Rapid digitalisation of recent decades has propelled a new type of economic phenomenon: the digital platform (DP), which provides a digital infrastructure to facilitate economic, social, cultural, educational, and other forms of interactions. A DP can grow a vast ecosystem of multiple member groups who have incentives to create, exchange, and consume value generated from facilitated interactions. The remarkable success of the digital platform ecosystem (DPE) business model is evident with their rapid dominance across many economic activities. For example, at the time of writing this paper, five of the top ten companies by market capitalisation globally were DPEs including Apple, Microsoft, and Alphabet (the parent company of Google) 1.
Regulators and even business leaders themselves often struggle to find appropriate frameworks to describe, understand, and debate the economics of DPEs and their broader impact on our socio-environmental systems. Vividly illustrating the scale of this challenge is the frequently cited statement associated with Jack Ma, the founder of Alibaba Group, a prominent DPE in China, where he described himself as “a blind man riding on a blind tiger” 2. During the last decades, observations on the apparent parallels between natural ecosystems (NEs) and networked businesses have inspired business leaders to draw insights from ecology to explain the dynamics of DPEs. As a major milestone, Moore extended ecological metaphors to explain the changes in interactions between competing businesses over time3. Another important contribution came from Iansiti and Levien who adapted the notion of 'keystone species’ to the role of platform orchestrators in the value creation of business ecosystems4. Recently, Lianos and Carballa-Smichowski explored network-based indicators such as centrality, motivated by their relevance in ecology, to measure market power5. Such research has contributed insightful terminology and helped to explain some aspects of the complex dynamics underlying DPEs.
However, applications of ecological concepts to DPEs have, thus far, been reduced to casual exploitations of selected metaphors lacking a systematic, guided approach and hence yielding useful, yet limited insights. Conceivably, a deeper and more coherent understanding of the workings of DPEs could be achieved through a comprehensive transfer of concepts from NEs, combining knowledge on evolutionary and ecological dynamics. Such transfers require scientific analogical reasoning founded on systematic links that are due to the shared structures of source (nature) and recipient (digital economy) domains 6.
NEs have long been recognized as complex adaptive system (CAS) 7. The CAS theory stylizes systems as populations of multiple (semi-)autonomous agents engaging in interactions and adapting to the changing environment and behaviour of other agents. Agents in a CAS can pursue different strategies to maximize individual success. Individual organisms in NEs, for example, maximize their fitness, that is their survival, growth, and reproduction rates. Adaptation incorporates the processes of trial-and-error, as well as copying and passing on information – mutation, replication and inheritance in NEs 8. Furthermore, feedback loops, path-dependence, and synergism, non-linearity, self-organisation, and emergent patterns are important properties of CASs, emerging from complex, dynamic, and strategic interactions among their members.
The paper maintains that DPEs possess similar characteristics and thus can be viewed as CASs. Indeed, DPEs build on idiosyncratic interactions among multiple heterogeneous members and member groups. There may be various channels for these interactions, including flows of finances, knowledge or user data. Interactions often form feedback loops among DPE members and the orchestrator, and their complexity distinguishes DPEs from the conventional, linear supply chains. As the DPE business model relies on the already available, often underutilized assets or resources, it can rapidly reorient and expand into new markets, and thus exhibits a high degree of adaptability to dynamic surroundings. A DPE’s path dependence comes from their high reliance on technology and innovation whereby readily available technologies determine the direction of future developments. Last, but not least on this list of characteristics of DPEs is the synergistic creation of value as a collective good by DPE members 9,10.
Viewing DPEs as CASs provides a scientific basis for systematic linking of their attributes with those of NEs. Recognizing that DPEs are embedded in the context of the digital economy while NEs are an inherent part of nature, this research relies on a well-established hierarchy of biological organization to construct a corresponding hierarchical framework for the digital economy which hosts DPEs. The resultant framework organizes key elements and processes within, and surrounding the DPEs based on parallels across the two domains. The framework allows for the flexible yet consistent transfer of ideas, concepts, and methods from ecology and evolutionary biology to the context of the digital economy beyond the scope of metaphors toward theoretically grounded scientific analogies. This will foster a comprehensive ‘eco-logical’ understanding for the digital economy which is also capable of informing the design of efficient regulations to govern DPEs.
