Emergence in marketing: an institutional and ecosystem framework

There are a number of contentious issues in the emergence literature. Here we will briefly introduce two of the most foundational ones: (1) whether emergence is a fundamentally ontological or epistemological phenomenon or whether there are two distinct types of emergence, and (2) the legitimacy of the concepts of “upward causation” and “downward causation.” We do this to clarify the position taken in this article, rather than for the purpose of conducting a deep exploration of the issues.

The question of ontological versus epistemological emergence is essentially concerned with whether emergent properties are truly novel and unpredictable or whether their existence, at least in principle, could be reduced to the constituent elements and some set of fundamental laws and therefore predicted, provided we had sufficient knowledge and information (Alexander, 1920). Epistemological emergence implies a reductionist explanation, in which emergent properties are novel only at the level of description and understanding (Silberstein, 2006). In other words, it is our lack of knowledge of them that causes them to appear to be new—for instance, the idea that chemical phenomena could be fully explained by the laws of physics, though at present we might not fully know how to do that. This implies, in turn, that the causal capacities—the ability to influence other elements or systems—of an emergent element can be wholly determined by the intrinsic properties of the system from which they emerged (Clayton, 2006). Pure epistemological interpretations of many human and social phenomena have been particularly problematic. For example, the difficulty of explaining the phenomenon of consciousness, based solely on physiological processes (Bedau, 2008; Chalmers, 2006; Clayton, 2006; Ellis, 2006; McLaughlin, 2008), has been noted as far back as Aristotle. Likewise, academic marketing would be hard-pressed to offer a comprehensive explanation of customer behavior based solely on a psychological stimulus-response model.

Conversely, ontological emergence implies that emergent properties are, in fact, truly novel and not reducible to the intrinsic properties of the constituent elements of the systems from which they emerge because “… complex systems are too complex to be explained by reductionistic practices” (Kauffman, 2007, p. 907). It follows that ontological emergentists see emergent phenomena to be causally empowered (Clayton, 2006; Goldstein, 1999)—capable of acting on and interacting with other elements—beyond what is directly attributable to the elements of the systems from which they emerged. Thus, emergent phenomena can interact with the other constituent elements of the system from which they emerged. This relational interaction condition is a fundamental principle of ontological views of emergence (DeLanda, 2006).

It is also possible to view epistemological and ontological emergence as separate types, related to phenomena in systems with different degrees of complexity. That is, there might be both weak (or epistemological) and strong (or ontological) emergence (Capra & Luisi, 2014; Kaufman and Clayton, 2006). We generally assent to this orientation but maintain that, at least in relation to the complex systems which are of interest in academic marketing, ontological emergence should be assumed. First, this is because it is unrealistic to expect that we can conceptualize, measure and compute all of the variables needed to be able to predict the outcome of all possible inter-relationships that might account for emergent properties in complex marketing systems (Davies, 2006), even if in principle that were possible. Second, it is hard to imagine that a phenomenon, such as an instance of customer experience, could be reduced to contextual and physiological variables alone. Finally, the very act of observing complex systems for the purpose of predicting future states changes the nature of the future state (Deacon, 2006; Ellis, 2006; Goldstein, 1999; Goldstein, 2000). Therefore, attempts at prediction lead to indeterminacy.

Emergence is also often characterized by the two related processes of upward and downward causation. Upward causation, or supervenience (Silberstein & McGeeve, 1999), means that the emergent, higher-order properties depend on (emerge from) the properties and interactions of their lower order constituent elements (McLaughlin, 2008). This implies different levels of organization in which wholes at one level (i.e., customers) function as parts at the next (and at all higher) levels (e.g., brand community/culture) (Wimsatt, 1994). In the emergence literature, the notion of dependence of higher-order emergent properties on their lower order constituent elements (i.e., supervenience) is relatively unproblematic.

