension, inter-business-model complementarity, refers to the degree to which respective products, customers, partners, and underlying technology can benefit from the presence of additional business models in the corporate portfolio. In other words, the existence of inter-business-model complementarity suggests that there are collaboration opportunities between the firm’s business models, and synergies can result from exploiting these existing complementarities. As an example, at HP, similar components were used for PCs and enterprise servers, allowing back-office collaboration in sourcing between the two business models and stronger bargaining power vis-à-vis the firm’s suppliers, enabled by HP’s integrated operation of two different business models in its portfolio.

Increasing inter-business-model complementarity is a useful strategic force to understand the success of today’s diversified firms, such as Google or Amazon. Digitalization and platformization, combined with artificial intelligence (AI) and machine learning (ML) from large swathes of data, create opportunities for new business models and cross-boundary disruptions to augment product- or component-derived complementarities, which were the main source of complementarities in the past. These technological advances extend corporate strategic leadership capability to effectively support an enlarged business-model portfolio by augmenting inter-business-model complementarity. However, inter-business-model complementarity can also decrease – for instance, due to the bifurcation of evolving business ecosystems. This happened in the case of HP when personal and enterprise computing began to bifurcate in late 2000s.

The second dimension in our framework is intra-business-model complexity. In line with Reference SimonSimon’s (1962: 468) definition of complexity in a system as “a large number of parts that interact in a non-simple way,” intra-business-model complexity is generated by a high number of activities, customers, and partners; technological sophistication; and the interdependencies between them within business models. Intra-business-model complexity makes high demands on managerial information processing and attention capabilities. For instance, at HP, the shift of computing to the cloud in late 2000s significantly increased the complexity of activities and the number of partners involved in the firm’s enterprise consulting business model.

When inter-business-model complementarity is high and intra-business-model complexity is low, there is no need to reconfigure the corporate portfolio and the company is better keeping all its constitutive parts “together.” As inter-business-model complementarity decreases (while intra-business-model complexity stays low), pursuing strategic integration becomes harder because executives, managers, and employees within the company find it more difficult to define jointly attractive cross-business opportunities.

For example, in the case of HP, this meant that collaboration opportunities between the PC, printing, and enterprise consulting business models became harder for corporate managers to spot and quickly execute. In the medium to long term, this situation jeopardizes corporate survival chances due to the lack of renewal opportunities in an ever-evolving environment, particularly a highly dynamic one faced by HP. This lack of collaboration can eventually result in divestment, just as IBM successively divested its printer, PC, and semiconductor businesses. Other recent examples of splits include GE, GSK, Johnson & Johnson, Siemens, and Toshiba.

As intra-business-model complexity increases (while inter-business-model complementarity stays high), the evaluation of new opportunities becomes challenging. This is because it becomes more difficult for executives, managers, and employees to understand, support, and evaluate multiple complex business strategies and associated resource requirements, particularly regarding the extension of corporate strategy to new technological frontiers of their product markets. This difficulty is due to executives’ bounded rationality and imagination, and computational problems related to cognitively and behaviorally dealing with increasing complexity.

In the HP case, this meant that spotting and implementing radical innovative technologies, such as 3D printing for the printing business-model or cloud-based solutions for the Internet of Things within the enterprise consulting business model, became more challenging under conditions of increasing intra-business-model complexity. High intra-business-model complexity also provides a potentially helpful explanation for Google’s and Microsoft’s previously unsuccessful efforts to develop health-care-related businesses, as those were characterized by extreme complexity. In the medium to long term, this situation jeopardizes corporate survival due to reduced corporate renewal when the organization is no longer capable of processing external signals about its environment and adapting to evolving circumstances.

Simultaneously decreasing inter-business-model complementarity and increasing intra-business-model complexity within a business-model portfolio therefore impose limits on corporate strategic leadership’s capability to keep the organization whole and well-functioning. Such simultaneous and interacting forces increase internal pressure (from managers and employees) and external pressure (from capital markets and investors) to formalize the diverging business models through a decision to scale down – that is, split the company into independent and smaller constitutive parts. Such a decision is most often undertaken by the top strategic leader: the organization’s CEO.

In the case of HP, the exogenous evolution and growth of cloud computing and mobile communications, and the endogenous lateness of HP in recognizing the importance of these changes, contributed to decreasing inter-business-model complementarity and increasing intra-business-model complexity in its business-model portfolio, intensifying pressure to reconfigure the firm’s portfolio by splitting the company. The effects of the two forces are not simply additive; they also amplify each other, as increasing complexity makes it much harder to leverage complementarity benefits.

Our framework suggests that some CEOs might be better at detecting when the timing for the split is right. In the case of Whitman, she was an outsider with less emotional commitment to keeping HP together and an excellent strategist. Existing research has criticized short-tenured outsider CEOs for underappreciating interdependencies in legacy business divestitures (Reference FeldmanFeldman, 2014). We suggest, however, that skillful outsider CEOs might be ideally qualified for undertaking corporate splits, as they can interpret and articulate organizational business-model divergence and ecosystem bifurcation and are the ultimate decision-makers activating the corporate split.