Framework to organize ecological thinking for the digital economy and digital platform ecosystems
The dynamics of nature emerges from physical, biological, and biogeochemical processes and phenomena, such as primary production, growth, competition, and nutrient cycles. They together ensure that NEs persist, provide ecosystem services, and a living environment for inhabitants.
NEs and in fact the entire biosphere have been recognized as complex adaptive systems (CASs), wherein higher-level units and patterns emerge from interactions and dynamics occurring at lower levels 7. The living hierarchy of NEs includes cells, tissues, organs, organisms, populations, species, communities, and ecosystems themselves. Cells can live as independent organisms or integrate into clusters such as tissues and organs, organisms form populations of species, which themselves cluster into communities as the living components of ecosystems. The foundation of this living hierarchy lies at the genetic level accommodating genes, the fundamental units of heredity that encode the “instructions” for building and regulating living organisms and play a pivotal role in defining their unique traits and characteristics. On the top of the hierarchy are biomes hosting multiple ecosystems in a particular area and being defined by a set of physical characteristics of the environment such as humidity or intensity of sunlight.
By identifying DPEs as CASs, the paper transfers the principles of biological organization into the domain of the digital economy. The paper focuses on genes (Micro), species (Meso), ecosystems (Macro), and biomes (Mega) as basic levels of nature’s hierarchy, which are the most relevant for building systematic parallels between nature and the digital economy, and propose that processes and entities of the digital economy be structured similarly. For example, products (goods and services) and other entities that are involved in the co-creation of value by the DPE, can be compared to the communities of various species, who, likewise, participate in co-creation of ecosystem services by NEs. Consequently, their co-opetative (cooperative and competitive) interactions form DPEs comparable to ecosystems in nature. Equivalent to genes, technological know-how and business strategy emerge as underlying fundamentals of the DPE’s specific value-creation model. Finally, ecosystems are embedded in the society including economic and other relevant contexts, just like NEs are embedded into their respective biomes. The four levels considered here provide a broad structure through which objects from the digital economy can be related to their counterparts from nature.
As both NEs and DPEs are CASs, interactions between agents at each level and across levels play a crucial role in determining their dynamics and characteristics. Recognizing the importance and universal nature of interaction types, the paper introduces a Meta level, which is comprised of interactions between components at each level, in the digital domain and in the domain of nature, respectively. Together, Micro, Meso, Macro, Mega, and Meta levels allow for the study of digital economies through the lens of their ecological analogies. The resultant approach is referred to as an Eco-evolutionary rooted framework for the digital economy, or the 5M Framework derived from the initials of the major levels throughout which DPEs are compared to NEs in this paper (see Fig. 1).
1. At the Micro Level: Comparing the fundamentals of digital businesses to genes
The survival and thriving of individual agents in a CAS depend on their ability to adapt to a dynamic environment through innovation. In nature, innovations arise from changes in the genetic makeup of individuals of species. Key phenomena which determine evolutionary changes at the species level are variation (arising from processes including mutation and recombination), selection, and inheritance. Innovations in economic or technological systems have already been analogized to the evolutionary processes 11, and the Micro level of the framework seeks to formally organize such comparisons in the context of the digital economy. Genetic elements, whose combined information determines a living organism, can be compared to the underlying fundamentals of the DPE offerings. These fundamentals can include elements of technology, knowledge (including knowledge related to user behaviour), and business strategy. For example, digital streaming services such as Spotify and Netflix emerged from various innovations in smartphone capabilities, high-volume data transfer, digital payments, and the introduction of machine learning to determine user preferences, among other elements.