Downward causation concerns the emergent property acting back upon its constituent elements (Davies, 2006) and this relationship has been considered more problematic because both the bottom-up and top-down causal explanations would then compete with each other (Davies, 2006). That is, the higher-level organization would both cause and be caused by the lower level–e.g., culture caused by and causing the individual customers’ behaviors. However, if upward and downward “levels” are understood as levels of complexity within a single system, with parts and wholes mutually constitutive, and if “downward” is understood metaphorically (i.e., such “levels” are analytical perspectives only), the multiple causal explanations no longer compete with each other. This is the approach taken by Silberstein (2006, p. 204), who advocates the notion of systemic causation, based on relatedness and complexity because, “The universe is not ordered as a hierarchy of closed autonomous levels such as atoms, molecules, cells, and the like. Rather, the universe is intrinsically nested and entangled.” As Capra and Luisi (2014, p. 68) similarly state, “in nature there is no ‘above’ and ‘below.’” We concur with these views and propose the interdependent, nested, and overlapping service-ecosystems perspective of S-D logic (as discussed in the next section) as a means of avoiding this problematic conceptualization.

It is important to note that emergence is a systemic phenomenon and critical to the self-organizing processes characteristic of all complex adaptive systems (Corning, 2002; Ladyman & Wiesner, 2020). That is, it is a vital part of the process of adaptation; it is the interplay between emergent phenomena and system structure that provide the adaptive, or self-adjusting, capabilities for these complex adaptive systems. In social systems, these structures are usually captured as institutions or institutional arrangements (sometimes called “institutional logics”), the essential role of which is increasingly becoming accepted and understood in academic marketing (e.g., Dolbec & Fischer, 2015; Humphreys, 2010; Vargo & Lusch, 2016). Self-organization (also referred to as autopoiesis, Ladyman & Wiesner, 2020; Merali & Allen, 2011) also involves positive and negative feedback (Arthur, 2015; Deacon, 2006), as will be discussed in a later section.

Complex adaptive systems can be seen at various levels of scale, such as molecular, cellular, organ, and bodily systems in biological systems; or customers and firms, “industries” and markets, and general social systems in marketing systems. It is also important to emphasize that emergent properties are not just outcomes of the systems from which they emerge; they become essential constituents in the functioning of the new whole.

Furthermore, the impact of emergent properties is not restricted to the system from which they emerged. They become what Kauffman (1996) calls the adjacent possible—new structural elements that can also be appropriated by other systems. This process is known as exaptation (Dew et al., 2004), the co-opting of an element with a role in a system for a role (similar or different) in another system. For example, in technological innovation, the exaptation of electrical charges in natural systems takes place when it is applied to lighting and communication technology. Arthur (2009) essentially argues that all technological advancement occurs through this process of exaptation, which he calls combinatorial evolution—fundamentally, newness always emerges from what is. In the following sections, we further explore the concept, process, and implications of emergence in the context of markets and marketing. To do this, we use an S-D logic theoretical framework.

First-order emergence accounts for the appearance of novel outcomes from ad-hoc resource integration and service exchange. This is illustrated in Fig. 1 by the single upward arrow. Here, ad-hoc resource integration and service exchange generate the essential interactions between actors and resources from which novel outcomes can emerge. First-order emergence, thus, operates through supervenience; the emergent outcome depends on, but differs from, the constituent elements (McLaughlin, 2008; Silberstein & McGeeve, 1999). At the heart of our model, then, is the basic logic of combinatorial evolution (Arthur, 2009). While the same outcome may re-emerge if the actors, resources, and interactions are the same, there is no additional mechanism to sustain regularity; what emerges may therefore be ephemeral and lack continuity or stability (Sawyer & Sawyer, 2005). By itself, then, first-order emergence is not capable of generating higher-level, complex structures (Ellis, 2006; Martin & Sunley, 2012). For this reason, we simply refer to what emerges at the first-order as novel outcomes.