Our split framework also suggests that a company’s size per se is a static concept and not necessarily always the source of economies of scope or the best way to evaluate company performance and prospects, especially in cases when the company manages a portfolio of diverse business models. Applying this insight to the case of HP leads to the conclusion that, at certain points in the process of organizational evolution, it may be better to reduce the size of the company than try to increase it.

We term this problem organizational fragility, defined as a greater downside than upside potential associated with corporate scale and scope. Organizational fragility at HP meant greater negative than positive potential in its adaptive capacity associated with the diverging adaptive requirements of its PC, printing, and enterprise consulting business models. This organizational fragility arose because top management’s ability to manage the diverging business-model portfolio was increasingly thinly stretched. By 2015 the strategic integration challenges had outgrown HP’s corporate-level strategic leadership capability, and the company was now pursuing overambitious strategic integration efforts that would drive it inexorably toward reduced adaptive capacity. After the HP split was finalized in 2015, both the resulting companies were half the size of the presplit HP, which fostered a significantly narrower corporate strategy at each firm and provided clearer and speedier ways to pursue new opportunities.

Our evolutionary approach suggests that there are moments during the corporate trajectory when it becomes clear that the company is being overambitious about the strategic integration of diverging business models within its portfolio. This can exacerbate organizational fragility, where the upside from integrating different business models in a corporate portfolio is minimal, threatening long-term survival. At times like these, adaptive portfolio reconfiguration might be warranted, in order to reduce organizational fragility and enable renewal through radically splitting the organization into smaller entities.

4.3 Conclusion

The analysis offered in this section suggests that strategy-making in organizations should be viewed as an open-ended evolutionary process that does not depend on the ex ante teleological vision of any given corporate leader. To sustain the company’s evolution, each successive CEO must be able to act as an agent who links the company’s close-ended past and its open-ended future. This, in turn, requires analyzing how the internal ecology of strategy-making, driven by the strategic leadership of successive CEOs in relation to the company’s evolving ecosystem, shapes the composition of the company’s business-model portfolio and its various induced or autonomous strategic initiatives.Footnote ⁸

Considering this, our analysis also revealed that a particular CEO in that evolving process may recognize that the company has reached a point of singularity at which it is clear that the company’s ecosystem and business-model portfolio have inexorably bifurcated, and that it is no longer possible to determine a viable corporate strategy for the company. This then requires a corporate portfolio split and the “speciation” of smaller, independent new entities, each more adaptive to its differentiated ecosystem.

5.1 Foundational Conceptual Frameworks of Strategy-Making

Perhaps the most important contribution of the evolutionary perspective consists in showing that the internal ecology of strategy-making – and the associated internal selection environment – is an emergent property of organizations (Section 2). This emergent aspect depends on strategy-making as featured particularly in the process models of ICV, SBE, and CVC (Section 3). These process models also explain how novelty can emerge and how organizations can renew and adapt from within. Furthermore, the research findings show that this strategic renewal process may be strongly influenced by the approach of the CEO and top management team during any period of the company’s evolution (Section 4). The various tools presented in the preceding sections are valuable for managerial practice to develop firm strategy at the CEO level, as well as to understand and participate in strategy-making at the operational and middle-management levels of the organization.

5.1.3 Guided Evolution

Reference Lovas and GhoshalLovas and Ghoshal (2000) reported a case study of a Danish company that suggested to the authors the possibility of organizational evolution “guided” by top management. In some ways, guided evolution suggests the possibility of inducing autonomous strategic action, which brings it back more closely to the traditional rational-actor model of strategy-making. Since we define autonomous strategic initiatives in terms of their difference from the product-market and competency categories encompassed by the corporate strategy, top management can of course initiate autonomous strategic initiatives. However, as the study of Andy Grove’s strategic leadership at Intel indicates (Reference BurgelmanBurgelman, 2002a), top management is likely to impose the logic of the existing corporate strategy on such autonomous initiatives (inertia I), or to become rapidly impatient and emaciate the strategic context determination process (inertia II). Figure 18 illustrates these two forms of strategic inertia, building on Tool II from Figure 2.

Figure 18 Inertial effects of a strategy vector on the internal ecology of strategy-making.

Source: Adapted from Reference BurgelmanBurgelman (2002a)

Also, top management may “forget” that the initiative emerged from below and claim ownership of its initiation if it is successful retrospectively (Reference BurgelmanBurgelman, 2016). For instance, whereas Steve Jobs reportedly told his biographer that he asked his team to come up with a multi-touch screen, Jonathan Ive, the team leader, had a different memory of the development:

He said his design team had already been working on a multi-touch input that was developed for the trackpads of Apple’s MacBook Pro, and they were experimenting with ways to transfer that capability to a computer screen. They used a projector to show on a wall what it would look like. “This is going to change everything,” Ive told his team. But he was careful not to show it to Jobs right away, especially since his people were working on it in their spare time and he didn’t want to squash their enthusiasm. “Because Steve is so quick to give an opinion, I don’t show him stuff in front of other people,” Ive recalled. “He might say, ‘This is shit,’ and snuff the idea. (…).”