In nature, each genetic innovation creates a new selection pressure which may prompt further innovations in response, a perpetual contest termed the Red Queen Mechanism 12. For example, organisms continuously develop new defence mechanisms against pathogens or new foraging strategies against competitors, prompting counteractive innovations in response. This regularly escalates into an “evolutionary arms race" which draws resources from both sides. Similarly, successful innovations in competitive economic systems provide strategic advantages, which pressures other agents toward further innovation, often via costly R&D 13. Digitalization and platformization, however, have altered the course of the evolutionary arms race. On the one hand, an open innovation model adopted by many DPEs facilitates innovation within their ecosystems 14, whilst on the other hand, dominating DPEs can suppress nascent innovations through mergers, acquisitions, or extensive competition.
A constant emergence of rich genetic diversity is a pre-requisite for species adaptation. In changing environments, even originally neutral or disadvantageous mutations may become advantageous (the phenomenon of ‘preadaptation’ 15. Myopic ecosystem management may harm the inherent diversity of NEs and thus prove counterproductive for the sustainable provision of ecosystem services in the long run. Monoculture is a well-known example: despite providing greater yield in the short run, genetically homogenous crops do not allow for the replenishment of key soil nutrients and increases susceptibility to pests, diseases, or climate shocks. To maintain yield, farmers apply costly pesticides and fertilisers, which in the long run further degrades the ecosystem 16. Likewise, technological or economic innovations are not always beneficial or profitable at their onset, yet maintained diversity of business fundamentals can indeed serve as the driver of economic growth 17. This is especially true for the digital economy where innovations are constantly being produced, restored, and replaced 18.
Modern management puts emphasis on protecting genetic diversity to ensure ecosystem resilience and sustainable provision of services in the long run; a similar objective should apply to the digital economy. Currently, innovation diversity is not sufficiently emphasized in regulation, and often contrasts directly with the self-vested objectives of DPE orchestrators 19. Thus, ecological reasoning can be useful for understanding the processes and necessities of innovation in, and for, innovation-favourable regulation of the digital economy.
2. At the Meso Level: Comparing products to species
Species are fundamental units for understanding ecological phenomena. Each species is characterized by a unique combination of traits which allows it to occupy a certain ecological niche, i.e., a set of resources and range of environmental parameters that the species relies on. In digital markets, products can be conceived as analogous to ecological species. In economic systems, products (goods and services) occupy ‘niches’ 17, which are shaped by the product’s position in the value creation network, as well as its reliance on finances, non-financial resources, and customers. The Meso level of the 5M Framework therefore hosts species in nature and products in the digital economy.
Species and products can have flexible definitions. For example, narrower classifications such as subspecies or ecotypes help differentiate groups of individuals of the same species that exhibit some variations, often due to geographical or ecological factors. Conversely, when species boundaries are blurred, broader groupings like genera (which includes sister species) or functional guilds (comprising species with similar functions) may be used. Likewise, boundaries of products and services can be sometimes ambiguous when the focus of analysis requires finer-or coarser-grained categories. For example, the focus can be one online payment service, or a whole portfolio management of financial services. Flexibility of the product definition in the digital realm is a pre-requisite to explain the economics of DPEs. Certain activities such as, for example, facilitation of interactions by the ecosystem orchestrator, provision of attention 20, or supplying data by the user are novel forms of products produced and exchanged within the DPE.
Survival and proliferation are the integral objectives of all species and driven by a dynamic feedback loop between species traits emerging from genetic innovation and environmental niches. These dynamics shape three existential processes for species: i) foraging for food, ii) reproduction, and iii) avoiding fatal situations such as premature death or injury arising from unfavorable environmental conditions, predatorial threats, or competitive pressure. In the digital economy, such feedback loops are equally present. For instance, foraging can be seen akin to seeking resources necessary for growth such as finances, users’ attention, and data of users and their interactions. Reproduction of species can be likened to businesses releasing new versions of their products. In both nature and economy, reproduction provides an opportunity to establish innovation. Life-threatening pressures call for adaptation of species to avoid extinction; similarly digital products must continuously adapt to survive market fluctuations or restrictive regulation.