One example of a novel outcome arising from resource integration and service exchange would be the farmer, Cornelius Dircksen, who in the 1630’s agreed on an ad-hoc basis to use his small skiff to take passengers across the East River in New York in exchange for shells (Bowery Boys History, 2015; Valentine, 1853). It should be noted that the same emergent property could be realized in several ways using different combinations of resources and interactions (Sawyer, 2005). In our example, “transportation over water” could come about in ways other than “Dircksen using his small skiff in exchange for shells,” such as acquiring and using your own vessel or building a bridge. Each of these would draw upon different resources and resource integration processes to co-create value.

The novel outcomes of first-order emergence critically include value—i.e., a change (positive or negative) in the viability of a focal system (Vargo & Lusch, 2014)—but also resourceness of resources. More generally, the interactions required for ad hoc resource integration and service exchange allow actors to cocreate novel outcomes, which may form the basis for repeated value co-creation. As such, the outcomes of first order emergence can be seen as what is sometimes referred to as “bilateral understandings” in classical works on the origins of institutions (e.g., Berger & Luckmann, 1967) and are therefore possible predecessors of institutions, subject to the parallel process of institutionalization.

In summary, first-order emergence accounts for how novel outcomes that depend on, but differ from (i.e., supervene on), the properties and interactions of their constituent elements might arise from ad-hoc resource integration. These outcomes may be either ephemeral or repeated. In accordance with S-D logic, the model posits that novelty is generated through the service-exchange-enabled interactions, rather than through the actions of individual actors or system elements.

Second-order emergence introduces a potential for greater regularity as the emergent property acts back on (interacts with) its constitutive elements (Goldstein, 1999, 2000). This is illustrated in Fig. 1 by the addition of a downward arrow, reflecting feedback (positive or negative) between the novel outcomes and the resource integration and service exchange that brought them forth in the first place. First-order outcomes (e.g., value-creating activities) thus become part of the context in which actors engage in additional resource integration and service exchange. This closes a feedback loop that enables and constrains how emergent properties reoccur, allowing for patterns of actor behavior to emerge that make interactions more ordered (Fromm, 2005; Martin & Sunley, 2012). The result is a form of self-organization through which some first-order outcomes are repeated while others are not.

It is important to stress that first-order emergence is subsumed into second-order emergence. Taken together, supervenience at the first-order and feedback at the second-order of emergence close a causal loop between the actors and the emergent property, which affords systemic causation (Silberstein, 2006). This creates a qualitatively different source of variation based on the unpredictable dynamics of positive and negative feedbacks (Arthur, 1999), which may either reinforce or weaken first-order emergent properties. This allows for the emergence of endogenously generated, recurrent patterns of resource integration and service exchange that are self-reinforcing (Martin & Sunley, 2012). Continuing our example, after some time as an ad hoc ferrier, Dircksen would allegedly drop whatever other chores he was doing when a prospective passenger sounded a horn he had hung in a nearby tree. Here, first-order emergent outcomes came to coordinate further interactions, encouraging iteration (MacIntosh & MacLean, 1999) and the emergence of spontaneous order (cf. Lansing, 2003) through feedback.

Feedback, then, allows for the formation and diffusion of patterns in a service ecosystem. In our example, since an increasing number of people engaged in service-for-service exchanges of “transportation over water for shells” with Dircksen, he decided in 1642 to undertake ferrying fulltime. This led to the identification of Dircksen as the first New York ferryman and with that (“brand”) identity came expectations about his behavior. Through iteration, then, common practices are established that can “… endow participants with cultural capital, produce a repertoire for insider sharing, generate consumption opportunities, evince brand community vitality, and create value” (Schau et al., 2009, p. 30). As part of the process of institutionalization, such a pattern would correspond to what Lawrence et al. (2002, p. 7) refer to as a proto-institution—a narrowly diffused and weakly entrenched “institution in the making.” That pattern may develop into a full-fledged institution (Vargo & Lusch, 2016), but this is not a given; the properties that result from first- and second-order emergence may either support further institutionalization by reducing variation or disrupt institutionalization by generating novelty through alternative outcomes. In our example, Dircksen was successful, but the services offered by other ferrymen who followed his lead proved unreliable, unsafe, and expensive (Bowery Boys History, 2015).