(Reference IsaacsonIsaacson, 2011: 468)

5.2 Methodological and Theoretical Implications

Our longitudinal grounded research methods (discussed in the Preface) have generated conceptual frameworks that stand between the narratives of historical science and the mathematical/statistical models of reductionist social science (Reference BurgelmanBurgelman, 2011; Reference GaddisGaddis, 2002). We have shown that such frameworks are useful to conceptualize substantive areas of strategy-making in large, complex organizations that involve multiple levels of management both sequentially and simultaneously. They help “identify the relevant variables and the questions that the user must answer in order to develop conclusions” (Reference Spender, Kraaijenbrink, Huggins and IzushiSpender & Kraaijenbrink, 2011: 52). However, they cannot resolve the uncertainty – the “objective ignorance” (Reference Kahneman, Sibony and SunsteinKahneman et al., 2021: 144–147) – associated with strategic issues and the challenges facing strategy-making. Nevertheless, if “models do better than people, but not by much … Still better is good, and models are better” (Reference Kahneman, Sibony and SunsteinKahneman et al., 2021: 147), we suggest that the value of developing frameworks based on longitudinal grounded research should be assessed, at least in part, in terms of their ability to help other scholars develop better models in further research (see also Reference FiresteinFirestein, 2016) and advance process research in strategic management (Reference Kouamé and LangleyKouamé & Langley, 2018).

5.3 Future Research Directions

The insights produced by the evolutionary lens link the theory of strategy-making more closely to evolutionary organization theory and provide several avenues for cumulative knowledge development. Further research, for instance, could focus on the characteristics of the internal selection environment that are associated with high adaptiveness – or, conversely, engender inertial tendencies in established companies. Future studies could also focus on the role of strategic context determination and dissolution processes in helping established companies cope with internal and external variation and associated dilemmas identified by research in organizational ecology and the economic theory of the firm. Further study of the leadership activities that constitute these processes, and of the external and internal forces that shape them, offers the prospect of adding meaningfully to the knowledge base that can strengthen the strategic leadership capability of large, complex organizations. Additionally, scholars have only just begun to examine the intricate processes involved in ecosystem bifurcation and corporate splits, and their implications for organizations and their strategy-making. Increasing dynamism, regulatory, climate, and technological changes suggest a need for additional research in this direction, as many organizations need to adapt to ecosystem evolution and pressing environmental concerns.

Our work has implications for research on the role of CEO strategic leadership in evolving organizations. First, we identify successive CEOs as a novel unit of observation for organization scholars, suggesting a need for additional studies examining CEO populations rather than only the impact of individual CEOs. Second, as noted earlier, Reference SelznickSelznick’s (1957) seminal analysis of the nature of institutional leadership foreshadowed the important role of CEO strategic leadership for built-to-become organizations. While strategic management researchers have provided evidence of the impact of CEOs on firm innovation and performance, it is not yet clear how successive CEO strategic leadership maintains or changes the distinctive character of an organization and sustains its longevity. Our exploratory study of this important question in relation to HP’s evolution indicates that the strategic leadership of founders and successive CEOs is fundamental to sustaining the process of organizational evolution. However, future comparative research into the sustaining and nonsustaining processes of organizational evolution is necessary to gain further insight into the consequences of interfaces between CEO legacy and the efforts of new CEOs to renew the organization and sustain its evolution. For instance, our research points to the important role of inter-CEO vicarious learning, and further research could explore the factors that determine the relative strength of this inter-CEO learning process, and whether and how it could be augmented – for instance, through machine-generated learning algorithms.

Our work also highlights a divergence between the fates of CEOs and organization-level evolution. While managing HP was a tough personal challenge for successive CEOs, together they contributed to HP’s longevity. This suggests an underlying conflict between forces associated with leadership and forces associated with determinism and inertia. More research focused on examining the interface between individual-level leadership, CEOs’ career imperatives, and evolutionary organization-level challenges would be valuable.

Business models offer an additional level of analysis to help focus on organizational learning problems (e.g., due to increasing complexity when running several business models in parallel), as is also highlighted in such recent approaches as the lean start-up (Reference Bocken and SnihurBocken & Snihur, 2020; Reference RiesRies, 2011) or the strong focus on learning in the development of AI. How organizations can learn well and quickly – for instance, using experiments and algorithms in real time, facilitated by digitalization and the increasing ease of data collection – remains an important problem to ensure organizational adaptation and longevity in turbulent environments.

We hope that the tools and frameworks set out in this Element provide a firm foundation for further development of the evolutionary theory of organizational strategy-making, particularly considering strategy-making in this process.