Theories of life history evolution explain how species optimize their traits to balance across existential processes. For example, the r/K selection theory stylizes the observation that species face a trade-off in the number of offspring they produce and the parental investments they may expend. Likewise, the orchestrator of an open DPE faces a tradeoff between the number and quality of complementors and products hosted on the platform. Just as mature, well-functioning ecosystems include both r-type (high number of offspring, low parental investment) and K-type species (low number of offspring, high parental investment) 21 diversity of complementors in terms of their product quality and the pace of development is an important factor for the DPE success. The concept of r- and K-strategy is yet an unexplored approach that can facilitate a better understanding of the product dynamics in digital economies.
Optimal foraging theory (OFT) may be another useful source of analogies and models. OFT explains how species adjust their strategies to optimize effort spent on acquiring food of different value and availability 21. Similarly, one may expect that the platform’s strategies of ‘foraging’ new domains of resources can be explained by optimizing metrics describing the required ‘effort’. For example, Uber often refrains from extending its operations into rural areas where the density of the ‘resource’, i.e., potential clients, is low. This is akin to animals spending less time in resource-poor than in resource-rich patches, despite similar travelling time between patches. Policies can affect the optimized metric through monetary or non-monetary instruments and thus the government can opt to steer platform’s foraging for resources in the direction of higher societal welfare.
3. At the Macro Level: Comparing natural ecosystems to digital platform ecosystems
The Macro level of the 5M Framework focuses on ecosystems. Both NEs and DPEs emerge from the dynamic interactions of their building blocks - communities of individuals from several species in NEs, and arrays of products (goods and services) in DPEs. For example, species are involved in feeding interactions, such as the predator-prey dynamics, or symbiosis, for example between the gut microbiome and the host. Complementors, the orchestrator, and other DPE members supply specific goods and provide specific services to members of an ecosystem which allows for the production of a collective output. Facilitating this are webs of financial flows, data flows, and influence relationships, among other relevant types of interactions. Figure 2 illustrates such interconnections for a stylized online travel agency.
The interactions within ecosystems are powered by boundary inputs. In NEs, they can take the form of solar energy inputs to the ecosystem's primary producers such as plants. Growth of DPEs is facilitated by financial investments and other inputs such as R&D and data. Tracking boundary inputs throughout the ecosystem allows for the assessment of efficiency with which energy or material are processed within the ecosystem. In NEs, boundary inputs are passed on from primary producers to their consumers and further up to top predators. This organises species into trophic levels, which measures the remoteness of species from the primary source of input. Considering financial or data inputs in a DPE, its members can also be assigned into ‘trophic levels’ according to their distance from the initial boundary inputs. As interconnections among the ecosystem parts can be complex, the trophic structure can help reveal the complexity and assist in understanding of the dependence of complementors and users on critical inputs into a DPE.
The degree of network complexity is determined by the number and types of interactions in an ecosystem. It has long been recognized that the networked structure is indispensable for the functioning of the system itself 22. This also speaks true for DPEs 23. Furthermore, complexity has been demonstrated to both promote and hinder stability in NEs (a so-called stability-complexity debate) 24. For instance, high species diversity may lower stability of some sub-systems while it may increase stability of the wider ecosystem 25. Similarly, some DPE network complexities provide for higher resilience while others hinder it, hence one objective of regulators may be to manage the complexity of the interaction structure of socially critical DPEs to increase their resilience.
Resilience is indeed an important characteristic of CASs and also increasingly an objective for DPE managers. Internal or external perturbations can result in a collapse or reorganisation of an ecosystem if stretched beyond its limit of adaptability. Flexible responses to perturbations are key in this phase to prevent a common ‘rigidity trap’ 26 characterized by low diversity and high connectivity of agents that begets lower resilience and may trigger collapse of the ecosystem. If collapse is avoided, a reorganization and subsequent new growth phase can reutilize the released material and elements 27. Regulation that recognizes, and is tailored to, different stages of ecosystem dynamics from growth to development and re-organisation, will be more effective than one that ignores the resilience cycle (for details, see Table 2).
4. At the Mega Level: Comparing biomes to societies
NEs are nested within even larger units, known as biomes, which combine the physical and biogeochemical environment and the totality of the biota therein. Likewise, DPEs are one form of an economy embedded into a wider society, which provides the political, economic, social, technological, environmental, legal, and other relevant contexts. Thus, the Mega level focuses on comparing biomes with the wider society.