In summary, second-order emergence explains how actor behavior can endogenously generate durable patterns of resource integration and service-for-service exchange that might start to institutionalize (that is, become rules, norms, beliefs, etc.), as indicated in S-D logic (e.g., Vargo & Lusch, 2016). The way actors interact with each other and with the resulting emergent outcomes can generate and reinforce patterns of resource integration and service exchange that result in specific outcomes. As such, second-order emergence also offers an account of how fledgling service ecosystems (markets, marketing channels, brand communities, service platforms, etc.) may emerge among multiple actors engaging in service-for-service exchange.

In systems capable of third-order emergence, actors are able to reproduce their resource integration and service exchange based on emergent patterns. Such actors exhibit a persistence of internal structures or a type of memory (Ladyman et al., 2013; Martin & Sunley, 2012, see the zoomed-in actor illustration at third-order in Fig. 1) that enables pattern recognition. This in turn allows emergent patterns to be reproduced, which is illustrated in Fig. 1 by the addition of a second upward arrow. Driving this reproduction is the anticipation of consequences of specific interactions (Holland, 1992) and the resulting value creation (Barile & Polese, 2010).

As before, the operation of first-order and second-order emergence is subsumed into third-order emergence. The third-order adds the possibility of second-order emergent patterns (e.g., value, resourceness) being recognized as such and reproduced by the actors (Ellis, 2006; Martin & Sunley, 2012; Vargo et al., 2020). While third-order emergence is characterized by pattern recognition and goal-directed behavior (Ellis, 2006), this behavior is reactive only. The patterns of resource integration and service exchange that endure are thus contingent on the nature of the system and its components (Cuff et al., 2007; Kauffman, 1996, 2007). For example, in 1654 the city of New York reacted to the noted problems associated with the increase in providers of ferry services by issuing licenses to ferry operators. This placed specific requirements on the operators, thus standardizing practices within the fledgling industry and contributing to their institutionalization.

The interplay between third-order emergence and the process of institutionalization can generate fully-fledged institutions that coordinate value cocreation. The critical step is the possibility of internalization of emergent structures, which can generate hyper-recurrent behavior (Deacon, 2006). This, in turn, makes those structures appear as “…given, unalterable and self-evident” parts of an external reality (Berger & Luckmann, 1967 p. 59), or in short, as institutionalized service-ecosystem properties (Vargo & Lusch, 2016). For example, by the 1800s the stabilized patterns of ferry transportation in and around New York were perceived as an external and given market by entrepreneurs like Cornelius Vanderbilt. To better compete in this market, they incrementally adjusted their operations by investing in new boats and ferry-technologies (e.g., horse-powered, steam-powered, and double-screw ferryboats). However, all such improvements essentially reproduced the institutionalized solution of ferry transportation.

The internal reproduction of system structure makes resource integration and value co-creation more predictable (Vargo & Lusch, 2016). The combined effect over time of third-order emergence and institutionalization is a growing number of institutional arrangements supporting pattern reproduction over alternative behavior (Vargo et al., 2020). That is, because of the performative effects of institutional arrangements, actions become quasi-predictable (Lusch & Vargo, 2014; Vargo, 2018). With third-order emergence, then, we get the full S-D logic narrative of actors engaging in value cocreation by integrating resources and exchanging service facilitated by endogenously generated institutions and institutional arrangements.

But the third order also introduces a new source of novelty compared to previous orders. It allows actors to (de)select specific actions based on emergent properties, which may induce further (yet unintended) changes in the system. Such adjustments can reinforce existing patterns. For instance, the gradual introduction of ferry services to other destinations around New York from the 1650s onwards established ferrying as a commonly accepted solution for transportation over water. But they can also weaken existing patterns. For instance, a toll on ferrying introduced by the British when they took control over the city in 1664 led to a dramatic decline in regular ferry services as people reverted to private ride-sharing and other ad hoc solutions.