Natural biomes are categorised mainly by their temperatures, precipitation patterns, and nutrient levels. Evidence shows that biota and ecosystems, even if separated geographically, display patterns of convergent evolution if they reside in the same biome type 28. For example, the succulent biome provides an arena for the evolution of different drought-adapted plant species across various continents. Likewise, convergence patterns may be observed in the formation and features of DPEs, which emerge from similarities of their societal contexts. For example, in the absence of good digital infrastructure, m-Pesa, a digital banking platform (DBP) providing PIN-secured SMS services for basic banking activities was launched in Kenya. Similar DBP models were launched across Latin America, another region with high mobile usage but poor digital infrastructures.
Furthermore, the abundance and variety of fundamental resources in NEs are major drivers of the evolution of the biome and its biota. For example, biomes such as tropical rainforests with abundant sunlight and precipitation provide suitable habitat for diverse biota with high biomass production whereas only drought-resistant biota with low production of biomass can survive the harsher conditions of succulent biomes. Like NEs, societies which provide abundant flow of resources such as finances and data are likely to generate a richer diversity of large DPEs and extensive innovations, compared to restricted or resource poor regions 29.
Therefore, biomes act as a theatre of evolution directing the evolutionary trajectories and the emerging functionalities of species and of ecosystems 30. Although there are several pragmatic studies ranking national economies on their digital developments based on various parameters of digital competitiveness, e.g., the IMD World Digital Competitiveness Index 29 or the Ease of Doing Digital Business 31, focused research is needed to analyze in greater detail the impact of digital and analog factors including resources, infrastructure, etc. on the DPE development and convergence patterns within their societal settings.
5. At the Meta Level: Comparing interactions in nature to interactions in the digital economy
Components at each level described above do not function or evolve in isolation. In the NEs, genes may influence the expression of other genes within an organism (Micro level); species feed on and compete for prey and engage in mutually beneficial relationships with other species within an ecosystem (Meso level); ecosystems can influence neighbouring ecosystems through exchanging resources (Macro level); and ultimately a biome can also influence other biomes (Mega level). Likewise, in the digital economy, technology, knowledge, and business strategy components often interact with each other (Micro level); products usually interact with other products (Meso level); DPEs may interact with other DPEs (Macro level); and finally, societies can influence other societies (Mega level). In both natural and digital contexts, at each level, interactions with other components complement the dynamically changing environment in forming pressures on or providing benefits to all participating components. The impact of interactions between components may ripple into the adjacent levels or even beyond. Their universal nature, a unified terminology, and existing tools for their analysis in ecological systems posit interactions between components as a separate level in the framework (Meta level).
Broadly, ecological interactions at all levels can have either a positive, neutral, or negative effect on the participating components, which results in five types of outcomes of pairwise interactions (a mutually neutral interaction is equivalent of no interaction, and called neutralism; see Fig. 3). Importantly, with time, interactions may transform. For example, a mutually beneficial interaction may gradually become parasitic if no controls or sanctions are placed to prevent exploiting the partner, and conversely, parasitism has often evolved into neutral or even beneficial interactions in nature.
Interactions creating positive effect on both involved components at all levels from Micro to Mega, are key enablers of life (see Table 1 for details). For example, bacteria cooperate by synthesizing various compounds to create a community matrix using molecular-based signalling systems. This cooperation enables the growth and maintenance of a biofilm that protects the members from external perturbations. Such interactions are extremely successful, such that biofilms cover almost all surfaces from rocks to human teeth. Likewise, in the digital economy, mutually beneficial cooperation plays a profound role. This is represented by the emergence of the term ‘complementors’ to refer to multiple decentralized firms that can co-create value on a massive scale. Such value co-creation became possible due to efficient coordination enabled through digitalization of firm interactions and transactions as well as artificial intelligence which allows efficient use of big data (see Table 1 for details).
However, it is often challenging to understand how mutualism and cooperation can exist as such interactions are prone to exploitation or free riding. Evolutionary game theory has deepened our understanding of the mechanisms and dynamics underlying the evolution and maintenance of cooperative behaviour in various ecological contexts from the gene- to species- levels and beyond 32,33. For example, only few species engage in the cooperative act of synthesizing costly compounds of biofilm while the benefit of the protective habitat it provides is enjoyed by the whole community. Defectors, i.e., individuals or even entire species that do not contribute to this process, can free ride on the production of others. Evidence indicates that a certain amount of free riding can be tolerated up to a threshold before the growth and even the survival of the whole community is jeopardized. Similar dynamics can be observed in the digital arena. For example, Wikipedia benefits the public, and is produced and edited by a comparatively very small number of volunteers. The quality of the content is maintained by the cooperative act of supervising, editing, and cleansing. If the amount of false or misleading contributions exceeds a certain limit, the popularity of this online platform is likely to drop with declining reliability of the information it offers.
Table 1
Positive interactions (+/+) in each M-level of NEs and their corresponding examples in DPEs
M-Level
|
Nature
|
Digital economy
|
Micro
|
Via synergistic interactions between genetic elements, called epistasis, the effect of one genetic element is enhanced by another genetic element.
Example: lung tumour growth is prominent if three mutations occur together, whereas the solitary mutations alone cause insignificant tumour growth.
|
Different components of knowledge, technology or business strategy can reinforce each other.
Example: Uber’s service is enabled by the combination of smartphones, mobile internet, and search-and-match algorithms which create a multiplicative effect.
|
Meso
|
Species can benefit from the presence of other species.
Example: corals and their symbiotic algae which benefit from the coral habitat and in turn provide nutrients.
|
Products can create a facilitating effect on each other.
Example: Collaboration between content producers on YouTube facilitates a synergistic growth through transfers of audiences across complementary contents.
|
Macro
|
Ecosystems can positively re-enforce each other.
Example: a freshwater lake ecosystem receives nutrients from the neighbouring forest ecosystem, while predators inhabiting the forests can visit the lake to feed on the fish.
|
Ecosystems can positively re-enforce each other.
Example: platforms such as Spotify, E-bay rely on the mobile operating system (MOS) infrastructures of Google and Apple to be downloaded on smart devices while the MOS need such apps for their value proposition.
|
Mega
|
Biomes can have positive effects on each other.
Example: global nutrient cycles or global circulation systems, such as the Gulf Stream or the El Niño connect distinct ecosystems from different biomes influencing the local niche conditions. Such currents can lead to milder temperatures during the winter or affect rainfall in some regions leading to higher or lower productivity of biota.
|
National economies can synergize each other.
Example: Governance of multinational DPEs benefits from cooperation of national policy agencies including informational exchange and uniting approaches.
|
On the other end of the spectrum are mutually negative interactions. In ecology, they stem from competing for limited resources, such as building blocks (amino acids) or energy (ATP) for protein synthesis during gene expression at the Micro level; food, mating partner, and territory at the Meso level; and habitat and nutritional resources at the Macro and Mega levels. Mutually negative interactions decrease expression of genes, fitness of species, and productivity of an ecosystem or biome in the short run, while in the long run such processes are crucial in promoting resilience and ensuring survival of the fittest. Numerous models have been developed in ecology to study the effects of competition 34. For example, the competitive Lotka–Volterra model demonstrates how one species competitively excludes another if both compete for a common resource, i.e., habitat or food. Similarly, digitalization amplifies the tendency of markets to tip in favour of one incumbent rather than allowing for co-existence of several competitors. Specific ecological models and insights regarding competition may be useful for a better understanding and regulation of DPEs; one novel analogy might be between intraguild predation, i.e., the killing and sometimes eating a potential competitor, and mergers and acquisitions (M&A) in certain, complex settings.
Other types of interactions include amensalism, a type of interaction, in which a component is inhibited or destroyed as a by-product of another organisms’ life cycle, while the opposite, positive effect as a by-product is called commensalism; in these two types of interactions another interacting component is not affected. Finally, under parasitism and predation one component is harmed by another that benefits from this interaction. Interactions in which at least one party suffers a negative effect, such as competition, predation, parasitism, and amensalism, are often referred to as antagonisms.