In summary, third-order emergence supports institutionalization of value cocreation by allowing reproduction of observed patterns, but also generates imperfections and unintended consequences in the reproduction of such patterns. Taken together, the first three orders of emergence and the interrelated institutionalization process are enough to account for many marketing-related phenomena, including the formation of markets, the generation and diffusion of innovations, market orientation, the creation of brands, the reactive response to competitors’ pricing, etc. However, to fully account for the complexity of such phenomena, requires a fourth order of emergence.

Out of the infinite number of possible outcomes of the first three orders of emergence, systems may emerge that are capable of intentionality beyond reproduction (Deacon, 2006). To accommodate such systems, we propose a fourth order of emergence at which (some) actors reflexively shape resource integration and service exchange to influence service-ecosystem properties. This is illustrated by the addition of a second downward arrow in Fig. 1, which completes a second causal loop so that emergent service-ecosystem properties can form the basis for, and be the subject of, intentional action. This requires actors with the capacity to envision how their interactions with others affect service-ecosystem properties (Ellis, 2006). Both the emergence literature and S-D logic recognize this characteristic as reflexivity (Ellis, 2006; Kjellberg, 2019; Martin & Sunley, 2012).

In our model, reflexivity denotes actors’ awareness not only of service-ecosystem properties but also of their own role in producing such properties (see the zoomed-in actor illustration at fourth-order in Fig. 1). Such awareness typically results from a combination of human and non-human elements in a service ecosystem (cf. Barile et al., 2018). The resulting reflexivity allows actors to intentionally shape their interactions with others to change (or maintain) service-ecosystem properties, or in short, engage in service-ecosystem design (Vink et al., 2021). For example, in an attempt to address faltering ferry services around 1700, the city of New York decided in 1708 to grant a charter to one ferry operator giving it the privilege to operate ferries across the East River. This was a drastic intentional effort of one actor to modify the service ecosystem by placing all service provision under the control of a single provider.

As before, the lower orders of emergence are subsumed into the fourth-order emergence. In addition, the outcomes, patterns, and service-ecosystem properties that emerge from the first three orders and that generate observable regularities across many marketing phenomena (e.g., the operation of markets, consumer behavior, pricing, advertising) become subject to intentional shaping. This introduces a new dynamic as a result of reflexive actors engaging in parallel but different efforts to change or maintain service-ecosystem properties (Martin & Sunley, 2012). In our historical example, there are competing intentional efforts to shape the transportation service ecosystem based on alternative envisioned designs, such as the construction of railroad bridges across and tunnels under the East and Hudson Rivers. Such reflexive shaping may also elicit direct defensive responses from those who seek to preserve the existing system. In the case of the ferry operators in and around New York, many were in fact acquired by the growing railroad companies, who for a time operated both kinds of services. If we fast forward to present time, the reactions of taxi operators to the disruptions caused by Uber and other ride-sharing platforms are an explicit example of such direct defensive responses. As a result, the changes in the service ecosystem at fourth-order supervene on both aligned and conflicting shaping and non-shaping efforts by multiple actors (Vink et al., 2021).

The interrelation between fourth-order emergence and institutionalization is based on the ability of actors to engage in “… conscious effort[s] to transform situations by challenging existing institutions” (Kjellberg, 2019, p. 469). Fourth-order emergence thus provides the necessary conditions for what Lawrence et al. (2011) term institutional work—the purposeful creation, maintenance, and disruption of institutions. This is nicely exemplified by how Uber disrupted taxi service provision by removing the need for professional drivers and changing the relationship between drivers and riders by allowing them to interact directly. However, Uber also maintained existing market institutions such as payment per distance traveled, rating of drivers and riders, and electronic payment systems (Vargo et al., 2020). The overall effect on a service ecosystem thus results from complex interrelations between the outcomes of lower orders of emergence (e.g., novel outcomes, recurrent patterns, service-ecosystem properties and/or institutions) and parallel reflexive shaping.

In summary, fourth-order emergence results from actors reflexively engaging in interactions with others to intentionally change (or maintain) service-ecosystem properties. Besides being subject to the dynamics of all lower orders of emergence, service ecosystems that exhibit fourth-order emergence are characterized by supervenience of multiple parallel design efforts. Fourth-order emergence thus adds a critical proactive component to our model, aligning it with S-D logic and its insistence on the primacy of effectual processes over predictive processes (Vargo & Lusch, 2014, p. 243).

As conceptualized above, emergence applies (though is not limited) to all situations in which resource integrating actors engage in service-for-service exchange. In other words, emergence permeates essentially all marketing contexts and phenomena. By identifying four distinct orders of emergence that pertain to such settings the conceptualization offered here allows for a more fine-grained analysis of specific instances of emergence in marketing. Below, we discuss some implications for marketing scholarship related to innovation and uncertainty. These implications cut across essentially all marketing concepts.

Embracing emergence and recognizing its interrelations with institutionalization have significant implications for marketing practice. Arguably, none is more fundamental than the need for actors in service ecosystems (marketers, customers, policymakers, etc.) to adopt a drastically different stance regarding certainty and uncertainty in their planning and strategizing. The entrepreneurship literature has usefully argued for the combined use of causal and effectual logics in organizational decision-making. Effectual decision-making allows decision-makers to respond flexibly to changes in their operating environments while causal logic allows them to stay focused and enact strategic plans when regularities permit (Matalamäki, 2017). As we argued above, the latter kind of quasi-predictability is likely to become an even rarer case, given ongoing technological, environmental, social, and political developments and the exponential opportunities for novel outcomes they trigger across all orders of emergence. Hence, the future of marketing is likely to be less rather than more predictable, making an entrepreneurial approach to marketing the general case and managerial approaches applicable only under increasingly rare instances of quasi-predictability (Koskela-Huotari & Vargo, 2018; Vargo & Lusch, 2014).

This shift to a more entrepreneurial approach to marketing, underpinned by emergence, has four implications for marketing practice. First, the interrelations between emergence and institutionalization imply that managers should consider effectuation (Read et al., 2009) as the primary approach for overseeing and strategizing organizational activities. In practice, this requires letting go (to a certain extent) of managerial control and knowledge. For example, managers may need to allow swift action by their teams in specific task environments even if the underlying processes are not fully understood. In parallel, managers need to control for the possible downside risks and constraints by continually assessing possible changes at each order of emergence. In situations in which institutionalization provides quasi-predictability, traditional marketing management approaches, such as segmenting, targeting, and positioning may still be used, at least as short-run tactics. As a complement, managers should strive to develop an in-depth understanding of the goals of the value-cocreating actors, as this will improve their ability to identify possible adaptations in the service ecosystem. Awareness of the specific conditions for emergence at each order and the current levels of institutionalization within the service ecosystem allows managers to navigate quasi-predictable and unpredictable service-ecosystem characteristics and use appropriate tools and practices. The key question becomes: how does one achieve awareness of the current level of unpredictability? This requires continually gauging the relative importance of each order of emergence in generating novelty, given the stabilizing effect of institutionalization. This will help managers create context-specific indicators to determine whether to opt for nonpredictive or traditional marketing strategies or tactics.

Second, managers should rethink and expand environmental scanning. Emergence requires managers to cover both the “verbs” and the “nouns”-that is, both processes and outcomes. Recognizing and making sense of emergence also necessitates engaging the organization at large in environmental scanning to foster an ability to oscillate between different levels of aggregation (Chandler & Vargo, 2011). This is because emergence cannot be fully understood by only focusing on the emergent outcomes or processes at a particular order. Rather, it is important to consider outcomes at all four orders of emergence, from ad hoc service exchanges to explicit attempts at ecosystem redesign. Much of environmental scanning today focuses on aggregating occurrences at the micro-level (e.g., customer preferences) into macro-level patterns (trends). However, emergence does not follow the logic of aggregation, in which summative effects alone can explain more complex phenomena. Instead, novel properties (e.g., macro-level trends) that emerge from the interactions between system elements (e.g., micro-level customer preferences) must also be strengthened or stabilized by subsequent interactions between system elements (i.e., meso level institutionalization) in order to endure. Environmental scanning should therefore strive to link developments at all four orders of emergence. This will include zooming in on micro-level interactions and zooming out to macro-level context, and connecting these to meso-level organization. Here, the meso level (market, service ecosystem, etc.) should be both the starting point and focus of analysis, as both micro- and macro-level changes and impacts emanate from the meso level (Arthur, 2015). As the relevant meso level is actor-specific, environmental scanning should be conducted-or at least coordinated-by the focal organization. External agencies and experts can be used to provide specific insights, but the core of the environmental scanning should not be outsourced. For many organizations this may require re-organizing and investing more resource in strategic market and marketing research.

Third, and related, organizations need to capitalize on emergence in relation to innovation. As emergence is always “on,” no managerial effort is needed to initiate or foster it. However, traditional business processes and organizational structures are ill-equipped to capitalize on the fact that innovations may result from all four orders of emergence. Thus, managers need to expand their definitions of innovation to include not only new technology but also, e.g., patterns of resource-integration and service exchange, business models (Wieland et al., 2017), service-ecosystem and market properties (Nenonen et al., 2019, Vargo et al., 2015). Given that emergence is non-linear, managers should also move from time-bound (i.e., annual or quarterly) to continual processes related to strategy, innovation, and business planning – or else the organization runs the risk of being too slow to recognize emergent changes. In practice this could mean moving from annual marketing plans to issues-based marketing strategies. In a similar vein, transitioning to self-organizing teams or otherwise allowing employees high levels of autonomy creates an organization that is better able to deal with change arising at different orders of emergence. However, to be capable of constructively questioning the current status quo of a service ecosystem and envisioning alternative states (fourth-order emergence) requires a vantage point not always available to task-specific teams, especially if these follow functional organizational silos. To fully capitalize on emergence, then, firms need self-organizing teams that span different functions over, for example, the marketing-sales interface or the marketing-R&D interface. Indeed, such teams should not follow current organizational boundaries, which likely reflect previously emergent and institutionalized solutions. Instead, they should involve an expanded network of innovation partners, as implied by concepts such as open innovation, innovation networks or ecosystems. Emergence increases the importance of these notions further, but with a twist: rather than trying to manage suitable innovation partners, firms should encourage others to come to them as “self-selected” innovation companions.

Fourth, the interrelated nature of emergence and institutionalization also has implications for meso-level policy initiatives (by firms and authorities alike). In particular, policymakers should recognize that this relationship creates potential tensions for any policy initiative. While institutions (formal and informal), such as laws, regulations, norms, are necessary for service-ecosystem viability, policymakers must understand that despite however “good” an institution is, emergence is omnipresent. Indeed, the successful creation of an institution also generates new opportunities for emergence at all four orders. It introduces a new resource that can be creatively combined with existing ones to generate novel outcomes (first-order). It establishes a new conduit for positive or negative feedback that can reinforce or weaken existing patterns of service exchange (second-order). It leads some actors to adjust their current behavior to ensure specific outcomes, which in turn can generate new and unforeseen consequences for the system as a whole (third-order). Finally, it can trigger resistance and concerted efforts by (groups of) concerned actors to explicitly design alternatives (fourth-order). Thus, it is usually just a matter of time before new emergent outcomes come into conflict with existing institutions. This is especially true of rigidly constructed institutions. The implication is that policy initiatives should strive to establish codified institutions that allow flexible adjustment over time and in changing contexts. This echoes Callon’s (2009) proposal that a well-functioning market is one capable of taking care of its own overflows. To improve the resilience of service ecosystems at any scale in the face of severe disruption, it is paramount to ensure processes and structures that allow flexible and responsive adjustment of institutions.